DHH: Future of Programming, AI, Ruby on Rails, Productivity & Parenting | Lex Fridman Podcast #474

0:00 No one anywhere who's serious believes

0:03 that cookie banners does anything good for anyone,

0:05 yet we've been unable to get rid of it.

0:07 This is the thing that really gets me about cookie banners too.

0:10 It's not just the EU.

0:12 It's the entire world.

0:13 You can't hide from cookie banners anywhere on this planet.

0:16 If you go to goddam Mars on one

0:18 of Elon's rockets and you try to access a webpage,

0:22 you'll still see a cookie banner.

0:23 No one in the universe is safe from this nonsense.

0:26 It sometimes feels like we're barely better off.

0:29 Like webpages aren't that different from what

0:32 they were in the late '90s, early 2000s.

0:34 They're still just forms.

0:35 They still just right to databases.

0:37 A lot of people, I think,

0:39 are very uncomfortable with the fact that they are essentially CRUD monkeys.

0:43 They just make systems that create read, update or delete row in a database,

0:49 and they have to compensate

0:51 for that existential dread by overcomplicating things.

0:56 That's a huge part of the satisfaction of driving a race car

0:59 is driving in at the edge of adhesion, as we call it.

1:02 Where you're essentially just a tiny movement away from spinning out.

1:06 Doesn't take much, then the car starts rotating.

1:08 Once it starts rotating, you lose grip and you're going for the wall.

1:13 That balance of danger and skill is what's so intoxicating.

1:21 The following is a conversation with David

1:24 Heinemeier Hansson, also known as DHH.

1:27 He is a legend in the programming and tech world,

1:31 brilliant and insightful, sometimes controversial, and always fun to talk to.

1:37 He's the creator of Ruby on Rails,

1:39 which is an influential web development framework

1:42 behind many websites used by millions of people,

1:44 including Shopify, GitHub, and Airbnb.

1:49 He is the co-owner and CTO of 37signals that created Basecamp, HEY and ONCE.

1:57 He is a "New York Times" bestselling author,

2:00 together with his co-author, Jason Fried of four books,

2:04 "Rework," "Remote," "Getting Real," and "It Doesn't Have to Be Crazy

2:09 at Work." And on top of that, he's also a race car driver,

2:13 including being a class winner at the legendary 24-hour Le Mans Race.

2:19 This is a Lex Fridman Podcast.

2:21 To support it, please check out our sponsors

2:23 in the description and consider subscribing to this channel.

2:27 And now, dear friends, here's DHH.

2:32 For someone who became a legendary programmer,

2:35 you officially got into programming late in life,

2:39 and I guess that's because you tried to learn

2:42 how to program a few times and you failed.

2:44 So can you tell me the the full story,

2:48 the saga of your failures to learn programming?

2:50 Was Commodore 64 involved?

2:53 Commodore 64 was the inspiration.

2:55 I really wanted a Commodore 64.

2:57 That was the first computer I ever sat down in front.

3:00 And the way I sat down in front of it was I was five years

3:03 old and there was this one kid on my street who had a Commodore 64.

3:07 No one else had a computer,

3:08 so we were all the kids just getting over there and we were

3:12 all playing "Yie Ar Kung-Fu." I dunno if you've ever seen that game.

3:14 It was one of the original fighting games.

3:16 It's really a great game.

3:17 And I was playing that for the first time at five years old.

3:20 And we were like seven kids sitting up in this one kid's bedroom,

3:24 all taking our turn to play the game.

3:27 And I just found that unbelievably interesting.

3:30 And I begged and I begged and I begged my dad, could I get a computer?

3:35 And he finally comes home.

3:37 He's like, "I got your computer." I was like, yes, my own Commodore 64.

3:42 And he pulls out this black green and blue keyboard.

3:47 That's an Amstrad 464.

3:49 I was like, dad, what's this?

3:54 The disappointment.

3:55 This is not a Commodore 64.

3:57 But it was a computer.

3:58 So I got my first computer at essentially six years old that Amstrad 464.

4:04 And of course, the first thing I wanted to do, I wanted to play video games.

4:07 And I think the computer, which he by the way,

4:10 had traded for TV and a stereo recorder or something like that.

4:15 Came with like two games.

4:16 One was this "Frogger" game where you had to escape from underground.

4:20 It was actually kind of dark.

4:22 Like this frog, you're trying to get it out from underground.

4:24 And I was pretty bad at it.

4:25 And I only had those two games.

4:27 And then I wanted more games.

4:29 And one way to get more games when you're a kid who don't have a lot of money,

4:33 I can't just buy a bunch of games, is to type them in yourself.

4:37 Back in '84, '85, magazines would literally print source code at the back

4:44 of their magazines and you could just sit and type it in.

4:46 So I tried to do that and it would take

4:48 like two hours to print this game into the Amstrad.

4:53 And of course I'd make some spelling

4:55 mistake along the way and something wouldn't work, and the whole thing.

4:58 I wasn't that good of English.

5:00 I was born in Denmark.

5:01 So I was really trying to get into it because I wanted all these games.

5:05 I didn't have the money to buy 'em.

5:07 And I tried quite hard for quite a while to get into it,

5:10 but it just never clicked.

5:11 And then I discovered the magic of piracy.

5:14 And after that I kind of basically just

5:16 took some time off from learning to program because,

5:19 well, now suddenly I had access to all sorts of games.

5:22 So that was the first attempt, like around six, seven years old.

5:26 And what's funny is I remember these fragments.

5:30 I remember not understanding the purpose of a variable.

5:34 If there's a thing and you sign something,

5:36 why would you assign another thing to it?

5:37 So for some reason, I understand it constants.

5:40 A constants made sense to me, but variables didn't.

5:43 Then maybe I'm 11 to 12, I've gotten into the Amiga at this point.

5:48 The Amiga, by the way, still perhaps my favorite computer of all time.

5:53 I mean, this is one of those things where you're like,

5:54 people get older and they're like, "Oh, the music from the '80s amazing." To me,

5:58 even as someone who loves computers, who loved new computers,

6:01 the Amiga was this magical machine that was made

6:07 by the same company that produced the Commodore 64.

6:10 And I got the Amiga 500, I think in '87.

6:16 Look at this sexy thing that is a sexy machine right there.

6:19 This is from an age by the way, where computing wasn't global in the same sense.

6:23 The different territories had different computers that were popular.

6:27 The Amiga was really popular in Europe,

6:29 but it wasn't very popular at all in the US as far as I understand it.

6:33 It wasn't popular in Japan.

6:34 There were just different machines.

6:35 The Apple II was a big thing in the US.

6:38 I'd never even heard of Apple in the '80s in Copenhagen.

6:42 But the Amiga 500 was the machine that brought me to want to try it again.

6:46 And you know what's funny?

6:47 The reason I wanted to try it again was I remembered the first time to learn.

6:51 And then there was this programming language that was literally called EasyAMOS,

6:57 like the easy version of AMOS.

6:59 I'm like, if it's EasyAMOS, how hard can it be?

7:01 I gotta be able to figure this out.

7:04 And this time I tried harder.

7:07 I got into conditionals, I got into loops,

7:09 I got into all these things and I still, I couldn't do it.

7:13 And on the second attempt, I really got to the point of like,

7:17 maybe I'm not smart enough,

7:19 maybe programming is just not for me, maybe it's too much math.

7:23 Like I like math in this sort of superficial way.

7:26 I don't like it in the deep way.

7:27 That's some on my perhaps slightly nerdier

7:29 friends did who I had tremendous respect for.

7:31 Like I'm not that person.

7:34 I'm not the math geek who's gonna figure it all out.

7:37 So after that attempt with EasyAMOS and failing to even get,

7:43 I don't even think I completed one even very basic game.

7:47 I thought the program's just not for me.

7:49 I'm gonna have to do something else.

7:50 I still love computers.

7:51 I still love video games.

7:53 I actually at that time had already begun

7:55 making friends with people who knew how to program,

7:57 who weren't even programming EasyAMOS.

7:59 They were programming freaking assembler.

8:02 And I would sit down and just go, how do you,

8:05 the moves and the memories and the copies, how do you even do this?

8:09 I don't even understand how you go from this to Amiga demos, for example.

8:14 That was the big thing with the Amiga.

8:16 It had this wonderful demo scene in Europe.

8:19 It's this really interesting period of time in Amiga's history where you had all

8:25 these programmers spread out mostly all over

8:27 Europe who would compete on graphic competitions,

8:32 where you could probably bring one of these different...

8:35 On that thing?

8:35 On this thing, they would make these little almost like music videos,

8:41 combining some midi music, combining some cool graphics.

8:44 And they would do all of it in like 4K, four kilobytes.

8:48 That is not 4Ks of revolution, four kilobytes of memory.

8:52 And I just thought that was such a cool scene.

8:54 This was obviously pre-internet.

8:56 It was even pre BBS, bulletin board systems to some extent.

9:02 It was you swap your demo software with someone else

9:05 by sending them a disk in the mail, like the 3.5s.

9:11 And I was enamored with that whole scene.

9:13 I was enamored with what they were able to create and I just wanted to be

9:16 a part of it even though I kind of didn't have any skills to contribute.

9:20 And that's how I got into running BBSs.

9:22 I didn't learn programming then and I

9:25 wouldn't learn programming until much later, until I was almost 20 years old.

9:29 The bulletin board systems existed in this funny

9:31 space where they were partly a service to the demo scenes allowing all

9:35 these demo groups to distribute their amazing demos.

9:40 And then it was also a place to trade piracy software, pirated software.

9:44 And I ended up starting one of those when I

9:47 was 14 years old in my tiny little bedroom in Copenhagen.

9:51 I had my, at that point, Amiga 4000.

9:54 I had three telephone lines coming in into my tiny room.

9:59 [Lex] Nice.

10:00 Which is funny because again, I'm 14 years old.

10:02 By the time I was in Sterling, my third line,

10:05 you had to get someone from the telephone company to come do it.

10:07 I get this guy and he's just looking around like, "What is this?

10:11 Why the hell is a 14-year-old having

10:13 three phone lines into their tiny little bedroom?

10:16 What's going on here?

10:18 Why are all these modems blinking red and black

10:22 and making funny sounds?"- Did your parents know?

10:25 They did and they didn't.

10:27 They knew I had the phone lines.

10:28 They knew I had the computer.

10:30 I don't think they really understood that I

10:32 was trading pirated software that was both illegal,

10:35 and whatever else was going on.

10:38 Oh, we should probably say that in Europe,

10:40 maybe you can comment on this in, especially in Eastern Europe,

10:43 but Europe in general,

10:44 piracy I think was more acceptable than it was in the United States.

10:48 I don't know there maybe it's just my upbringing.

10:51 Even that conversation wasn't present.

10:53 I never spoke to anyone growing up in Denmark.

10:56 The piracy is wrong.

10:57 Had any moral qualms whatsoever about piracy.

11:00 It was just completely accepted that you're a kid, you won a lot of games.

11:05 You don't have a lot of money, what do you do?

11:07 You trade.

11:08 Yeah.

11:09 Some people would occasionally buy a game.

11:10 I mean, I once bought a second Master System, and I bought one game.

11:16 Because that was what I could afford.

11:17 I got "After Burner II." I dunno if you've ever played that game.

11:20 It's pretty bad implementation on the second Master System,

11:25 but it was like 600 crowners.

11:27 And I was making money at that time doing newspaper delivery.

11:32 I had to do that for a month to afford one game.

11:36 I like video games way too much to wait a month just to get one game.

11:40 So piracy was just the way you did it.

11:42 And that was how I got into running this bulletin board system,

11:46 being part of the demo scene, being part of the piracy scene to some extent.

11:50 And then also at some point realizing, oh,

11:53 you can actually also make money on this and this can fund

11:56 buying more phone lines and buying more modems and buying more Amigas.

12:01 Oh yeah, that was one of the demo parties.

12:03 These were amazing things.

12:04 What am I looking at?

12:06 Isn't that amazing?

12:07 Look at all those CRT monitors.

12:08 All these CRT monitors.

12:09 Again, when I was 14, I don't understand fully why my parents allowed this.

12:15 But I traveled from Copenhagen,

12:18 the capital of Denmark to Aarhus this tiny little town in Jutland on the train

12:23 with a bunch of dudes who were like late teens and their 20s.

12:28 I'm 14 years old.

12:29 I'm lugging my 14-inch CRT monitor with my computer

12:32 in the back to go to the party.

12:35 That was what it was called.

12:36 That was the biggest demo scene party at that time.

12:38 And it was exactly as you see in that picture,

12:40 thousands of people just lining up with their computers,

12:43 programming demos all day long, and trading these things back and forth.

12:48 That's kind of awesome.

12:49 Not gonna lie.

12:50 It's a little ridiculous.

12:52 It's totally awesome.

12:53 And I miss it in ways where the internet has connected people in some ways,

12:58 but the connection you get from sitting right

13:00 next to someone else who has their own

13:02 CRT monitor who's lugged at halfway around

13:04 the country to get there is truly special.

13:07 Because it was also just this burst of creativity.

13:10 You're constantly running around.

13:11 You're constantly surrounded by people who are

13:13 really good at what they could do.

13:14 They're really good at programming computers.

13:15 It's infectious.

13:17 It was part of that pang I felt then going like,

13:20 oh man, why can't I figure this out?

13:22 I mean, why can't I even figure out EasyAMOS?

13:26 It's kind of frustrating.

13:28 But on your third attempt, you a little more successful?

13:31 So third attempt is when I start getting it,

13:33 this is when I start helping out, let's say build things for the internet.

13:39 So around '95, I think it is, or '96, I discovered the internet.

13:45 Actually, in ninth grade.

13:46 That was my first experience.

13:47 I went to some university in Denmark.

13:52 And in ninth grade, we had this excursion and they sat us

13:56 down in front of a computer and the computer had Netscape Navigator,

13:59 the first version, or maybe it was even the precursor to that.

14:03 And they had a text editor.

14:05 And us kids just got like, "Hey,

14:06 built something on the internet." And it was just

14:07 HTML and the first thing you do is like,

14:09 oh, I can make the text blink by just putting in this tag and saving it.

14:14 That moment, that was actually when I reawaken the urge

14:20 to wanna learn the program because I got a positive experience.

14:23 All the other experiences I had with programming was I'd spend

14:26 hours typing something in, I'd click Run and it wouldn't work.

14:30 And I'd get an error message that made no sense to me as a kid,

14:34 either at six or seven or at 12.

14:36 And here I am sitting in front of a computer

14:39 connected to the internet and I'm making text blink.

14:42 I'm making it larger.

14:43 I'm turning it into an h1 or an h2.

14:46 And these guys out here,

14:47 "We just did it for like an hour and a half." And suddenly I go, oh,

14:51 I can make things for the internet that someone

14:53 in Germany can be able to access and see,

14:56 and I don't have to ask anyone for permission.

14:59 This is super cool.

15:00 I gotta do more of this.

15:01 So I got into to the internet.

15:02 I got into working with HTML and I

15:05 still had all these friends from these demo parties.

15:07 And I started working with them on creating gaming websites.

15:11 I'd rather buy the video games, I'd review him.

15:13 This was another good way of getting new video games

15:15 was to walk down to some store and say like, hey, I'm a journalist.

15:19 I'm like this 15-year-old kid.

15:20 And they're looking at you,

15:22 "You're a journalist?" Yeah, can I borrow some games?

15:25 Because this was when games moved on to the PlayStation and these other things.

15:28 You couldn't just as easily pirate, at least not at first.

15:31 So I went down there,

15:33 did all that, and that started the journey of the internet for me.

15:37 It started working on these gaming websites, working with programmers,

15:40 figuring out that I could do something, I could work on the HTML part.

15:44 It's not really programming, but it kind of smells like it.

15:46 You're talking to a computer.

15:48 You're making it put text on the screen

15:50 and you're communicating with someone halfway around the world.

15:53 So that became my pathway back into programming.

15:56 And then slowly I picked up more and more of it.

15:59 First website I did with someone,

16:02 one of these programs from the demo scene that was dynamic was ASP.NET.

16:07 I wasn't even actually called .NET.

16:08 That was what we started on.

16:10 And then we moved on to PHP, and PHP was when I finally got it.

16:14 When it finally clicked.

16:15 When conditionals and loops and variables and all of that stuff started

16:21 to make sense enough to me that I thought I can do this.

16:26 So would it be fair to say that we wouldn't have DHH without PHP,

16:29 and therefore you owe all of your success to PHP?

16:33 100% that's true.

16:35 And it's even better than that because as PHP to me didn't

16:39 just give a start in terms of making my own web applications,

16:43 it actually gave me a bar.

16:44 In many ways, I think the pinnacle of developer,

16:48 web developer ergonomics is late '90s PHP,

16:52 you write this script, you FTP it to a server and instantly it's deployed.

16:58 Instantly it's available.

16:59 You change anything in that file and you reload.

17:02 Boom, it's right there.

17:03 There's no web servers, there's no setup.

17:06 There's just an Apache that runs mod PHP.

17:10 And it was essentially the easiest way to get a dynamic webpage up and going.

17:15 And this is one of the things I've been

17:17 chasing that high for basically the rest of my career.

17:20 That it was so easy to make things for the internet in the mid to late '90s.

17:26 How did we lose the sensibilities that allowed us to not just work this way

17:31 but get new people into the industry to give

17:34 them that was success experiences that I had.

17:37 Adding a freaking blink tag to an HTML page, FTPing a PHP page to an Apache

17:44 web server without knowing really anything about anything,

17:47 without knowing anything about frameworks, without knowing anything about setup.

17:51 All of that stuff have really taken us to a place

17:54 where it sometimes feels like we're barely better off.

17:58 Like webpages aren't that different from what

18:01 they were in the late '90s, early 2000s.

18:03 They're still just forms.

18:04 They still just write to databases.

18:06 A lot of people, I think are very uncomfortable

18:08 with the fact that they are essentially CRUD monkeys.

18:12 They just make systems that create read, update or delete row in a database,

18:18 and they have to compensate

18:20 for that existential dread by overcomplicating things.

18:25 Now that's a bit of a character.

18:26 There's more to it.

18:27 And there's things you can learn

18:28 from more sophisticated ways of thinking about this.

18:30 But there's still an ideal here, which is why I was so happy you had

18:34 Pieter Levels on, because he still basically works like this.

18:36 And I look at that and go like, man, that's amazing.

18:39 Yeah, you're chasing that high.

18:40 He's been high all along using PHP, jQuery and SQLite.

18:48 I think it's amazing because he's proving

18:50 that this isn't just a nostalgic dream.

18:53 He's actually doing it.

18:54 He's running all these businesses.

18:56 Now some of that is, as he would admit up first upfront,

19:00 is that he's just one guy.

19:02 And you could do different things when you just one guy.

19:04 When you're working in a team, when I started working on the team,

19:08 when I started working with Jason Fried on Basecamp,

19:11 we at first didn't use version control together.

19:16 I used version control for myself.

19:17 And then I thought, do you know what?

19:19 Designers, ah, they probably not smart enough to figure out CBS.

19:24 And therefore I just like, no, no, no, you just FTP it up.

19:27 You just FTP it.

19:28 I knew they knew how to do FTP.

19:30 And then after the third time I had overridden their changes.

19:33 I was like, goddammit.

19:34 I gotta teach Jason CBS to not do that again.

19:38 But I think there's still way more truth to the fact

19:43 that we can work the way we did in the '90s,

19:45 work the way Pieter works today, even in the team context.

19:49 And that we've been far too willing to hand over far

19:53 too much of our developer ergonomics to the merchants of complexity.

19:57 And you've been chasing that with Rails 8.

20:00 So how do you bring all the cool features

20:03 of a modern framework and make it no build,

20:06 make it as easy to create something and to ship

20:10 it as it was in the '90s with just PHP?

20:13 It's very difficult for me to beat the Pieter Levels approach.

20:19 It's so easy to just ship some PHP.

20:22 And it should be.

20:23 Why should it be harder than that?

20:25 Our computers today are almost infinitely faster

20:29 than what they were in the '90s.

20:30 So shouldn't we be able to work in even easier ways?

20:34 We should be looking back on the '90s ago,

20:36 like, oh, that was way too complicated.

20:38 Now we have more sophisticated technology that's way faster and it

20:41 allows us to work in these easier to use ways.

20:44 But that's not true.

20:46 But now you can see the line I draw in my work with Ruby on Rails,

20:50 and especially with Rails 8.

20:51 No build to me is reaching back to that '90s feeling and going,

20:57 now we can do some of those things without giving up on all the progress.

21:02 Because I do think you can get too nostalgic.

21:04 I do think you can start just

21:07 fantasizing that everything was better in the '90s.

21:09 It wasn't.

21:10 I mean, I was there.

21:11 There was a lot of things that sucked.

21:13 And if we can somehow find a way to combine the advantages and advances we've

21:18 had over the past 20 years with that ease of developer ergonomics, we can win.

21:24 No build is a rejection of the part of web

21:28 development I've hated the most in the past 10,

21:31 15 years, which is the JavaScript scene.

21:35 And I don't say that as someone who hate JavaScript.

21:38 I mean, I often joke that JavaScript.

21:39 It's my second favorite program language.

21:42 It's a very distant second.

21:43 Ruby is by far and away number one, but I actually like JavaScript.

21:47 I don't think it's a bad language.

21:49 It gets a lot of flack.

21:51 People add a string of two plus a one and it gives something nonsense.

21:55 And I just go like, yeah, but why would you do that?

21:58 Just don't do that.

21:59 The language is actually quite lovely, especially the modern version.

22:02 ES6 that really introduced a proper class syntax to it.

22:06 So I could work with JavaScript in many of the same

22:09 ways that I love working with Ruby made things so much better.

22:13 But in the early 2010s until quite recently,

22:19 all of that advancement happened in pre-processing, happened in built pipelines.

22:25 The browsers couldn't speak a dialect

22:27 of JavaScript that was pleasant to work with.

22:29 So everyone started to pre-filing their JavaScript to be able

22:33 to use more modern ways of programming with a browser

22:37 that was seen as stuck with an ancient version

22:40 of JavaScript that no one actually wanted to work with.

22:43 And that made sense to me, but it was also deeply unpleasant.

22:46 And I remember thinking during that time, the dark ages,

22:49 as I refer to them with JavaScript, that this cannot be the final destination.

22:56 There's no way that we have managed to turn

22:59 the internet into such an unpleasant place to work.

23:03 Where I would start working on a project

23:06 in JavaScript using webpack and all of these dependencies.

23:09 And I would put it down for literally

23:11 five minutes and the thing wouldn't compile anymore.

23:14 The amount of churn that the JavaScript community,

23:17 especially with its frameworks and its tooling went through

23:20 in the decade from 2010 to 2020 was absurd.

23:25 And you had to be trapped inside of that asylum to not

23:31 realize what an utterly perverse situation we had landed ourselves in.

23:38 Why does everything break all the time?

23:41 I mean, the joke wouldn't be just that the software would break,

23:44 that would annoy me personally.

23:45 But then I'd go on Hacker News and I'd see

23:48 some thread on the latest JavaScript release of some framework.

23:51 And the thread would be like, someone would ask, "Well,

23:55 aren't we using the thing we just used

23:57 three months ago?" And people would be like, "That thing is so outdated.

24:00 That's so three months ago.

24:02 You gotta get with the new program.

24:04 We're completely rewriting everything for the oomph-teen

24:07 time." And anything you've learned

24:09 in the framework you've been spending the last

24:12 amount of time on, it's all useless.

24:14 You gotta throw everything out and you gotta start over.

24:16 Why aren't you doing it, stupid idiot?

24:18 Is that a kind of mass hysteria

24:20 that took over the developer community you think?

24:22 Like where you have to keep creating new frameworks and new frameworks.

24:25 And are we past that dark age?

24:29 I think we're getting out of it,

24:31 and we're getting out of it because browsers have gotten so much better.

24:35 There was a stagnation in browser technology.

24:38 Some of it was an overhang all the way back from IE5.

24:42 So IE5 essentially put the whole

24:45 internet development experience into a deep freeze,

24:48 because Microsoft won the browser wars in the mid 2000s,

24:53 and then they basically disbanded their browser development team.

24:56 'Cause they're like, "All right, job done.

24:57 We don't need any more innovation on the internet.

25:00 Can we just go back to writing Windows forms or something now that we control

25:03 everything?" And it really wasn't until obviously Firefox

25:09 kind of kindled a little bit of something.

25:11 Then Chrome got into the scene and Google

25:14 got serious about moving to web forward.

25:16 That you had a kindling of maybe the browser could be better.

25:20 Maybe the browser wasn't frozen in time in 2005.

25:23 Maybe the browser could actually evolve like

25:27 at the development platform that it is.

25:30 But then what happened was you had a lot

25:32 of smart people who poured in to the web because the web turned out to be

25:38 the greatest application development platform of all time.

25:41 This was where all the money was being made.

25:43 This was where all the billionaires were being minted.

25:45 This was where the Facebooks and whatever of the world came to be.

25:49 So you had all of this brain power applied

25:53 to the problem of how to work with the web.

25:56 And there were some very smart people with some,

25:59 I'm sure very good ideas who did not

26:04 have programmer happiness as their motivation number one.

26:08 They had other priorities and those priorities allowed them

26:15 to discount and even rationalize the complexity they were injecting everywhere.

26:21 Some of that complexity came from organizational structure.

26:24 When you have a company like Facebook, for example,

26:26 that does depend on the web and wanna push it forward,

26:28 but have sliced the development role job into these tiny little niches.

26:35 I'm a front-end glob pipeline configurator.

26:41 Oh yeah, well, I'm a front-end whatever engineer.

26:44 And suddenly the web developer was no longer one person.

26:48 It was 15 different roles.

26:50 That in itself injected a ton of complexity.

26:52 But I also wanna give it the bold case here,

26:56 which was that some of that complexity was

26:58 necessary to get to where we are today.

27:01 That the complexity was a bridge.

27:04 It wasn't the destination,

27:05 but we had to cross that bridge to get to where we are today,

27:08 where browsers are frankly incredible.

27:11 The JavaScript you can write in a text file and then

27:14 serve on a web server for a browser to ingest is amazing.

27:18 It's actually a really good experience.

27:19 You don't need any pre-processing.

27:20 You could just write text files,

27:22 send them to a browser, and you have an incredible development.

27:25 And we should also say that it can kind of be broken, at least the HTML,

27:29 but even the JavaScript could be a little bit broken

27:31 and it kind of still works like maybe a half-ass works.

27:35 But like, just the amount of mess of smelly

27:40 code that a browser has to deal with is insane.

27:44 This is one of the hardest problems

27:46 in computing today is to parse the entire internet.

27:50 Because thankfully for us as web developers,

27:54 but perhaps not so much for the browser developers,

27:57 every webpage that has ever been created minus

28:00 the brief period with Flash still runs today.

28:03 The webpage I did in ninth grade would

28:06 render on a modern browser today, 30 years later.

28:10 That's crazy.

28:11 That is completely crazy.

28:12 When you think about the amount of evolution we've had with the web,

28:16 how much better we've made it, how many more standards browsers have adopted.

28:20 It's essentially an Apollo project today to create a new browser,

28:24 which is why it doesn't happen very often,

28:26 which is why even companies like Microsoft had to throw into towel and say,

28:30 "We can't do it." Now I actually don't think that's good for the web.

28:34 There is the danger of the monoculture if we

28:36 just get a single browser engine that runs everything,

28:38 and we are in danger of that.

28:40 I love the fact that the Ladybird project, for example,

28:42 is trying to make a new browser engine from scratch.

28:46 I've supported that project.

28:47 I would encourage people to look into that.

28:49 It's really a wonderful- Nice.

28:52 thing.

28:52 It's staffed by a bunch of people who

28:54 worked on other browser projects in the past.

28:57 Truly independent web browser.

28:59 We really need that.

29:01 But I can hold that thought in my head.

29:03 At the same time, I hold the thought in my head that Google's

29:08 Chrome was pivotal to the web surviving as the premier web development platform.

29:14 If it had not been for Google and their entire

29:18 business depending on a thriving open web, Apple, Microsoft,

29:24 I think would've been just as fine to see

29:27 the web go away to disappear into being something

29:30 that's just served native web applications or native mobile

29:34 applications and native desktop applications that they could completely control.

29:38 So I have all sorts of problems with Google, but it's not Chrome.

29:43 Chrome is a complete gift to web developers everywhere,

29:47 to the web as a development platform.

29:49 And they deserve an enormous amount of credit, I think for that.

29:54 Even if it's entangled with their business model and half

29:58 of Chrome is code that spies on you or informs targeted ads.

30:03 And a bunch of things, I'm not a big fan of.

30:06 I can divorce that from the fact that we need champions in the corner

30:09 of the web who have trillions of dollars

30:12 of market cap value riding on the open web.

30:16 We're gonna take tangents upon a tangents upon a tangents.

30:19 So let's go to Chrome.

30:20 I think Chrome positive impact on humanities

30:24 is immeasurable for reasons that you just described.

30:27 On the technology front, the features that present the competition they created,

30:31 it's spurred on this wonderful flourishing of web technologies.

30:35 But anyway, I have to ask you about the recent stuff with the DOJ,

30:40 trying to split up Chrome and Google.

30:43 Do you think this is a good idea?

30:45 Do you think this does harm?

30:47 It's a disaster.

30:49 And I say that as someone who's been very sympathetic to the antitrust fight,

30:54 because I do think we have antitrust problems in technology.

30:58 But the one place where we don't have them, by and large,

31:03 is with browsers, is with the tools we use to access the open web.

31:08 First of all, we have Firefox.

31:10 Now Firefox is not doing all that great.

31:13 And Firefox has been propped up by Google for many years to deter from exactly

31:18 what's going on with the DOJ that they were the only game in town.

31:22 Apple has Safari.

31:24 I have a bunch of problems with Apple too, but I love Safari.

31:27 I love the fact that we have a premier browser running on a premier

31:31 operating system that people can't turn the web into just a Chrome experience.

31:36 But I also think that the open web needs this trillion dollar champion,

31:42 or at least benefits from it.

31:44 Maybe it doesn't need it, but it certainly benefits from it.

31:47 And of all the things that are wrong with monopoly formation in technology,

31:52 Chrome is the last thing.

31:54 And this is why I get so frustrated sometimes about

31:56 the anti or the monopoly fight that there are real problems.

32:01 And we should be focusing on the premier problems first.

32:03 Like the toll booths on our mobile phones.

32:08 They're far bigger problems.

32:09 It's not the open web.

32:10 It's not the tools that we use to access the open web.

32:12 If I don't want to use Chrome, if my customers of my businesses that run

32:18 on the internet don't wanna use Chrome, they don't have to.

32:21 We're never forced to go through it.

32:23 The open internet is still open.

32:25 So I think it's a real shame that the DOJ

32:29 has chosen to pursue Google in this way.

32:32 I do think there are other things you

32:33 can nail Google for and their ad monopoly maybe,

32:36 or the shenanigans done in controlling both sides of the ad

32:41 ledger that they both controlled the supply and the demand.

32:45 There are problems.

32:46 Chrome, isn't it?

32:47 And you end up making the web much worse.

32:50 And this is the thing we always gotta remember when we think about legislation,

32:54 when we think about monopoly fights is you may not like how things look today.

32:59 And you may wanna do something about it, but you may also make it worse.

33:04 The good intentions behind the GDPR in Europe currently has amounted to what?

33:11 Cookie banners that everyone on the internet hates.

33:15 That helps no one do anything better,

33:17 anything more efficient that saves no privacy in any way,

33:21 shape or form has been a complete boondoggle

33:23 that has only enriched lawyers and accountants and bureaucrats.

33:30 Yeah, you said that the cookie banner is a monument for why

33:34 Europe is losing is doing the worst of all the regions in tech.

33:41 It's a monument to good intentions leading straight to hell.

33:47 And the Europe is actually world class

33:50 in good intentions leading straight to hell.

33:53 So hell is the cookie accept button that you have to accept all cookies.

33:58 That's what hell looks like over and over.

34:00 You don't actually ever get to the web page.

34:03 Just on a human scale,

34:04 try to imagine how many hours every day are wasted clicking that away.

34:10 And how much harm we've done to the web

34:13 as a platform that people enjoy because of them.

34:17 The internet is ugly in part because of cookie banners.

34:21 Cookie banners were supposed to save us from advertisement,

34:25 and advertisement can make the web ugly.

34:27 There's plenty of examples of that, but cookie banners

34:29 made the entire internet ugly in one felt swoop.

34:33 And that's a complete tragedy.

34:34 But what's even worse,

34:36 and this is why I call it out as a monument to everything the EU gets wrong,

34:40 is that we have known this for a decade.

34:42 No one anywhere who's serious believes

34:44 that cookie banners does anything good for anyone.

34:48 Yet we've been unable to get rid of it.

34:50 There's this one piece of legislation that's now I think 10 or 12 years old.

34:54 It's complete failure on every conceivable metric.

34:57 Everyone hates it universally, yet we can't seem to do anything about it.

35:01 That's a bankruptcy declaration for any body of bureaucrats who

35:07 pretend or pretend to make things better for not just citizens,

35:12 but people around the world.

35:14 This is the thing that really gets me about cookie banners too.

35:17 It's not just the EU.

35:19 It's the entire world.

35:20 You can't hide from cookie banners anywhere on this planet.

35:23 If you go to goddamn Mars on one

35:26 of Elon's rockets and you try to access a webpage,

35:29 you still see a cookie banner.

35:30 No one in the universe is safe from this nonsense.

35:34 Probably the interface on the rocket.

35:36 It'll be slower.

35:37 You'll have basically 150 second ping time.

35:42 So it'll take you 45 seconds just to get through the cookie banners from Mars.

35:47 All right, let's walk back up the stack

35:49 of this recursive tangents we've been taking.

35:51 So Chrome, we should say, at least in my opinion, is not winning unfairly.

35:58 It's winning in the fairway by just being better.

36:03 It is, if I was gonna steal man the other side just for a half second,

36:07 people would say, well, maybe yes,

36:10 most people do sort of begrudgingly agree that is a pretty good browser.

36:15 But then they'll say the reason it got dominance was distribution.

36:19 And the reason it got distribution was because Google also controls Android,

36:23 and therefore can make Chrome the default browser on all these phones.

36:27 Now I don't buy that.

36:29 And the reason I don't buy that is because on Android,

36:32 you are actually allowed to ship a different browser

36:36 that has a browser engine that's not the same as Chrome.

36:38 Unlike on iOS, where if you wanna ship a browser, Chrome, for example,

36:43 ships for iOS, but it's not Chrome, it's Safari wrapped in a dress.

36:47 And every single alternative browser on iOS have to use the Safari web engine.

36:54 That's not competition.

36:55 That's not what happened on Android.

36:57 Again, I think there are some nuances to it,

36:59 but if you zoom out and you look at all

37:02 the problems we have with big tech, Chrome is not it.

37:05 Chrome won on merits.

37:07 I begrudgingly have switched to Chrome on that realization alone.

37:12 As a web developer, I just prefer it.

37:14 I like Firefox in many ways.

37:16 I like the ethos of it, but Chrome is a better browser than Firefox full stop.

37:21 And by the way, we've never mentioned Edge.

37:23 Edge is also a good browser- Because it's also Chrome in a dress.

37:28 But it never gets to love.

37:29 I don't think I've ever used Bing, and I'm sure Bing is really nice.

37:34 Maybe you have, because you know what?

37:35 Is Bing in a dress?

37:37 What?

37:38 DuckDuckGo.

37:38 Which is actually- What?

37:39 the search engine that I use.

37:41 DuckDuckGo gets its search results from Bing, or at least it used to.

37:45 If they change that, that would be news to me.

37:48 Well, maybe everything is just a wrap or a dress.

37:52 Everything is wearing a dress underneath.

37:54 There's some other.

37:54 There's some of that.

37:56 Turtles, all the dress is all the way down.

37:57 Okay, what were we talking about?

37:58 They, we got there from JavaScript and from you learning how to program.

38:03 So eventually, the big success story is when you built a bunch

38:09 of stuff with PHP and you were like actually shipping things.

38:14 And that's when the Ruby story came.

38:17 So what your big love affair with programming began there?

38:23 So can you take me there?

38:24 What is Ruby?

38:25 Tell the story of Ruby.

38:27 Explain Ruby to me.

38:28 PHP was what converted me from just being able to fondle HTML and turn

38:34 out some web pages to actually being able to produce web applications myself.

38:40 So I owe a tremendous gratitude to PHP in that regard.

38:43 But I never thought of PHP as a calling.

38:47 I never thought, I'm a professional programmer who writes PHP,

38:51 that's who I am and that's what I do.

38:53 I thought of PHP as a tool I needed to smack

38:57 the computer with until it produced web applications I wanted.

39:01 It was very much a means to an end.

39:04 I didn't fall in love with PHP.

39:07 I'm very grateful that it taught me the basics of programming.

39:10 And I'm very grateful that it set the bar for the economics.

39:14 But it really wasn't until Ruby that I

39:18 started thinking of myself as a programmer.

39:20 And the way that came about was that the first time I ever

39:23 got hired as a professional programmer

39:26 to write code was actually by Jason Fried, my business partner still.

39:31 All the way back in 2001, I had been working on these gaming websites

39:36 in PHP for essentially 18 months at that point.

39:39 No one had been paying me to do code in that regard.

39:42 And I connect with Jason Fried over an email sent from Copenhagen,

39:49 Denmark to Chicago, Illinois to a person who didn't know who I was.

39:52 I was just offering solicited advice.

39:55 Jason had asked a question on the internet and I

39:58 had sent him the answer and he was asking in PHP.

40:00 And I'd sent him the answer to that question.

40:02 And we started talking and then we started working.

40:04 Which by the way, is a miracle of what the internet can allow.

40:07 How can a kid in Copenhagen who's never met this guy

40:11 in Chicago connect just over email and start working together?

40:15 And by the way, we're still working together now 24 years later.

40:20 That's incredible.

40:21 But we started working together and we

40:23 started working together on some client projects.

40:25 Jason would do the design, 37signals would do design,

40:28 I would bring the programming PHP.

40:30 And after we'd work on I think two or three client projects together in PHP,

40:36 we kept hitting the same problem.

40:38 That whenever you work with a client, you start that project off an email.

40:43 Oh yeah, let's work together.

40:44 Here's what we're building.

40:46 And you start trading more and more emails.

40:48 And before a few weeks have passed, you gotta add someone to the project.

40:53 They don't have the emails.

40:54 They don't have the context.

40:55 You send it, where's the latest file?

40:57 Oh, I've uploaded on the FTP.

40:59 It's like finalfinal_v06_2.0, right?

41:03 That's the one to get.

41:04 It's just a mess.

41:05 A beautiful mess in some ways,

41:06 a mess that still runs the vast majority of projects to this day.

41:10 Email is the lowest common denominator.

41:12 That's wonderful.

41:13 But we had dropped the ball a couple of times

41:16 in serious ways with customers and we thought we can do better.

41:20 We know how to make web applications.

41:22 Can't we just make a system that's better than email for managing projects?

41:27 It can't be that hard.

41:28 We've been doing blogs.

41:30 We've been doing to-do lists.

41:31 Let's put some of these together and just make a system where

41:34 everything that anyone involved in the project needs is on one page.

41:39 And it has to be simple enough that I'm not

41:41 gonna run a seminar teaching you how to use the system.

41:45 I'm just gonna give you the login code.

41:46 You're gonna jump into it.

41:47 So that's Basecamp.

41:49 And when we started working on Basecamp,

41:52 I for the first time in the experience I

41:56 had with Jason had the freedom of technology choice.

42:00 There was no client telling me, "Yeah, PHP, that sounds good.

42:04 We know PHP.

42:05 Can you build it in PHP?" I had free reigns.

42:08 And at that time I'd been reading IEEE magazine

42:13 and a couple of other magazines back from the early

42:16 2000s where Dave Thomas and Martin Fowler had been

42:22 writing about programming patterns and how to write better code.

42:27 And these two guys, in particular, were both using Ruby to explain

42:32 their concepts because Ruby looked like pseudo code.

42:37 Whether you were programming in C or Java or PHP,

42:40 all three constituencies could understand Ruby

42:42 because it basically just read link English.

42:45 So these guys were using Ruby to describe the concepts.

42:48 And first of all, I would read

42:50 these articles for just the concepts they were explaining.

42:52 And I'd be like, what is this programming language?

42:56 I mean, I like the concept you're explaining,

42:57 but I also wanna see the programming language.

43:00 Why haven't I heard of this?

43:02 So I started looking into Ruby and I realized at that time,

43:07 Ruby might not be known by anyone,

43:09 but it's actually been around for a long time.

43:11 Matz, the Japanese creator of Ruby had started working on Ruby back in '93.

43:17 Before the internet was even a thing.

43:19 And here I am in 2003, 10 years later,

43:25 picking up what seems like this hidden gem

43:28 that's just laying in obscurity in plain sight.

43:32 But Dave Thomas and Martin Fowler,

43:34 I think successfully put me and a handful of other people on the trail

43:41 of a programming language that hadn't been used much in the west, but could be.

43:47 So I picked up Ruby and I thought, this is very different.

43:52 First of all, where are all the semicolons?

43:55 I'd been programming in PHP, in ASP,

43:59 I'd even done some Pascal, I'd looked at some C.

44:01 There were semicolons everywhere.

44:04 And that was the first thing that struck me is where are the damn semicolons?

44:08 And I started thinking, actually, why do we have semicolons in programming?

44:13 They're to tell the interpreter that there's a new line of instructions,

44:18 but I don't need 'em as a human.

44:20 How?

44:21 Oh, someone is looking out for the human here, not for the machine.

44:26 So that really got me interested.

44:28 And then I thought to myself, do you know what?

44:31 I know PHP quite well.

44:32 I'm not an amazing programmer.

44:34 I haven't been working in programming for all that long,

44:38 but maybe I can figure it out.

44:41 I'm gonna give myself two weeks.

44:43 I'm gonna write a proof of concept where I talk to a database,

44:47 I pull some records, I format them a bit, and I display 'em on an HTML page.

44:51 Can I figure that out in a couple weeks?

44:54 It took about one weekend and I was completely mesmerized.

44:59 I was completely mind blown because Ruby was made for my brain

45:07 like a perfect tailored glove by someone I'd never met.

45:12 Like, how is this even possible?

45:14 We should say maybe like paint the picture

45:17 of the certain qualities that Ruby has, maybe even compared to PHP.

45:20 We should also say that there's a ridiculous thing that I'm

45:24 used to that I forget about that there's dollar signs everywhere.

45:27 PHP.

45:28 Yes, yes, there's line noise.

45:30 That's what I like to call.

45:32 There's line noise.

45:33 Line noise, that's such a beautiful phrase.

45:35 Yeah, so there's all these things that look like programs.

45:39 And with Ruby, I mean there's some similarities in Python there.

45:43 It just looks kind of like natural language.

45:45 You can read it normally.

45:47 Here's a wild loop that does a five iterations.

45:50 You can literally type the number five, dot,

45:55 now I'm calling a method under number five.

45:57 By the way, that's one of the beautiful aspects

45:59 of Ruby that primitives like integers are also objects.

46:04 And you can call 5.times, start brackets.

46:10 Now you're iterating over the code in that bracket five times, that's it.

46:15 [Lex] Okay, that's nice.

46:16 That's not just nice, that's exceptional.

46:18 There's literally no other programming language that I know

46:21 of that has managed to boil away the line noise

46:24 that almost every other programming language would inject into a five

46:29 time iteration over a block of code to that extent.

46:32 Wow, that's a really nice, well, thank you for giving that example.

46:35 That's a beautiful example.

46:37 Wow, I don't think I know what programming language that does that.

46:40 That's really nice.

46:41 Ruby is full of that.

46:43 So let me dive into a couple of examples.

46:44 Because I really think it helps paint the picture.

46:47 And let me preface this by saying I actually, I like the ethos of Python.

46:51 I think the Ruby and the Python community share a lot of similarities.

46:56 They're both dynamic interpreted languages.

46:59 They're both focused on immediacy and productivity

47:04 and ease of use in a bunch of ways.

47:05 But then they're also very different in many other ways.

47:08 And one of the one ways they're very different is aesthetically.

47:11 Python, to me, I hope I don't offend people too much.

47:15 I've said this before, it's ugly.

47:18 And it's ugly in its space, because it's full of superfluous instructions

47:25 that are necessary for legacy reasons

47:28 of when Guido made Python back in '87 that are still here in 2025.

47:35 And my brain can't cope with that.

47:37 Let me give you a basic example.

47:38 When you make a class in Python,

47:41 the initialize a method, the starting method is def.

47:44 Okay, fair enough.

47:45 That's actually the same as Ruby, D-E-F, definition of a method.

47:48 Then it is underscore, not one, underscore, two, init, underscore,

47:56 underscore, parentheses start, self, comma, and then the first argument.

48:03 [Lex] Yeah, the whole self thing.

48:06 I look at that and go, I'm sorry I'm out.

48:08 I can't do it.

48:10 It's everything about it offends my sensibilities to the core.

48:14 Here you have the most important method

48:17 that all new objects or classes have to implement.

48:21 And it is one of the most aesthetically offensive

48:24 ways of typing initialized that I've ever seen anywhere.

48:27 And you guys are okay with this?

48:29 Hey, you're making me, you know,

48:30 where you're like talking about my marriage or something like

48:33 this, and I'm not realizing I've been in a toxic relationship all along.

48:37 Yet, I just get used to it.

48:39 That to me, by the way, was the magic of Ruby.

48:41 It opened my eyes to help beautiful program could be.

48:46 I didn't know I'd been working in ASP.

48:49 I'd been working in PHP.

48:50 I didn't even have the concept that aesthetics

48:53 beautiful code was something we could optimize for that.

48:56 Something we could pursue.

48:58 And even more than that, that we could pursue it above other objectives.

49:02 That Ruby is as beautiful as it is.

49:05 It's not an accident and it's not easy.

49:08 Ruby itself is implemented in C.

49:10 It's very difficult to parse Ruby code,

49:14 because Ruby is written for humans and humans are messy creatures.

49:19 They like things in just the right way.

49:21 I can't fully explain why the __init___ make me repulse, but it does.

49:30 And when I look at the Ruby alternative, it's really instructive.

49:34 So it's def, same part, D-E-F, space, initialize, parentheses.

49:41 Not even parentheses.

49:41 If you don't need to call it within the arguments,

49:43 there's not even a parentheses.

49:44 That in itself is actually also a major part.

49:46 If the human doesn't need the additional characters,

49:49 we're not just gonna put them in because

49:51 it'd be nicer to parse for the computer.

49:54 We're gonna get rid of the semicolons.

49:55 We're gonna get rid of the parentheses.

49:57 We're gonna get rid of the underscores.

50:00 We're gonna get rid of all that ugliness,

50:02 all the line noise and boil it down to its pure essentials.

50:05 And at the same time, we're not gonna abbreviate.

50:09 This is a key difference in the aesthetics between Ruby and Python as well.

50:14 init, short of type, it's only five characters.

50:17 Initialize is a lot longer,

50:19 but it looks a lot better and you don't type it very often.

50:23 So you should look at something pretty.

50:24 If you don't have to do it all the time, it's okay that it's long.

50:29 Those kinds of aesthetic evaluations are rife all over the Ruby language.

50:34 But let me give you an even better example.

50:36 The if conditional.

50:38 That's the bedrock of all programming languages.

50:42 They have the if conditional.

50:44 If you take most programming languages, they all have if.

50:47 That's basically the same in almost every language.

50:49 Space, start parentheses, we all do that.

50:52 And then you have perhaps,

50:55 let's say you're calling a object called user.isadmin, close parentheses,

51:01 close parentheses, start brackets,

51:07 and here's what we're gonna do if the user is an admin, right?

51:11 That would be a normal programming language.

51:13 Ruby doesn't do it like that.

51:15 Ruby boils almost all of it away.

51:16 We start with the if.

51:17 Okay, that's the same.

51:19 No parenthesis necessary because there's no ambiguity for the human

51:23 to distinguish that the next part is just a single statement.

51:27 So you do if, space, user, dot, admin, question mark.

51:38 No open brackets, no parentheses, no nothing.

51:42 Next, open line.

51:44 Here's your conditional.

51:45 That question mark means nothing to the computer,

51:50 but it meets something to the human.

51:53 Ruby put in the predicate method style

51:58 purely as a communication tool between humans.

52:02 It's actually more work for the interpreter to be

52:06 able to see that this question mark is there.

52:08 Why is this question mark in here?

52:10 Because it just reads so nicely.

52:12 If user.admin?

52:15 That's a very human phrase, but it gets better.

52:20 You can turn this around.

52:22 You can have your statement you want to execute before the conditional.

52:27 You can do user.upgrade.

52:30 Say you're calling an upgrade method on a user.

52:33 Space, if, space, user.admin?.

52:38 We do the thing if the thing is true.

52:41 Instead of saying, if the thing is true, do the thing.

52:44 But it gets even better.

52:45 This is why I love this example with the conditional,

52:47 because you can keep diving into it.

52:48 So let's flip it around.

52:51 user.downgrade if !not user.admin, right?

52:59 That'd be a typical way of writing it.

53:02 Ruby goes, that exclamation point is light noise.

53:06 Why do we have if and then an exclamation point, that's ugly.

53:09 We could do user.downgrade unless user.admin?

53:17 That to me is an encapsulation of the incredible beauty that Ruby affords

53:24 the programmer through ambiguity that is only

53:28 to serve the human reader and writer.

53:31 All of these statements we've just discussed, they're the same for the computer.

53:34 It'll combine down to the same C code.

53:37 That'll compile down to the same assembly code.

53:39 It makes no difference whatsoever.

53:41 In fact, it just makes it harder to write an interpreter.

53:44 But for the human who gets to choose whether

53:47 the statement comes before the conditional or the predicate method has,

53:51 it's just incredible.

53:52 It reads like poetry at some point.

53:55 It's also incredible that, you know,

53:56 one language designer's creating that, you know, Guido van Rossum also,

54:01 it's like one person gets to make these extremely difficult decision

54:05 because you have to think about how does that all get parsed?

54:10 And you have to think about the thousands,

54:13 if it's a popular language that millions of people that end up using this.

54:17 And what they feel, what that question mark for the if statement,

54:21 what does that feel like for the user?

54:24 That's what Matz thought about because he started his entire mission off

54:28 a different premise than almost every

54:30 programming language designer that I'd heard, at least articulate the vision.

54:34 That his number one goal was programmer happiness.

54:38 That his number one goal was the affordances that would allow

54:43 programmers to articulate code in ways that not just execute it correctly,

54:49 but were a joy to write and were a joy to read.

54:54 And that vision is based on a fundamentally different view of humanity.

55:01 There's no greater contrast between Matz and James Gosling,

55:05 the designer of Java.

55:06 I wanted to listen to James talk about the design of Java.

55:11 Why was it the way it was?

55:12 Why was it so rigid?

55:14 And he was very blunt about it, which, by the way, I really appreciate.

55:17 And I think Gosling has done a tremendous job with Java.

55:20 But his view of humanity is rather dark.

55:24 His view of humanity was programmers at the average are stupid creatures.

55:30 They cannot be trusted,

55:33 which sophisticated programming languages because they're gonna

55:37 shoot their foot off or their handoff.

55:40 And that would be kind of inconvenient to the regional development office

55:45 of a mid-tier insurance company writing code that has to last for 20 years.

55:52 Now it's actually a very Thomas-Sowell view of constrained capacity

55:58 in humans that I've come to appreciate much later in life.

56:01 But it's also a very depressing view of programmers that there

56:04 are just certain programmers who are too dumb to appreciate code poetry.

56:11 There are too ignorant to learn how to write it well.

56:14 We need to give them a sandbox where they just won't hurt themselves too much.

56:20 Matz went the completely opposite direction.

56:23 He believes in humanity.

56:25 He believes in the unlimited capacity of programmers to learn and become better.

56:31 So much so that he's willing to put the stranger at his own level.

56:38 This is the second part I truly appreciate about Ruby.

56:42 Ruby allows you to extend base classes.

56:46 You know how we just talked about 5.times is

56:49 a way to iterate over a statement five times?

56:53 That five is obviously a base class.

56:56 It's a number.

56:57 Do you know what?

56:58 You can add your own methods to that.

57:01 I did, extensively.

57:03 In Rails, we have something called active support,

57:06 which is essentially my dialect of Ruby for programming web applications.

57:11 And I'll give you one example.

57:13 I've added a method called days to the Number.

57:17 So if you do 5.days, you get five days in seconds.

57:23 Because seconds is the way we set

57:25 cache expiration times and other things like that.

57:27 So you can say cache expires in 5.days.

57:32 And you're gonna get whatever's five times- That's nice.

57:37 24 times 60 times 60 is, or whatever the math is, right?

57:41 Very humanly readable.

57:43 In a novel programming language,

57:44 you would type out the seconds and then you would

57:47 have a little comment above it saying this represent five days.

57:51 In Ruby, you get to write five days.

57:53 But even better than that, Matz didn't come up with it.

57:56 Matz didn't need the five days.

57:58 I needed that because I needed to expire caches.

58:01 I was allowed by Matz to extend his story with my own chapters unequal footing,

58:10 such that a reader Ruby could not tell the difference

58:13 between the code Matz wrote and the code that I wrote.

58:16 He trusted me as a complete stranger from Denmark

58:20 who had never met to mess with his beautiful story.

58:24 That level of trust is essentially unheard of.

58:27 I know there are other program languages

58:29 that allow things with macros and so forth.

58:31 But none do it in a way like Ruby does it.

58:34 None does it with an articulated vision of humanity,

58:38 a trust in humanity like Matz does,

58:42 that is the opposite end of the spectrum of Java.

58:46 Yeah, I mean for my aesthetic sensibilities,

58:49 just the way you describe 5.days, that's really pleasant to me.

58:53 Like I could see myself sitting alone, sleep deprived, and just writing that.

59:01 It's just an easy thing.

59:02 You can write it in a long way with a comment.

59:04 You can write multiple lines.

59:05 You could do, and now with AI, I'm sure it's gonna generate it correctly,

59:09 but there's something really pleasant about the simplicity of that.

59:11 I'm not sure what that is.

59:13 But you're right, there is a good feeling there.

59:16 And I'm sure we'll talk about happiness from all kinds of philosophical angles,

59:20 but you know, that is what happiness is made of.

59:26 That little- Exactly.

59:27 good feeling there.

59:29 It's the good feeling that come out of a concept compressed to its pure essence.

59:35 There's nothing you can take away from that statement that's superfluous.

59:39 But see, I also wanna push back a little

59:41 bit 'cause I also a program named Perl a bunch.

59:46 Just be cool.

59:49 So like it's not all about compression.

59:51 No, you can compress it too far.

59:53 Perl golf is a thing where you can

59:56 turn programs into something that's unreadable for humans.

59:59 Now the great thing about Perl was that it came out before Ruby.

1:00:03 Matz was a great student of Wall, was a great student of Perl,

1:00:09 was a great student of Python and Smalltalk and Lisp.

1:00:12 He took inspiration from all of these prior attempts at creating good

1:00:17 programming languages and really edited down the very best bits into this.

1:00:21 So he was able to learn from his lessons.

1:00:24 But what I found incredible about Ruby is that here we are 2025,

1:00:29 Ruby has been worked on for over 30 years,

1:00:33 and essentially the first draft is 90% of what we're still using.

1:00:38 There was almost a sense of divine inspiration

1:00:41 possible in wherever Matz was writing that initial version

1:00:45 of Ruby that transcended time to such a degree

1:00:49 that no one has still even begun to reach it.

1:00:51 This is the other thing I always find fascinating.

1:00:54 I generally believe in the efficient market theory that if

1:00:58 someone comes up with a better mousetrap or better idea,

1:01:00 others will eventually copy them to such an extent

1:01:03 that perhaps the original mousetrap is no longer even remembered.

1:01:07 No one has been able to copy that essence of Ruby.

1:01:10 They borrowed elements, that's totally fine,

1:01:13 but Ruby still stands taller than everyone else on these metrics,

1:01:17 on this trust in humanity and programmers.

1:01:21 And we should also say like, you know,

1:01:23 maybe the perfect programming language that metric and then

1:01:27 there's the successful language and those are often different.

1:01:30 There's something wonderful about the Brendan Eich story of creating JavaScript.

1:01:36 There's something truly beautiful about the way JavaScript took over the world.

1:01:43 I've recently got to visit the Amazon jungle and just one of my favorite

1:01:46 things to do is just to watch the ants take over anything, everything.

1:01:50 And it's just like it's a nice distributor system.

1:01:52 It's a messy thing that doesn't seem to be ordered,

1:01:55 but it just works and the machinery of it.

1:01:58 Worse is Better.

1:01:58 Worse is Better.

1:01:59 I mean that's actually the name of a pattern in software

1:02:03 development and other ways of how do is the pattern of Linux.

1:02:07 Linux was quantifiably worse than I think it was Minx at the time.

1:02:11 Other ways of it that were more cathedral less bizarre and it's still

1:02:18 one that there's something to it

1:02:20 that the imperfections can help something go forward.

1:02:23 It's actually a trick I've studied to the degree that I

1:02:28 now incorporated in almost all open source that I do.

1:02:30 I make sure that when I release the first version

1:02:32 of any new thing I work on, it's a little broken.

1:02:35 It's a little busted in ways that invite people to come in and help me.

1:02:40 Because there's no easier way to get the collaboration of other programmers

1:02:46 than to put something out that they know how to fix and improve.

1:02:49 Yeah, that's awesome.

1:02:50 But Ruby is somehow, or was at least a little bit different in that regard.

1:02:54 Not in all regards.

1:02:55 Matz got the ethos of the language, the design of language just right.

1:03:00 But the first versions of Ruby were terribly slow.

1:03:03 It's taken, I mean hundreds of man years to get Ruby

1:03:09 to be both this beautiful yet also highly efficient and really fast.

1:03:15 And we should say that the thing that made you

1:03:17 fall in love with this particular programming language is metaprogramming?

1:03:21 Yes, so that takes all of these elements we've

1:03:24 just talked about and turned them up to 11.

1:03:27 I'll explain metaprogramming real quick.

1:03:28 Yeah, please.

1:03:29 Metaprogramming is essentially a version of the 5.days.

1:03:34 You get to add keywords to the language.

1:03:37 Active record is the part of Rails that communicates with the database.

1:03:43 This is a system where every table in the database is represented by a class.

1:03:50 So if we take the user example, again,

1:03:51 you do class user descends from active record base,

1:03:57 and then the first line you can write is this.

1:04:00 I want my users to have many posts or have many comments.

1:04:04 Let's do that.

1:04:05 We're making some system where users can make comments.

1:04:07 The very next line is has_many, space, colon comments.

1:04:15 Now you've set up a dependency between

1:04:17 users and comments that will give your whole

1:04:21 host of access and factory methods for users to be able to own comments,

1:04:25 to create comments, to update comments in that line

1:04:30 alone has many looks like a keyword.

1:04:33 It looks like it's part of the Ruby language.

1:04:35 That's metaprogramming.

1:04:36 When Rails is able to add these elements to how you define a class

1:04:42 and then that runs code that adds a bunch of methods to the use of class,

1:04:47 that's metaprogramming.

1:04:47 And when metaprogramming is used in this way,

1:04:51 we call it domain specific languages.

1:04:53 You take a generic language like Ruby and you tailor it

1:04:57 to a certain domain like describing relationships

1:05:00 in a database at a object level.

1:05:05 And this is one of those early examples where you can do,

1:05:09 user has many comments.

1:05:11 Belongs _two_ :account.

1:05:15 Now you've set up a one-to-one

1:05:18 relationship before we had a one-to-many relationship.

1:05:21 Rails is rife with all these kinds of domain specific

1:05:25 languages where at sometimes it doesn't even look like Ruby.

1:05:29 You can't identify Ruby keywords.

1:05:31 You can just identify what looks like keywords in its own programming language.

1:05:37 Now again, I know that Lisp and others also do this stuff.

1:05:41 They just do it with the maximum amount of line

1:05:44 noise that can ever be crammed into a programming language.

1:05:48 And Ruby does it at a level where you cannot

1:05:51 tell my metaprogramming from Matz's keywords and with zero line noise.

1:05:56 Yeah, I should say that my first love was Lisp.

1:05:58 So there's a slow tear that you can't see.

1:06:02 I've actually never written any real Lisp myself.

1:06:04 Well, how can you judge it so harshly then?

1:06:07 'Cause I have two eyes and I can look at code,

1:06:09 and my aesthetic sensibilities forbid me to even go much further.

1:06:13 Which is a limitation, I know.

1:06:15 I should actually dive into Lisp because I've found that I've

1:06:18 learned a lot just diving into, maybe I'm insulting Lisp again here,

1:06:22 but the past of programming languages with Smalltalk, for example.

1:06:26 I think Smalltalk is a incredible experiment that also

1:06:32 worked but isn't suitable for today's programming environments.

1:06:36 I love that we're talking about Ruby so much and what

1:06:38 beautiful code is and what a beautiful programming language is.

1:06:41 So one of the things that is I think implied maybe you made

1:06:46 explicit in your descriptions there is

1:06:49 that Ruby is dynamic typing versus strict typing.

1:06:52 And you have been not just saying that it's a nice thing,

1:06:57 but that you will defend dynamic typing to the death.

1:07:00 Like that freedom is a powerful freedom to preserve.

1:07:05 It's the essence of what makes Ruby Ruby.

1:07:07 This is why I don't fully understand when

1:07:09 people call for Ruby to add static typing.

1:07:12 'Cause to me, it's the bedrock of what this is.

1:07:15 Why would you wanna turn one

1:07:17 of the most beautiful languages into something far uglier?

1:07:21 This is one of my primary objections to static typing.

1:07:25 It's not just that it limits you in certain ways.

1:07:28 It makes metaprogramming harder.

1:07:30 I write a bunch of metaprogramming.

1:07:32 I've seen what it takes to do metaprogramming TypeScript.

1:07:35 That was actually one of the things that just really sent me on a tear

1:07:38 of getting meta or getting TypeScript out

1:07:40 of some of the projects that I'm involved with.

1:07:42 We pulled TypeScript out of turbo.

1:07:45 One of the front-end frameworks that we have,

1:07:48 because I tried to write to metaprogramming

1:07:51 in TypeScript and I was just infuriated.

1:07:52 I don't want that experience,

1:07:55 but I also don't want it from an aesthetic point of view.

1:07:57 I hate repetition.

1:07:59 We've just talked about how much I love that Ruby

1:08:02 boils all of these expressions down to its essence.

1:08:06 You can't remove one dot.

1:08:08 You can't remove one character without losing something.

1:08:11 This moment you go for static typing that you declare,

1:08:14 at least I know there are ways to do implied typing and so forth.

1:08:17 But let's just take the stereotypical case of a example, for example.

1:08:22 Capital U, user, I'm declaring the type of the variable; lowercase user,

1:08:29 I'm now naming my variable; equals uppercase user or new uppercase user.

1:08:35 I've repeated user three times.

1:08:37 I don't have time for this.

1:08:40 I don't have sensibilities for this.

1:08:42 I don't want my Ruby polluted with this.

1:08:45 Now I understand all the arguments for why people like static

1:08:49 typing when the primary arguments is that it makes tooling easier.

1:08:53 It makes it easier to do auto complete in editors, for example.

1:08:57 It makes it easier to find certain kinds

1:08:59 of bugs because maybe you're calling methods that don't

1:09:03 exist on an object and the editor can

1:09:06 actually catch that bug before you even run it.

1:09:09 I don't care.

1:09:11 First of all, I don't write code with tools.

1:09:15 I write them with text editors.

1:09:17 I chisel them out of the screen with my bare hands.

1:09:22 I don't auto complete.

1:09:23 And this is why I love Ruby so much, and this is why I continue to be

1:09:27 in love with the text editor rather than the IDE.

1:09:31 I don't want an IDE.

1:09:33 I want my fingers to have to individually type out every element of it

1:09:38 because it will force me to stay in the world where Ruby is beautiful.

1:09:43 Because as soon as it gets easy to type a lot of boilerplate, well guess what?

1:09:47 You can have a lot of boilerplate.

1:09:49 Every single language basically that has great tooling support has

1:09:54 a much higher tolerance for boilerplate because the thinking is,

1:09:58 well, you're not typing it anyway, you're just auto completing it.

1:10:01 I don't want that at all.

1:10:02 I want something where the fabric I'm working in, it's just a text file.

1:10:08 There's nothing else to it.

1:10:10 So these things play together.

1:10:12 There's the aesthetic part,

1:10:13 there's the tooling part, there's the metaprogramming part.

1:10:16 There's the fact that Ruby's ethos of duck typing,

1:10:21 I dunno if you've heard that term before.

1:10:23 It's essentially not about, can I call this method,

1:10:27 if a object is of a certain class.

1:10:30 It is, can I call this method if the method responds?

1:10:34 It's very out of Smalltalk in that regard.

1:10:37 You don't actually check of whether that class has the method,

1:10:41 which allows you to dynamically add methods at runtime and do all sorts

1:10:46 of really interesting things that underpin all

1:10:48 the beautiful metaprogramming that we do in Ruby.

1:10:51 I don't wanna lose any of that.

1:10:52 And I don't care for the benefits.

1:10:55 One of the benefits I see touted over and over

1:10:57 again is that it's much easier to write correct software.

1:11:01 You can have fewer bugs.

1:11:02 You're gonna have less null pointer exceptions.

1:11:06 You're gonna have less all of this stuff.

1:11:07 Yeah, I don't have any of that.

1:11:09 It's just not something that occurs in my standard mode of operation.

1:11:13 I'm not saying I don't have bugs, of course I do,

1:11:15 but I catch those bugs with unit testing, with integration testing.

1:11:19 Those are the kinds of precautions that will catch logical bugs.

1:11:25 Things that compile but are wrong along with the uncompilable stuff.

1:11:30 So I've never been drawn into this world,

1:11:33 and part of it is because I work on a certain class of systems.

1:11:36 I fully accept that.

1:11:37 If you're writing systems that have five,

1:11:40 10, 50 million lines of code with hundreds,

1:11:44 thousands, or tens of thousands of programmers,

1:11:47 I fully accept that you need different methods.

1:11:49 What I object to is the idea that what's right

1:11:52 for a code base of 10 million lines of code

1:11:55 with a hundred thousand programmers working on it is also the same

1:11:58 thing I should be using in my bedroom to create Basecamp,

1:12:01 because I'm just a single individual.

1:12:03 That's complete nonsense.

1:12:05 In the real world, we would know that that makes no sense at all.

1:12:08 That you don't, I don't know, use your Pagani to go pick up groceries at Costco.

1:12:14 It's a bad vehicle for that.

1:12:16 It just doesn't have the space.

1:12:17 You don't wanna muddy the beautiful seats.

1:12:19 You don't wanna do any of those things.

1:12:21 We know that certain things that are very

1:12:24 good in certain domains don't apply to all.

1:12:26 In programming languages, it seems like we forget that.

1:12:29 Now, to be fair, I also had a little

1:12:31 bit perhaps of a reputation of forgetting that.

1:12:33 When I first learned Ruby, I was so head over heels in love

1:12:36 with this programming language that I almost found

1:12:39 it unconceivable that anyone would choose any other

1:12:42 programming language at all to write web applications.

1:12:45 And I kind of engaged the evangelism of Ruby

1:12:47 on Rails in that spirit as a crusade, as I just need to teach you the gospel.

1:12:54 I just need to show you this conditional code that we just

1:12:57 talked about and you will convert at the point of a sharp argument.

1:13:01 Now, I learned that that's not the way.

1:13:03 And part of the reason it's not the way is the programmers think differently.

1:13:06 Our brains are configured differently.

1:13:10 My brain is configured perfectly for Ruby.

1:13:13 Perfectly for a dynamically duck typed language that I

1:13:19 can chisel code out of a text editor with.

1:13:21 And other people need the security of an IDE.

1:13:27 They want the security of classes that won't

1:13:31 compile unless you call the methods on it.

1:13:33 I have come to accept that, but most programmers don't.

1:13:36 They're still stuck in, essentially, I like static typing.

1:13:40 Therefore, static typing is the only way to create reliable correct systems,

1:13:45 which is just such a mind blowing,

1:13:48 to be blunt, idiotic thing to say in the face of evidence,

1:13:52 mountains of evidence to the contrary.

1:13:55 This is one of the reasons I'm so in love

1:13:58 with Shopify as the flagship application for Ruby on Rails.

1:14:03 Shopify exists at a scale that most programmers will never touch.

1:14:10 On Black Friday, I think Shopify did 1 million requests per second.

1:14:15 That's not 1 million requests of images.

1:14:18 That's of dynamic requests that are funneling through the pipeline of commerce.

1:14:22 I mean, Shopify runs something like 30%

1:14:25 of all e-commerce stores on the damn internet.

1:14:29 A huge portion of all commerce in total runs

1:14:33 through Shopify and that runs on Ruby on Rails.

1:14:35 So Ruby on Rails is able to scale up to that level

1:14:42 without using static typing in all of what it does.

1:14:45 Now I know they've done certain experiments in certain ways because they

1:14:48 are hitting some of the limits that you will hit with dynamic typing.

1:14:51 And some of those limits you hit with dynamic typing are actually,

1:14:55 by the way, just limits you hit when you write 5 million lines of code.

1:14:58 I think the Shopify monolith is about 5 million lines of code.

1:15:01 At that scale, everything breaks because you're at the frontier

1:15:06 of what humans are capable of doing with programming languages.

1:15:09 The difference in part is that Ruby is

1:15:12 such a succinct language that those 5 million,

1:15:14 if they had been written in, let's just say Go or Java would've been 50 or 25.

1:15:20 Now that might have alleviated some of the problems that you

1:15:24 have when you work on huge systems with many programmers.

1:15:27 But it certainly would also have compounded them

1:15:29 trying to understand 25 million lines of code.

1:15:33 So the thing does scale.

1:15:34 That's a persistent myth that it doesn't scale Shopify and others.

1:15:39 But Shopify thinks a great example.

1:15:41 By the way, I love Shopify and I love Tobi, amazing.

1:15:45 You gotta have Tobi on.

1:15:46 Yeah, for sure.

1:15:47 Just talking to him this morning.

1:15:47 For sure, he's a brilliant.

1:15:49 I got to hang out with him in the desert somewhere, I forget in Utah.

1:15:52 He's just a brilliant human.

1:15:55 And Shopify, shopify.com/lex has been supporting

1:15:59 this podcast for the longest time.

1:16:00 I don't think actually Tobi knows that they sponsor this podcast.

1:16:04 I mean this is a big company, right?

1:16:05 It's a huge company.

1:16:06 I think just under 10,000 employees, market cap of 120 billion,

1:16:13 GMV of a quarter of a trillion every quarter.

1:16:16 And he's involved with the details though.

1:16:18 He is, very much so.

1:16:19 Funny story about Tobi.

1:16:21 Tobi was on the Rails core team back in the mid 2000s.

1:16:26 Tobi himself wrote Active Merchant,

1:16:29 which is one of the frameworks for creating shops.

1:16:32 He wrote the Liquid templating language that Shopify still uses to this day.

1:16:37 He has a huge list of contributions to the Rails ecosystem,

1:16:41 and he's the CEO of the company.

1:16:43 Yeah, I think it's just...

1:16:45 It's very inspiring to me because it's such

1:16:47 at the opposite end of what I like to do.

1:16:50 I like to chisel code with my own hands most of the day.

1:16:53 He runs a company of almost 10,000 people

1:16:57 that is literally like world commerce depends on it.

1:17:01 A level of criticality I can't even begin to understand.

1:17:05 And yet we can see eye to eye on so

1:17:07 many of these fundamental questions in computer science and program development.

1:17:12 That is a dynamic range to be able to encompass Rails being a great tool

1:17:20 for the one developer who's just starting out

1:17:23 with an idea who don't even fully know everything.

1:17:25 Who is right at the level where PHP would've

1:17:28 been a good fit in those late '90s because yeah,

1:17:30 I could probably upload something to an FTP server, and so on.

1:17:33 Rails does have more complexity than

1:17:35 that, but it also has so much longer runway.

1:17:37 The runway goes all the way to goddamn Shopify.

1:17:40 That is about the most convincing argument I can make for sort

1:17:43 of dynamic range that we can do a lot of it.

1:17:46 And even having said that, Shopify is the outlier of course.

1:17:51 I don't think about Shopify as the primary target when I write Rails.

1:17:56 I think of the single developer.

1:17:58 Actually, I do think about Shopify.

1:17:59 But I don't think about Shopify now.

1:18:01 I think of Shopify when Tobi was writing Snow Devil,

1:18:04 which was the first e-commerce store to sell snowboards that he created,

1:18:08 that was the pre-Shopify Shopify.

1:18:10 He created all by himself.

1:18:12 And that was possible because Ruby on Rails isn't just about beautiful code.

1:18:17 It's just as much about productivity.

1:18:19 It's just as much about the impact

1:18:21 that an individual programmer is able to have.

1:18:23 That they can build system where they can keep the whole

1:18:26 thing in their head and be able to move it forward such

1:18:28 that you can go from one developer sitting and working on something

1:18:33 and that something is Shopify and it turns into what it is today.

1:18:37 When we talk about programming languages and we compare 'em,

1:18:40 we often compare 'em at a very late stage.

1:18:43 Like what is the better programming language for let's

1:18:45 say Twitter in 2009 when it's already a huge success.

1:18:49 Twitter was started on Ruby on Rails.

1:18:51 They then hit some scaling problems.

1:18:53 It was a big debacle at the time.

1:18:55 They end up then I think writing it in some other language,

1:19:01 which by the way I think is the best advertisement ever for Ruby on Rails,

1:19:03 because nothing fucking happened for 10 years after they switched over, right?

1:19:08 Essentially zero innovation.

1:19:10 Some of that was because they were doing a long conversion,

1:19:14 and all of the early success in part came because

1:19:17 they had the agility quickly change and adopt and so forth.

1:19:20 That's what startups needs.

1:19:22 That's what Shopify needed.

1:19:23 That's what Twitter needed.

1:19:24 That's what everyone needs.

1:19:26 And that's the number one priority for Ruby on Rails.

1:19:29 To make sure that we don't lose that.

1:19:30 Because what happens so often when development tools and program language

1:19:35 driven by huge companies is that they mirror their org chart.

1:19:39 React and everything else needed to use that is

1:19:44 in some ways a reflection of how Meta builds Facebook.

1:19:48 'Cause of course it is.

1:19:49 Because of course it's an destruction of that.

1:19:50 I'm not saying React isn't a great tool and that can't be used by smaller teams.

1:19:54 Of course it can.

1:19:55 But it's born in a very different context than something like Ruby on Rails.

1:20:00 Let me say this a small aside because I

1:20:02 think we might return to Shopify and celebrate it often.

1:20:05 Just a sort of personal note.

1:20:08 This particular podcast has way more sponsors and sponsors

1:20:12 that want to be sponsors than I could possibly ever have.

1:20:16 And it's really, really important for me to not give a shit,

1:20:21 and to be able to celebrate people.

1:20:24 Like I celebrate people.

1:20:25 I celebrate companies.

1:20:27 And I don't care that they're sponsoring.

1:20:30 I really don't care.

1:20:32 I just wanna make that very explicit 'cause

1:20:34 we're gonna continue saying positive things about Shopify.

1:20:36 I don't care.

1:20:37 Stop sponsoring.

1:20:39 It doesn't really matter to me.

1:20:40 But yeah, I just wanna make that explicit.

1:20:42 So, but to linger on the scaling thing with the Twitter and the Shopify,

1:20:46 can you just explain to me what Shopify is doing with the YJIT?

1:20:52 What did they have to try to do to scale this thing?

1:20:57 Because that's kind of an incredible story, right?

1:20:59 Yeah, so one of the great contributions

1:21:01 that Shopify has made to the entire Ruby ecosystem,

1:21:04 not just Rails but in particular Rails is YJIT.

1:21:07 So YJIT is their compiler for Ruby.

1:21:11 That just makes everything a lot more efficient

1:21:13 and at Shopify scale eking out even a five,

1:21:17 10% improvement in Ruby's overhead and execution time is a huge deal.

1:21:23 Now, Shopify didn't need YJIT.

1:21:27 Shopify was already running on the initial version of Ruby that was,

1:21:31 I think 10 times slower than what we have today.

1:21:34 If you look back upon the Ruby 186,

1:21:38 that Tobi probably started on just as I started on, and that was

1:21:42 enough to propel Shopify to the scale that it has today.

1:21:46 A lot of the scaling conversation in is

1:21:49 lost in a failure to distinguish two things.

1:21:54 Scale is kind of one package we talk about

1:21:57 when there are really multiple packages inside of it.

1:21:59 One is runtime performance, latency.

1:22:03 How fast can you execute a single request?

1:22:05 Can it happen fast enough that the user will not notice?

1:22:09 If your Rails request takes a second and a half to execute,

1:22:12 the user's gonna notice.

1:22:14 Your app is gonna feel slow and sluggish.

1:22:16 You have to get that response time down below,

1:22:20 let's say at least 300 milliseconds.

1:22:21 I like to target a hundred milliseconds as my latency.

1:22:25 That's kind of performance.

1:22:27 How much performance of that kind of latency

1:22:30 can you squeeze out of a single CPU core,

1:22:32 that tells you something about what the price of a single request will be.

1:22:36 But then whether you can deal with 1 million

1:22:40 requests a second like Shopify is doing right now.

1:22:43 If you have one box that can do a thousand requests a second,

1:22:47 you just need X boxes to get up to a million.

1:22:50 And what you'll actually find is that when it comes to programming languages,

1:22:53 they're all the same in this way.

1:22:55 They all scale largely beautifully horizontally.

1:22:59 You just add more boxes.

1:23:01 The hard parts of scaling a Shopify is typically the program language.

1:23:07 It's the database.

1:23:08 And that's actually one of the challenges that Shopify has now is

1:23:13 how do you deal with MySQL at the scale that they're operating at?

1:23:18 When do you need to move to other databases to get worldwide performance?

1:23:23 All of these things.

1:23:25 The questions about scaling Ruby are economic questions.

1:23:28 If we are spending so and so much on application servers,

1:23:32 if we can get just 5% more performance out of Ruby,

1:23:35 well we could save 5% of those servers

1:23:37 and that could filter down into the budget.

1:23:39 Now that analysis concludes into basically one thing.

1:23:43 Ruby is a luxury language.

1:23:47 It's a luxury, the highest luxury in my opinion.

1:23:51 It is the Coco Chanel of programming languages.

1:23:55 Something that not everyone can afford.

1:23:57 And I mean this in the best possible way.

1:23:59 There are some applications on the internet

1:24:01 where each request has so little value,

1:24:05 you can't afford to use a luxurious language like Ruby to program in it.

1:24:09 You simply have to slum it with a C

1:24:12 or a Go or some other low level language or Rust,

1:24:16 talk about line noise there for...

1:24:17 It's like the thrift store of languages.

1:24:20 Exactly, where you need kind of just...

1:24:22 you need a very low level to do it.

1:24:24 You can't afford to use a luxury language to use to build it with.

1:24:29 That's not true of Shopify.

1:24:30 It wasn't true of Basecamp even back in 2004.

1:24:33 It's not been true of 99% of all web applications ever

1:24:38 created because the main cost component of 99% of web applications,

1:24:42 it's not CPU course.

1:24:44 It's wet course.

1:24:46 It's human course.

1:24:47 It's human capacity to understand and involve systems.

1:24:52 It's their personal productivity.

1:24:54 I did a calculation once when someone had for the 400th time said that, "Oh,

1:24:58 if you switch from Ruby to some faster language,

1:25:02 you could save a bunch of money." And I calculated it out that at the time,

1:25:05 and I think the last time I did this calculation was almost a decade ago.

1:25:09 We were spending about 15% of our operating budget on Ruby application service.

1:25:15 So for me to improve my cost profile of the business by seven percentage points,

1:25:24 I'd have to pick something twice as fast.

1:25:26 That's quite hard.

1:25:27 Versus if Ruby and Ruby on Rails

1:25:30 was even 10% more productive than something else,

1:25:33 I would move the needle far more because making

1:25:36 individual programmers more productive actually matters a lot more.

1:25:39 This is why people are so excited about AI.

1:25:40 This is why they're freaking out over the fact

1:25:44 that a single programmer in Silicon Valley who

1:25:46 makes $300,000 a year can now do the work of three or five, at least in theory.

1:25:52 I haven't actually seen that fully in practice,

1:25:54 but let's just assume the theory is correct if not now, then in six months.

1:25:58 That's a huge deal.

1:26:00 That matters so much more than whether you can

1:26:03 squeeze a few more cycles out of the CPU,

1:26:06 when it comes to these kinds of business applications.

1:26:08 If you're making unreal engine rendering stuff like Tim Sweeney you had on.

1:26:14 Yeah, he needs to really sweat all those details.

1:26:16 The Nanite engine can't run on Ruby.

1:26:19 It's never going to.

1:26:20 It would not meant for that, fine.

1:26:22 These kinds of business applications absolutely can and everything

1:26:26 that people are excited about AI for right now,

1:26:30 that extra capacity to just do more.

1:26:33 That was why we were excited about Ruby back in the early 2000s.

1:26:36 That was be because I saw that if we could

1:26:39 even squeeze out a 10% improvement of the human programmer,

1:26:44 we'd be able to do so much more for so much less.

1:26:48 Probably argue about this, but I really like working together with AI,

1:26:54 collaborating with AI.

1:26:55 And I would argue that the kind of code

1:26:57 you want AI to generate is human readable, human interpretable.

1:27:00 If it's generating Perl golf code, it's not a collaboration.

1:27:07 So it has to be speaking the human.

1:27:09 It's not just you're writing the prompts in English.

1:27:11 You also want to read the responses

1:27:14 in the human interpretable language like Ruby, right?

1:27:17 So that's actually is beneficial for AI too.

1:27:19 'Cause you've kind of said that for you the sculptor,

1:27:23 the sort of the elitist Coco Chanel sculptor,

1:27:27 you want to on your fancy keyboard to type

1:27:30 every single letter yourself of your own fingers.

1:27:32 But it's also that the benefit of Ruby also applies in once some

1:27:38 of that is written by AI and you're

1:27:40 actually doing with your own fingers the editing.

1:27:43 Because you can interact with it because it's human interpretable.

1:27:47 The paradigm I really love with this was something Elon actually said

1:27:49 on one of your shows when you guys were talking about Neuralink,

1:27:52 that Neuralink allows the bandwidth between you and the machine to increase.

1:27:59 That language either spoken or written is very low bandwidth.

1:28:03 If you are to calculate just how many bits we

1:28:06 can exchange as we are sitting here, it's very slow.

1:28:11 Ruby has a much higher bandwidth of communication revealed conveys so

1:28:18 much more concept per character than most other programming languages do.

1:28:24 So when you are collaborating with AI, you want really high bandwidth.

1:28:29 You want it to be able to produce programs with you,

1:28:32 whether you're letting it write the code or not,

1:28:36 that both of you can actually understand really

1:28:39 quickly and that you could compress a grand concept,

1:28:44 a grand system into far fewer parts that both of you can understand.

1:28:49 Now, I actually love collaborating with AI too.

1:28:52 I love chiseling my code and the way I use AI is in a separate window.

1:28:57 I don't let it drive my code.

1:28:58 I've tried that.

1:28:59 I've tried the cursors and the wind surfaces

1:29:01 and I don't enjoy that way of writing.

1:29:03 And one of the reasons I don't enjoy that way of writing

1:29:07 is I can literally feel competence draining out of my fingers.

1:29:12 Like that level of immediacy with the material disappears.

1:29:17 And the where I felt this the most was I did

1:29:21 this remix of Ubuntu called Omakub when I switched to Linux,

1:29:25 and it's all written in Bash.

1:29:27 I'd never written any serious amount of code in Bash before.

1:29:31 So I was using AI to collaborate,

1:29:32 to write a bunch of Bash with me, 'cause I needed all this.

1:29:35 I knew what I wanted.

1:29:37 I could express it in Ruby.

1:29:39 But I thought it was an interesting challenge to filter through

1:29:41 Bash because what I was doing was setting up a Linux machine.

1:29:45 That's basically what Bash was designed for.

1:29:47 It's a great constraint.

1:29:49 But what I found myself doing was asking

1:29:52 AI for the same way of expressing a conditional,

1:29:55 for example, in Bash over and over again.

1:29:58 That by not typing it, I wasn't learning it.

1:30:01 I was using it, I was getting the expression I wanted, but I wasn't learning it.

1:30:06 And I got a little scared.

1:30:08 I got a little scared like, is this the end of learning?

1:30:11 Am I no longer learning if I'm not typing?

1:30:13 And the way I, for me, recast that was, I don't wanna give up on the AI.

1:30:18 It's such a better experience as a programmer to look up APIs,

1:30:23 to get a second opinion on something, to do a draft.

1:30:28 But I have to do the typing myself because you learn with your fingers.

1:30:32 If you're learning how to play the guitar,

1:30:34 you can watch as many YouTube videos as you want.

1:30:36 You're not gonna learn the guitar.

1:30:38 You have to put your fingers on the strings to actually learn the motions.

1:30:44 And I think there is a parallel here to programming where

1:30:46 programming has to be learned in part by the actual typing.

1:30:51 I'm just really, this is fascinating.

1:30:53 Listen, part of my brain agrees with you.

1:30:55 100% part doesn't.

1:30:59 I think AI should be in the loop of learning.

1:31:02 Now current systems don't do that, but I

1:31:06 think it's very possible for cursor to say,

1:31:09 to basically force you to type certain things.

1:31:12 So like if you set the mode of learning,

1:31:14 I don't want to be this sort of give up on AI.

1:31:20 I think vibe coding is a skill.

1:31:24 So for an experienced programmer it's too

1:31:28 easy to dismiss vibe coding as a thing.

1:31:31 [DHH] I agree, I wouldn't dismiss it.

1:31:32 But I think you need to start building that skill and start to figure

1:31:37 out how do you prevent the competency

1:31:40 from slipping away from your fingers and brain.

1:31:43 Like how do you develop that skill in parallel to the other skill?

1:31:47 I don't know.

1:31:48 I think it's a fascinating puzzle though.

1:31:51 I know too many really strong programmers that just kind of avoid AI,

1:31:54 'cause it's currently a little too dumb.

1:31:57 Yes, it's a little too slow.

1:31:58 It's actually my main problem.

1:32:00 It's a little too dumb in some ways, but it's a little too slow in other ways.

1:32:03 When I use Claude's code, the terminal version of a Claude,

1:32:10 which is actually my preferred way of using it, I get too impatient.

1:32:14 It feels like I'm going back to a time where code

1:32:16 had to compile and I had to go do something else.

1:32:19 A boil some tea while the code is compiling.

1:32:21 Well, I've been working in Ruby for 20 years.

1:32:24 I don't have compiled weight in me anymore.

1:32:28 So there's that aspect of it.

1:32:30 But I think the more crucial aspect

1:32:31 for me is I really care about the competence.

1:32:34 And I've seen what happens to even great

1:32:36 programmers the moment they put away the keyboard.

1:32:39 Because even before AI, this would happen as soon as people would get promoted.

1:32:44 Most great programmers who work in large businesses stop writing code

1:32:48 on a daily basis because they simply have too many meetings to attend to.

1:32:51 They have too many other things to do.

1:32:53 And invariably they lose touch with programming.

1:32:57 That doesn't mean they forget everything.

1:33:00 But if you don't have your fingers in the sauce,

1:33:05 source, you are going to lose touch with it.

1:33:08 There's just no other way.

1:33:09 I don't want that because I enjoy it too much.

1:33:12 This is not just about outcomes.

1:33:14 This is what's crucial to understand programming for programmers who like

1:33:19 to code is not just about the programs they get out of it.

1:33:23 That may be the economic value.

1:33:25 It's not the only human value.

1:33:27 The human value is just much in the expression.

1:33:31 When someone who sits down on a guitar and plays "Stairway

1:33:35 to Heaven," there's a perfect recording of that that will last in eternity.

1:33:40 You can just put it on Spotify.

1:33:41 You don't actually need to do it.

1:33:42 The joy is to command the guitar yourself.

1:33:46 The joy of a programmer, of me as a programmer, is to type the code myself.

1:33:51 If I elevate myself, if I promote myself out of programming,

1:33:55 I turn myself into a project manager.

1:33:57 A project manager of a murder of AI crows as I wrote the other day.

1:34:02 I could have been become a project manager my whole career.

1:34:05 I could have become a project manager 20 years

1:34:07 ago if I didn't care to write code myself.

1:34:09 And I just wanted outcomes.

1:34:11 That's how I got started in programming.

1:34:12 I just wanted outcomes.

1:34:14 Then I fell in love with programming

1:34:16 and now I'd rather retire than giving it up.

1:34:19 Now that doesn't mean you can't have your cake and needed to.

1:34:22 I've done some vibe coding where I didn't care that I wasn't playing myself.

1:34:28 I just wanted to see something.

1:34:30 There was an idea in my head.

1:34:31 I wanted to see something.

1:34:33 That's fine.

1:34:34 I also use AI all day long.

1:34:37 In fact, I'm already at the point where if you took it away from me,

1:34:40 I'd be like, oh my god, how do we even look things up on the internet anymore?

1:34:43 Is Stack Overflow still around?

1:34:46 Or I'm still a thing?

1:34:47 Like how do I even find answers to some of these questions I have all day long.

1:34:51 I don't wanna give up AI.

1:34:52 In fact, I'd say the way I like to use AI,

1:34:55 I'm getting smarter every because of AI.

1:34:58 Because I'm using AI to have it explain things to me.

1:35:02 Even stupid questions.

1:35:04 I would be a little embarrassed to even enter into Google.

1:35:07 AI is perfectly willing to give me the ELI5 explanation of some Unix command.

1:35:14 I should have known already, but I don't, I'm sorry,

1:35:17 can you just explain it to me and now I know the thing.

1:35:19 So at the end of the day of me working with AI all day long,

1:35:23 I'm a little bit smarter.

1:35:25 Like 5%, sorry, not 5%, half a percent maybe.

1:35:29 That compounds over time.

1:35:31 But what I've also seen,

1:35:32 when I worked on the Omakub project and I tried to let AI drive for me,

1:35:36 I felt I was maybe half a percent dumber at the end of the day.

1:35:41 Okay, you said a lot of interesting things.

1:35:42 First of all, let's just start at the very fact that asking dumb questions.

1:35:46 If you go to Stack Overflow and ask a dumb question

1:35:49 or read somebody else's dumb question and the answer to it,

1:35:52 there's a lot of judgment there.

1:35:54 AI sometimes to an excessive degree has no judgment.

1:35:58 It usually says, oh, that's a great question.

1:36:00 [DHH] To a fault.

1:36:02 Yeah, oh, that's wonderful.

1:36:05 Yeah, I mean, it's so conducive to learning.

1:36:10 It's such a wonderful tool of learning and I too would miss it.

1:36:13 And it's a great basically search engine into all

1:36:17 kinds of nuances of a particular programming language,

1:36:19 especially if you don't know it that well

1:36:22 or like APIs you can load in documentation.

1:36:24 It's just so great for learning.

1:36:26 I, for me personally...

1:36:30 I mean, on the happiness scale, it makes me more excited to program.

1:36:35 I don't know what that is exactly.

1:36:37 Part of that is the, I'm really sorry.

1:36:40 Stack Overflow is an incredible website, but there is a negativity there.

1:36:44 There's a judgment there.

1:36:45 It's just exciting to be out like with a hype man next to me,

1:36:51 just like saying, yeah, that's a great idea.

1:36:53 And I'll say, no, that's wrong.

1:36:55 I'll correct the AI, and the AI will say, "You're absolutely right.

1:37:00 How did I not think about that?" You know, rewrite the code.

1:37:03 I'm like, holy shit, I'm having, that's like a buddy.

1:37:06 That's like really being positive and is

1:37:09 very smart and is challenging me to think.

1:37:12 And even if I never use the code it generates, I'm already a better programmer.

1:37:16 But actually the deeper thing is for some reason I'm having more fun.

1:37:21 That's a really, really important thing.

1:37:23 I like to think of it as a pair programmer for exactly that reason.

1:37:26 Pair programming came a vogue in like the 2000s.

1:37:30 Where you'd have two programmers in front of one

1:37:32 machine and you'd push the keyboard between you.

1:37:35 One programmer would be driving, they'd be typing in.

1:37:38 The other programmer would essentially sit and watch the code,

1:37:41 suggest improvements, look something up.

1:37:43 That was a really interesting dynamic.

1:37:46 Now unfortunately, I'm an introvert,

1:37:47 so I can do that for about five minutes before I want to jump off a bridge.

1:37:51 So it doesn't work for me as a full-time occupation.

1:37:54 But AI allows me to have all the best of that experience all the time.

1:37:59 Now I think what's really interesting, what you said about it makes it more fun.

1:38:02 I hadn't actually thought about that, but what it's made

1:38:05 more fun to me is to be a beginner again.

1:38:09 It made it more fun to learn Bash successfully for the first time.

1:38:14 Now I had to do the detour where I let

1:38:16 it write all the code for me and I realized I

1:38:18 wasn't learning nearly as much as I hoped I would

1:38:21 and that I started doing once I typed it out myself.

1:38:24 But it gave me the confidence that, you know what?

1:38:27 If I need to do some iOS programming myself,

1:38:29 I haven't done that in probably six years was the last time I dabbled in it.

1:38:33 I never really built anything for real.

1:38:35 I feel highly confident now that I could sit down with AI and I

1:38:39 could have something in the app store by the end of the week.

1:38:41 I would not have that confidence unless I

1:38:44 had a pair of programming body like AI.

1:38:46 I don't actually use it very much for Ruby code.

1:38:49 I'm occasionally impressed whenever I try it,

1:38:51 they're like, oh, it got this one thing right.

1:38:53 That is truly remarkable and it's actually pretty good.

1:38:55 And then I'll ask you two more questions and I go like, oh yeah, okay.

1:38:59 If you were my junior programmer,

1:39:01 I'd start tapping my fingers and going like, you gotta shape up.

1:39:05 Now the great thing of course is we can just wait five minutes.

1:39:08 The Anthropic CEO seems to think that 90% of all code

1:39:12 by the end of the year is gonna be written by AI.

1:39:14 I am more than a little bit skeptical about that, but I'm open-minded

1:39:17 about the prospect that programming potentially will

1:39:22 turn into a horse when done manually.

1:39:25 Something we do recreationally is no longer

1:39:28 a mode of transportation to get around LA.

1:39:30 You're not gonna saddle up and go to the grocery

1:39:33 store and pick up stuff from whole foods in your saddlebacks.

1:39:35 That's just not a thing anymore.

1:39:37 That could be the future for programming,

1:39:40 for manual programming entirely possible.

1:39:44 I also don't care.

1:39:45 Like even though we have great renditions of all the best songs,

1:39:49 as I said, there are millions of people who love to play the guitar.

1:39:52 It may no longer have as much economic value as it once did.

1:39:55 I think that I'm quite convinced is true that we perhaps have seen the peak.

1:40:01 Now I understand the paradox, when the price of something goes down,

1:40:04 actually the overall usage goes up and total spend on that activity goes up.

1:40:09 That could also happen maybe.

1:40:10 But what we're seeing right now is that a lot of the big shops,

1:40:15 a lot of the big companies are not hiring like they were five years ago.

1:40:19 They're not anticipating they're gonna need tons more programmers.

1:40:23 Controversially, Tobi actually put out a memo inside of Shopify

1:40:28 asking everyone who's considering hiring someone to ask the question,

1:40:31 could this be done by AI?

1:40:33 Now, he's further ahead on this question than I am.

1:40:35 I look at some of the coding trenches and I go like, I'd love to use AI more,

1:40:39 and I see how it's making us more productive,

1:40:41 but it's not yet at the level where I just go like, oh, we have this project.

1:40:44 Let me just give it to the AI agent and it's gonna go off and do it.

1:40:47 But let's just be honest.

1:40:49 You're like a Clint-Eastwood type character cowboy and on a horse

1:40:52 seeing cars going around and you're like, well.

1:40:56 That's part of it.

1:40:57 And I think that's, it is important to have that humility

1:40:59 that what you are good at may no longer be what society values.

1:41:05 This has happened a million times in history that you

1:41:07 could have been exceptionally good at saddle making, for example.

1:41:10 That's something that a lot of people used

1:41:12 to care about because everyone rode a horse.

1:41:14 And then suddenly riding a horse became this niche

1:41:18 hobby that there's some people care about it, but not nearly as many.

1:41:22 That's okay.

1:41:22 Now the other thing of this is I've had the good fortune to...

1:41:27 I've been a programmer for nearly 30 years.

1:41:28 That was a great run.

1:41:30 I try to look at life in this way

1:41:33 that I've already been blessed with decades of economically viable,

1:41:38 highly valuable ways of translating what I like

1:41:43 best in the working world to write Ruby code.

1:41:47 That that was so valuable that I could

1:41:49 make millions and millions of dollars doing it.

1:41:51 And if that's over tomorrow, I shouldn't look at that with regret.

1:41:55 I should look at it with gratitude.

1:41:57 But you're also a highly experienced, brilliant and opinionated human being,

1:42:04 so it's really interesting to get your opinion

1:42:07 on the future of the horse because it, you know,

1:42:11 there's a lot of young people listening to this who love programming

1:42:13 or who are excited by the possibility of building stuff with software,

1:42:19 with Ruby on Rails, that kind of language.

1:42:22 This and now the possibility.

1:42:24 But is it a career?

1:42:27 Is it a career?

1:42:29 And if indeed a single person can build more

1:42:32 and more and more with the help of AI, like how do they learn that skill?

1:42:36 Is this a good skill to learn?

1:42:38 I mean, that to me is the real mystery here,

1:42:41 because I think it's still absolutely true that you

1:42:44 have to learn how to program from scratch currently.

1:42:47 But how do you balance those two skills?

1:42:50 Because I too, as I'm thinking now,

1:42:53 there is a scary slipping away of skill that happens.

1:42:57 In a matter of like really minutes on a particular piece of code.

1:43:03 It's scary.

1:43:04 Not the way driving, you know,

1:43:08 when you have a car drive for you doesn't quite slip away that fast.

1:43:14 So that really scares me.

1:43:16 So when somebody comes up to me and asks me like, how do I learn to program,

1:43:20 I don't know what the advice is because I think it's

1:43:23 not enough to just use cursor or copilot to generate code.

1:43:29 It's absolutely not enough.

1:43:30 Not if you wanna learn, none of you want to become better at it.

1:43:33 If you just become a tap monkey, maybe you're productive in a second.

1:43:37 But then you have to realize, well, can anyone just tap?

1:43:40 If that's all we're doing is just sitting around all day long tapping?

1:43:43 Yes, yes, yes, yes, yes.

1:43:44 That's not a marketable skill.

1:43:46 Now, I always preface this both to myself and when I speak to others,

1:43:51 it is rule number note one, nobody fucking knows anything.

1:43:55 No one can predict even six months ahead.

1:43:57 Right now, we're probably at peak AI future hype because we see all the promise,

1:44:03 because so much of it is real and so many people have experienced it themselves.

1:44:07 This mind boggling thing that the silicon is thinking

1:44:13 in some way that feels eerily reminiscent of humans.

1:44:17 I'd actually say the big thing for me wasn't even ChatGPT.

1:44:21 It wasn't even Claude.

1:44:22 It was DeepSeek.

1:44:23 Running DeepSeek locally and seeing the think box where

1:44:27 it converses with itself about how to formulate the response.

1:44:33 I almost wanted to think, is this a gimmick?

1:44:35 Is it doing this as a performance for my benefit?

1:44:38 But that's not actually how it thinks.

1:44:40 If this is how it actually thinks, okay, I'm a little scared.

1:44:43 This is incredibly human how it thinks in this way.

1:44:47 But where does that go?

1:44:49 So in '95, 1 of my favorite movies,

1:44:53 one of my favorite B movies came out, "The Lawnmower Man."- Great movie.

1:44:57 Incredible movie about virtual reality.

1:44:59 Being an avatar and living in VR, like the story was a mess.

1:45:03 But the aesthetics the world had built up was incredible.

1:45:06 And I thought, we're five years away.

1:45:09 I'm gonna be living in VR now.

1:45:11 I'm just gonna be floating around.

1:45:13 I'm gonna be an avatar.

1:45:15 This is where most humans can spend most of the day that didn't happen.

1:45:19 We're 30 years later, VR is still not here.

1:45:25 It's here for gaming.

1:45:26 It's here for some specialized applications.

1:45:29 My oldest loves playing Gorilla Tag.

1:45:32 I dunno if you've tried that.

1:45:33 That's basically the hottest VR game.

1:45:35 Wonderful, it's great.

1:45:37 It's really hard to predict the future 'cause we just don't know.

1:45:41 And then when you factor in AI and you have even the smartest people go like,

1:45:47 I don't think we fully understand how this works.

1:45:49 But then on the flip side,

1:45:50 you have Moore's Law that seems to have worked for many,

1:45:54 many, many years in decreasing the size of transistor, for example.

1:45:58 So like you know, Flash didn't take over the internet, but Moore's Law worked.

1:46:05 So we don't know which one AI is.

1:46:07 Which one it is.

1:46:08 And this is what I find so fascinating too.

1:46:10 I forget who did this presentation, but someone in the web community,

1:46:13 this great presentation on the history of the airplane.

1:46:17 So you go from the Wright brothers flying in, what?

1:46:20 1903 or something like that.

1:46:22 And 40 years later you have a jet flight.

1:46:27 Just an unbelievable amount of progress in four decades.

1:46:31 Then in '56, I think it was, the halt the sign for the Boeing 747 essentially

1:46:39 precursor was designed and basically nothing has happened since.

1:46:43 Just minor tweaks and improvements on the flying experience since the '50s.

1:46:48 Somehow, if you were to predict where flying was gonna go,

1:46:52 and you were sitting in 42 and you'd seen,

1:46:55 you'd remember the Wright brothers flying in '03,

1:46:58 and you were seeing that jet engines coming, you're like,

1:47:01 we're gonna fly to to the stars in another two decades.

1:47:04 We're gonna invent super mega hypersonic flights

1:47:08 that's gonna traverse the earth in two hours.

1:47:11 And then that didn't happen.

1:47:12 It tapped out.

1:47:13 This is what's so hard about predicting the future.

1:47:16 We can be so excited in the moment because

1:47:18 we're drawing a line through early dots on a chart,

1:47:22 and it looks like those early dots are just going up into the right,

1:47:24 and sometimes it's just flatten out.

1:47:26 This is also one of those things where we have so much critical infrastructure,

1:47:30 for example, that still runs on COBOL.

1:47:32 That about five humans around the world really understand truly,

1:47:35 deeply that there's a lot, it's possible for society to lose a competence.

1:47:40 It's still needs because it's chasing the future.

1:47:44 COBOL is still with us.

1:47:46 This is one of the things I think about with with programming.

1:47:49 Ruby on Rails is at such a level now that in 50 years from now,

1:47:53 it's exceedingly likely that there's still a ton

1:47:55 of Ruby on Rail systems running around.

1:47:57 And very hard to predict what that exact world is gonna be like.

1:48:00 But yesterday's weather tells us that if there's still

1:48:03 COBOL code from the '70s operating social security today,

1:48:07 and we haven't figured out a clean way to convert that, let alone understand it,

1:48:12 we should certainly be humble about predicting the future.

1:48:16 I don't think any of the programmers who wrote that COBOL

1:48:19 code back in the '70s had any damn idea that in 2025,

1:48:23 checks were still being cut off the business

1:48:25 logic that they had encoded back then.

1:48:27 But that just brings me to the conclusion

1:48:29 on the question for what should a young programmer do?

1:48:32 You're not gonna be able to predict the future.

1:48:34 No one's gonna be able to predict the future.

1:48:36 If you like programming, you should learn programming.

1:48:38 Now, is that gonna be a career forever?

1:48:40 I don't know, but what's gonna be a career forever?

1:48:42 Who knows?

1:48:44 Like a second ago, we thought that it was

1:48:47 the blue collar labor that was gonna be extracted first.

1:48:49 It was the robots that were gonna take over.

1:48:52 Then GenAI comes out, and then all the artists suddenly look like,

1:48:55 "Holy shit, is this gonna do all animation now?

1:48:58 Is gonna do all music now?" They get real scared.

1:49:00 And now I see the latest Tesla robot going like,

1:49:03 oh, maybe we're back now to blue collar being in trouble.

1:49:06 Because if it can dance like that, it can probably fix a toilet.

1:49:11 So no one knows anything, and you have to then position yourself for the future

1:49:17 in such a way that it doesn't matter that you pick a profession or path where if

1:49:24 it turns out that you have to retool and re-skill,

1:49:28 you're not gonna regret the path you took.

1:49:30 That's a general life principle.

1:49:32 For me, how I look at all endeavors I involve

1:49:35 myself in is I want to be content with all outcomes.

1:49:39 When we start working on a new product at 37signals,

1:49:42 I set up my mental model for its success.

1:49:45 And I go, do you know what?

1:49:47 If no one wants this, I will have

1:49:50 had another opportunity to write beautiful Ruby code,

1:49:53 to explore Greenfield domain,

1:49:55 to learn something new, to build a system I want, even if no one else wants it.

1:49:59 What a blessing.

1:50:01 What a privilege.

1:50:03 If a bunch of people want it, that's great.

1:50:06 We can pay some salaries, we can keep the business running,

1:50:09 and if it's a blow away success, wonderful.

1:50:12 I get to impact a bunch of people.

1:50:13 I think one of the big open questions to me

1:50:15 is how far you can get with vibe coding,

1:50:19 whether an approach for a young developer to invest most

1:50:23 of the time it's into vibe coding or into writing code from scratch.

1:50:28 So vibe coding, meaning, so I'm leaning into the meme a little bit,

1:50:32 but the vibe coding, meaning you generate code,

1:50:36 you have this idea of a thing you want to create,

1:50:38 you generate the code and then you fix it

1:50:41 with both natural language to the prompts and manually.

1:50:44 You learn enough to manually fix it.

1:50:46 So that's the learning process.

1:50:48 How you fix code that's generated,

1:50:50 or you write code from scratch and have the LMS kind of tab,

1:50:58 tab, tab, tab, add extra code.

1:51:00 Like which part do you lean on?

1:51:03 I think to be safe, you should find the beauty

1:51:07 and the artistry and the skill in both right from scratch.

1:51:12 So like there should be some percent of your time

1:51:14 just writing from scratch and some percent vibe coding.

1:51:16 There should be more of the time writing from scratch.

1:51:19 If you are interested in learning how to program,

1:51:22 unfortunately you're not gonna get fit by watching fitness videos.

1:51:25 You're not gonna learn how to play the guitar by watching YouTube guitar videos.

1:51:29 You have to actually play yourself.

1:51:31 You have to do the sit ups.

1:51:33 Programming, understanding, learning almost anything requires you to do.

1:51:37 Humans are not built to absorb information in a way

1:51:42 that transforms into skills by just watching others from afar.

1:51:45 Now ironically, it seems AI is actually quite good at that, but humans are not.

1:51:49 If you wanna learn how to become a competent programmer, you have to program.

1:51:55 It's really not that difficult to understand.

1:51:58 Now I understand the temptation.

1:52:00 And the temptation is there because vibe coding can produce things.

1:52:04 Perhaps in this moment, especially in a new domain you're not familiar

1:52:07 with, with tools you don't know perfectly well,

1:52:09 that's better than what you could do.

1:52:10 Or that you would take much longer

1:52:12 to get at, but you're not gonna learn anything.

1:52:15 You're gonna learn in this superficial way that feels like learning,

1:52:19 but is completely empty calories.

1:52:22 And secondly, if you can just vibe code it, you're not a programmer.

1:52:26 Then anyone could do it, which may be wonderful.

1:52:29 That's essentially what happened with the Access database.

1:52:31 That's what happened with Excel.

1:52:33 It took the capacity of accountants to become

1:52:38 software developers because the tools became so accessible

1:52:41 to them that they could build a model

1:52:43 for how the business was gonna do next week.

1:52:45 That required a programmer prior to Excel.

1:52:47 Now it didn't because they could do it themselves by coding enables

1:52:53 non-programmers to explore their ideas

1:52:56 in a way that I find absolutely wonderful.

1:52:59 But it doesn't make your a programmer.

1:53:02 I agree with you, but I wanna allow for room for both of us to be wrong.

1:53:06 For example, there could be, vibe coding could actually be a skill.

1:53:11 That if you train it, and by vibe coding,

1:53:14 let's include the step of correction, the iterative correction.

1:53:18 It's possible if you get really good

1:53:20 at that, that you're outperforming the people that run from scratch,

1:53:24 that you can come up with truly innovative things,

1:53:27 especially at this moment in history.

1:53:30 While the LLMs are a little bit too dumb

1:53:33 to create super novel things and a complete product,

1:53:36 but they're starting to creep close to that.

1:53:38 So if you are investing time now into becoming a really good vibe coder,

1:53:43 maybe this is the right thing to do as if it's indeed a skill,

1:53:48 we kind of meme about vibe coding, just like sitting back and it's in the name.

1:53:53 But if you treat it seriously,

1:53:55 a competitive vibe coder and get good at riding the wave of AI

1:54:00 and get good at the skill of editing code versus writing code from scratch,

1:54:07 it's possible that you can actually get farther in the long term.

1:54:12 Maybe editing is a fundamentally different task than writing from scratch

1:54:16 if you take that seriously as a skill that you develop.

1:54:19 To me that's an open question.

1:54:21 I just think, I personally, now you're on another level, but just me,

1:54:27 just personally, I'm not as good at editing the code that I didn't write.

1:54:35 No one is.

1:54:36 That's one different.

1:54:37 No one is of this generation, but maybe that's a skill.

1:54:42 Maybe if you get on the same page as the AI,

1:54:45 'cause there's a consistency to the AI.

1:54:47 It's like it really is a pair programmer

1:54:50 with a consistent style and structure and so on.

1:54:53 Plus with your own prompting, you can control the kind of wrote code you write.

1:54:56 I mean, it could generally be a skill like.

1:54:59 That's the dream of the prompt engineer.

1:55:01 I think it's complete pipe dream.

1:55:03 I don't think editors exist that aren't good at writing.

1:55:08 I've written a number of books.

1:55:10 I've had a number of professional editors.

1:55:12 Not all of them wrote their own great books,

1:55:15 but all of them were great writers in some regard.

1:55:17 You cannot give someone pointers if you don't know how to do it.

1:55:21 It's very difficult for an editor to be able to spot what's

1:55:26 wrong with a problem if you that it couldn't make the solution themselves.

1:55:31 Editing, in my opinion, is the reward,

1:55:34 the capacity to be a good editor is the reward you get from being a good doer.

1:55:39 You have to be a doer first.

1:55:40 Now that's not the same as saying that vibe coding,

1:55:45 prompt engineering won't be able to produce

1:55:48 fully formed amazing systems even shortly.

1:55:51 I think that's entirely possible, but then there's no skill left,

1:55:55 which maybe is the greatest payoff at all.

1:55:57 Wasn't that the whole promise of AI anyway,

1:55:59 that it was just all natural language that even my clumsy

1:56:02 way of formulating a question could result in a beautiful, succinct answer.

1:56:07 That actually to me is a much more

1:56:09 appealing vision that there's gonna be these special

1:56:11 prompt engineering wizards who know how to tickle

1:56:14 the AI just write to produce what they want.

1:56:17 The beauty of AI is to think that someone who doesn't know the first thing about

1:56:22 how AI actually works is able to formulate

1:56:25 their idea and their aspirations for what they want,

1:56:28 and the AI could somehow take that messy

1:56:31 clump of ideas and produce something that someone wants.

1:56:35 That's actually what programming has always been.

1:56:37 There's very often been people who didn't know how to program,

1:56:41 who wanted programs, who then hired programmers,

1:56:43 who gave them messy descriptions of what they wanted

1:56:46 and then when the programmers delivered that back said,

1:56:49 "Oh no, actually that's not what I meant.

1:56:50 I want something else." AI may be able to provide that cycle.

1:56:56 If that happens to the fullest extent of it,

1:56:58 yeah, there's not gonna be as many programmers around, right?

1:57:01 But hopefully presumably someone still, at least for the foreseeable future,

1:57:06 have to understand whether what the AI is producing actually works or not.

1:57:11 As a interesting case study, maybe a thought experiment.

1:57:14 If I wanted to vibe code Basecamp or HEY,

1:57:20 and some of the products you've built, like what would be the bottleneck?

1:57:26 Where would I fail along the way?

1:57:28 What I've seen when I've been trying to do this, trying to use

1:57:32 vibe coding to build something real is you actually failed really early.

1:57:36 The vibe coding is able to build a veneer at the current

1:57:40 present moment of something that looks like it works, right?

1:57:43 But it's flawed in all sorts of ways.

1:57:44 There are the obvious ways, the meme ways that it's leaking all your API keys,

1:57:49 it's storing your password in plain text.

1:57:52 I think that's ultimately solvable.

1:57:54 Like it's gonna figure that out or at least it's gonna get better

1:57:57 at that, but its capacity to get lost

1:58:00 in its own labyrinth is very great right now.

1:58:03 It code something and then you wanna change

1:58:06 something and it becomes a game of whack-a-mole.

1:58:08 Real quick, Pieter Levels who've been doing this wonderful flight

1:58:11 simulator was talking to that where at a certain scale,

1:58:14 the thing just keeps biting its own tail.

1:58:16 You wanna fix something and it breaks five other things,

1:58:20 which I think is actually uniquely human

1:58:21 because that's how most bad programmers are.

1:58:23 At the certain level of complexity with the domain,

1:58:25 they can't fix one thing without breaking three other things.

1:58:30 So in that way I'm actually in some way it's almost a positive signal

1:58:34 for that the AI is gonna figure this out

1:58:36 because it's done an extremely human trajectory right now.

1:58:39 The kind of mistakes it's making are the kind

1:58:41 of mistakes the junior programmers make all the time.

1:58:44 Yeah, can we zoom out and look at the vision,

1:58:48 the manifesto, the doctrine of Rails?

1:58:51 What are some of the things that make a programming language a framework great?

1:58:57 Especially for web development.

1:58:58 So we talked about happiness.

1:59:00 The underlying objective of Ruby.

1:59:03 What else?

1:59:05 So you're looking at the nine points I wrote out in I think 2012.

1:59:08 And first, before we dive into 'em, I wanna say the reason I wrote it down

1:59:12 is that if you want a community to endure,

1:59:17 you have to record its values and you have to record its practices.

1:59:21 If you don't, eventually you're gonna get enough new people come

1:59:24 in who have their own ideas of where this thing should go.

1:59:27 And if we don't have a a guiding light helping us to make decisions,

1:59:33 we're gonna start flailing.

1:59:34 We're gonna start actually falling apart.

1:59:36 I think this is one of the key

1:59:37 reasons that institutions of all kinds start falling apart.

1:59:40 We forget why Chesterton's Fence is there.

1:59:43 We just go like, why is that fence there?

1:59:45 Let's yank it out.

1:59:46 Oh, it was to keep the wolves out.

1:59:48 Now we're all dead, oops.

1:59:50 So I wanted to write these things down.

1:59:52 And if we just take them quick one by one,

1:59:54 you talked about optimizing for programmer happiness.

1:59:57 I put that at number one in homage of Matz,

1:59:59 and that's a lot about accepting that there is occasionally a trade

2:00:03 off between writing beautiful code and other things we want out of systems.

2:00:07 There could be a runtime trade off,

2:00:09 there can be a performance trade off, but we're gonna do it nonetheless.

2:00:13 We're also going to allow ambiguity in a way that many programmers,

2:00:18 by default, are uncomfortable with.

2:00:20 I give the example actually here of in the interactive Ruby shell where

2:00:26 you can play with the language or even interact with your domain model.

2:00:29 You can quit it in two ways, at least that I found.

2:00:33 You can write exit.

2:00:34 Boom, you're out of the program.

2:00:35 You can write quit.

2:00:37 Boom, you're out of the program.

2:00:38 They do the same thing.

2:00:39 We just wrote both exit or the people who built that wrote both exit

2:00:42 and quit because they knew humans were likely to pick one or the other.

2:00:45 Python is the perfect contrast to this.

2:00:49 In the Python interactive protocol, if you write exit, it won't exit.

2:00:53 It'll give you a lesson.

2:00:55 It'll basically tell you to read the fucking manual.

2:00:58 It says use exit parentheses or Ctrl-D i.e.

2:01:03 end of file to exit.

2:01:06 I'm like, one is very human and another is very engineer.

2:01:11 And I mean that both of them in the best possible way.

2:01:15 Python is pedantic.

2:01:16 Python is the value from the start stated is that there should

2:01:22 be preferably one and only one way to do a certain thing.

2:01:26 Ruby is the complete opposite.

2:01:28 No, we want the full expression that fits different human brains,

2:01:33 such that it seems like the language is guessing just what they want.

2:01:38 And part of that is also described the principle of least surprise,

2:01:41 which is a difficult thing to engineer into a language

2:01:44 because you have to kind of, it's a subjective thing.

2:01:47 Which is why you can't do it in one way,

2:01:48 which is why I used the example of both exit and quit.

2:01:50 The principle of least surprise for some people would be like, "Oh, exit.

2:01:54 That's how I get out of the prompt." For other people it'd be quit.

2:01:58 Why don't we just do both?

2:02:01 Okay, so what's the convention over configuration?

2:02:04 That's a big one.

2:02:05 That's a big one.

2:02:06 That's a huge one.

2:02:06 And it was born out of a frustration

2:02:09 I had in the early days with especially Java

2:02:12 frameworks where when you were setting up a web

2:02:14 application framework for Java back in the day,

2:02:15 it was not uncommon to literally write hundreds

2:02:18 if not thousands of lines of XML configuration files.

2:02:23 Oh, I need this.

2:02:23 I want the database to use the foreign keys as post_id.

2:02:29 No, no, no, I wanted as post capital ID.

2:02:33 Oh, no, no, no, you have to do it capital PID.

2:02:37 There are all these ways where you can configure how foreign

2:02:40 relation keys should work in a database and none of them matter.

2:02:43 We just need to pick one, and then that's fine.

2:02:46 And if we pick one and we can depend on it, it becomes a convention.

2:02:50 And if it's a convention, we don't have to configure it.

2:02:52 And if we don't have to configure it,

2:02:53 you can get started with you actually care about much quicker.

2:02:57 So convention over configuration is essentially to take

2:03:00 that idea that the system should come preassembled.

2:03:04 I'm not just handing you a box of fucking

2:03:06 Legos and asking you to build the Millennium Falcon.

2:03:08 I'm giving you a finished toy.

2:03:10 You can edit, you can change it, it's still build out a Legos.

2:03:13 You can still take some pieces off and put in some other pieces,

2:03:15 but I'm giving you the final product.

2:03:17 And this cuts against the grain of what most programmers love.

2:03:21 They love a box of Legos.

2:03:22 They love to put everything together from scratch.

2:03:24 They love to make all these detailed

2:03:26 little decisions that just don't matter at all.

2:03:28 And I want to elevate that up so that hey,

2:03:30 I'm not trying to take the decisions away from you.

2:03:32 I just want you to focus on decisions

2:03:34 that actually matter that you truly care about.

2:03:37 No one cares about whether it's post_id or post ID or PID.

2:03:41 Yeah, great defaults.

2:03:43 Yes.

2:03:43 It's just a wonderful thing.

2:03:45 You have all these aspirations, they're gonna do some kind of custom,

2:03:48 most beautiful Legos castle that nobody's ever built from these pieces.

2:03:52 But in reality to be productive in most situations,

2:03:56 you just need to build the basic thing.

2:04:00 And then on top of that is where your creativity comes.

2:04:03 Absolutely, and I think this is one of those, part of the doctrine that a lot

2:04:08 of programmers who get to use Ruby

2:04:10 on Rails begrudgingly will acknowledge it's a nice thing,

2:04:14 even if they don't really like it.

2:04:16 Like it's hard to beat the sort of attraction

2:04:21 to building with Legos from scratch out of programmers.

2:04:23 That's just what we like.

2:04:24 This is why we're programmers in the first place

2:04:26 because we'd like to put these little pieces together,

2:04:28 but we can direct that instinct towards a more productive end of the stack.

2:04:33 Okay, what are some of the other ones?

2:04:36 The menu is omakase.

2:04:37 It actually comes out of the same principle that great defaults really matter.

2:04:41 If you look at everything that's wrong with the JavaScript ecosystem right now,

2:04:45 for example, it is that no one is in charge of the menu.

2:04:49 There are a billion different dishes and you

2:04:52 can configure just your tailored specific configuration of it,

2:04:56 but no one done the work to make sure it all fits together.

2:04:59 So you have all these unique problems in the JavaScript ecosystem,

2:05:02 for example, there's probably 25 major ways of just doing the controller layer.

2:05:07 And then as many of how to talk to the database.

2:05:10 So you get this permutation of n times n

2:05:13 times n of no one is using the same thing.

2:05:16 And if they are using the same thing,

2:05:18 they're only using the same thing for about five minutes.

2:05:20 So we have no retained wisdom.

2:05:23 We build up no durable skills.

2:05:25 Rails goes the complete opposite way of saying, do you know what?

2:05:28 Rails is not just a white framework.

2:05:32 It's a complete attempt at solving the web problem.

2:05:36 It's a complete attempt at solving everything

2:05:38 you need to build a great web application.

2:05:40 And every piece of that puzzle should

2:05:44 ideally be in the box pre-configured, preassembled.

2:05:48 If you want to change some of those pieces later, that's wonderful.

2:05:51 But on day one you'll get a full menu designed by a chef

2:05:55 who really cared about every piece of ingredient and you're gonna enjoy it.

2:06:01 And that's again, one of those things

2:06:03 where many programmers think like I know better.

2:06:07 And they do in some hyper local sense of it.

2:06:10 Every programmer knows better.

2:06:12 This is what Ruby is built on.

2:06:13 That every programmer knows better in their specific situation.

2:06:16 Maybe they can do something dangerous.

2:06:19 Maybe they think they know better and then

2:06:20 they blow their foot off and then they

2:06:22 truly will know better because they've blown

2:06:24 their foot off once and won't do it again.

2:06:26 But the the menu of omakase is that.

2:06:29 So you in general see the value in the in the monolith?

2:06:33 Yes, the integrated system.

2:06:35 Integrated.

2:06:35 That someone thought of the whole problem.

2:06:37 This is one of the reasons why I've been

2:06:39 on a crusade against microservices since the term was coined.

2:06:43 Microservices was born out of essentially a good idea.

2:06:47 What do you do at Netflix scale when you have

2:06:50 thousands of engineers working on millions of lines of code?

2:06:53 No one can keep that entire system in their head at one time.

2:06:56 You have to break it down.

2:06:57 Microservices can be a reasonable way to do that.

2:06:59 When you're at Netflix scale,

2:07:01 when you apply that pattern to a team of 20 programmers working

2:07:05 on a code base of half a million lines of code, you're an idiot.

2:07:08 You just don't need to turn method invocations into network calls.

2:07:14 It is the first rule of distributed programming.

2:07:17 Do not distribute your programming.

2:07:20 It is makes everything harder.

2:07:22 All the failure conditions you have to consider as a programmer

2:07:25 just becomes infinitely harder when there's a network cable involved.

2:07:28 So I hate the idea of premature decomposition and microservices is exactly that.

2:07:35 The monolith says.

2:07:36 Let's try to focus on building a whole system that a single

2:07:40 human can actually understand and push that paradigm as far

2:07:43 as possible by compressing all the concepts such that more

2:07:47 of it will fit into memory of a single operating human.

2:07:50 And then we can have a system where I can actually understand all the Basecamp.

2:07:53 I can actually understand all of HEY.

2:07:55 Both of those systems are just over a hundred thousand lines of code.

2:07:59 I've seen people do this that maybe twice,

2:08:00 maybe three times that scale and then it starts breaking down.

2:08:03 Once you get north of certainly half a million lines of code,

2:08:06 no individual human can do it and that's when you

2:08:08 get into maybe some degree of microservices can make sense.

2:08:12 Basecamp and HEY are both a hundred thousand?

2:08:13 A hundred thousand lines of code thereabouts.

2:08:15 Wow, it's small.

2:08:16 It is.

2:08:17 Considering the fact that Basecamp I think has something like 420 screens,

2:08:21 different ways and configurations.

2:08:23 Ah, do you include the front end in that?

2:08:25 No, that's the Ruby code.

2:08:26 Well, it's front end in the sense that some of that Ruby codes.

2:08:29 It's beneficial to the front, but it's not JavaScript for example.

2:08:32 Now the other thing we might talk about later is

2:08:34 we write very little JavaScript actually for all of our applications.

2:08:38 HEY, which is a Gmail competitor.

2:08:40 Gmail ships I think 28 megabytes of uncompressed JavaScript.

2:08:44 If you compress it, I think it's about six megabytes, 28 megabytes.

2:08:47 Think about how many lines of code that is.

2:08:48 When HEY launched, we shipped 40 kilobytes.

2:08:53 It's trying to solve the same problem.

2:08:55 You can solve the email client problem with either 28 megabytes

2:08:59 of uncompressed JavaScript or with 40 kilobytes if you do things differently.

2:09:03 But that comes through the same problem essentially.

2:09:06 This is why I have fiercely fought splitting front end and back end apart.

2:09:12 That, in my opinion, this was one of the great crimes against web development.

2:09:16 That we are still atoning for.

2:09:20 That we separated and divided what was

2:09:23 and should be a unified problem solving mechanism.

2:09:27 When you are working both on front end and back end,

2:09:30 you understand the whole system.

2:09:31 And you're not going to get into these camps

2:09:34 that decompose and eventually you end up with like GraphQL.

2:09:39 Okay, let's fly through the rest of the doctrine.

2:09:43 No one paradigm.

2:09:44 No one paradigm goes to the fact that Ruby

2:09:46 is a fiercely object-oriented programming language at its core,

2:09:49 but it's also a functional program language.

2:09:52 This five times I told you about,

2:09:55 you can essentially do these anonymous function calls and you can chain 'em

2:10:00 together very much in the spirit

2:10:01 of how true functional programming languages worked.

2:10:04 Ruby has even moved closer towards the functional

2:10:06 programming and of the scale by making strings immutable.

2:10:11 There are ideas from all different disciplines and all

2:10:15 different paradigms of software development that can fit together.

2:10:18 Smalltalk, for example.

2:10:19 With only object oriented, and that was just it.

2:10:22 Ruby tries to be mainly object oriented,

2:10:25 but borrow a little bit of functional programming,

2:10:27 a little bit of imperative programming, be able to do all of that.

2:10:30 Rails tries to do the same thing.

2:10:32 We're not just gonna pick one paradigm and run it through everything.

2:10:35 Object orientation is at the center of it,

2:10:38 but it's okay to invite all these other

2:10:40 disciplines and it's okay to be inspired.

2:10:42 It's okay to remix it.

2:10:43 I actually think one of the main benefits of Rails is that it's a remix.

2:10:48 I didn't invent all these ideas.

2:10:50 I didn't come up with active record.

2:10:52 I didn't come up with the NBC way of dividing an application.

2:10:56 I took all the great ideas that I had learned

2:10:58 and picked up from every different camp and I put it together.

2:11:02 Not because there was gonna be just one single overarching theory of everything,

2:11:06 but I was gonna have a cohesive unit that incorporated the best from everywhere.

2:11:10 Is that idea a bit at tension with the beauty of the monolith system?

2:11:15 I think the monolith can be thought of as quite roomy,

2:11:20 quite as a big tent that the monolith needs actually to borrow

2:11:25 a little bit of functional programming

2:11:27 for the kinds of problems that that excels,

2:11:29 that discipline excels at solving, and that paradigm excels at solving.

2:11:33 If you also want object orientation at its core.

2:11:36 I actually think when I've looked at functional programming languages,

2:11:39 there's a lot to love.

2:11:40 And then I see some of the crazy contortions they have to go

2:11:44 through when part of the problem they're solving calls for mutating something.

2:11:49 And you go like, holy shit, this is a great paradigm from 90% of the problem.

2:11:56 And then you're twisting yourself completely out of shape

2:11:58 when you try to solve the last 10.

2:12:00 Ooh, exalt beautiful code is in the next one.

2:12:03 We've talked about that at length and here's

2:12:04 a great example that really summarizes the domain specific

2:12:07 language quality of Ruby on Rails that you

2:12:11 can make code actually pleasant to write and read.

2:12:16 Which is really funny to me because

2:12:17 as we talked about when I started learning programming,

2:12:20 it wasn't even a consideration.

2:12:21 I didn't even know that that could be part of the premise,

2:12:23 that could be part of the solution.

2:12:26 That writing code could feel as good as writing a poem.

2:12:31 class Project, ApplicationRecord belongs_to :account has many participants.

2:12:36 Class_name, Person, validates_presence_of :name.

2:12:41 See, you could read it out.

2:12:42 You didn't even change.

2:12:44 Anything like a haiku or something.

2:12:45 Right?

2:12:46 Isn't that beautiful?

2:12:47 Yeah, it's nice.

2:12:48 It's really nice.

2:12:50 There's an intuitive nature to it.

2:12:52 Okay, so I have specific questions there.

2:12:54 I mean ActiveRecord, just to take that tangent.

2:12:57 That has to be your favorite feature.

2:13:00 It's the crown jewel of Rails.

2:13:02 It really is.

2:13:04 It is the defining characteristic of how to work with Ruby on Rails.

2:13:07 And it's born in an interesting level of controversy,

2:13:12 because it actually uses a pattern that had

2:13:14 been described by Martin Fowler in the "Patterns

2:13:17 of Enterprise Application Architecture." One of the greatest

2:13:20 books for anyone working on business systems.

2:13:23 And if you had not read it, you must pick it up immediately.

2:13:26 "Patterns of Enterprise Application Architecture," I

2:13:29 think it was published in 2001.

2:13:30 It is one of the very few programming books that I have read many times over.

2:13:35 It's incredible.

2:13:36 In it, Martin describes a bunch of different

2:13:40 patterns of how to build business systems essentially.

2:13:42 And ActiveRecord is a little bit of a footnote in there.

2:13:45 The pattern is literally called active record.

2:13:47 You can look it up.

2:13:48 Nice.

2:13:48 It's called active record.

2:13:49 I wasn't even creative enough to come up with a name of my own.

2:13:52 But it allows the creation,

2:13:55 the marriage of database and object orientation in a way

2:13:59 that a lot of programmers find a little off-putting.

2:14:03 They don't actually want to pollute the beautiful

2:14:07 object-oriented nature of that kind of programming with SQL.

2:14:12 There was a rant by Uncle Bob the other

2:14:14 day about how SQL is the worst thing ever.

2:14:16 Baba, okay fine, whatever, I don't care.

2:14:20 This is practical.

2:14:22 We are making CRUD applications.

2:14:24 You're taking things out of an HTML form

2:14:26 and you're sticking 'em into a damn database.

2:14:28 It's not more complicated than that.

2:14:30 The more abstractions you put in between those two ends of the spectrum,

2:14:34 the more you're just fooling yourself.

2:14:36 This is what we're doing.

2:14:37 We're talking to SQL databases.

2:14:39 By the way, quick aside,

2:14:40 SQL was one of those things that have endured the onslaught of NoSQL databases

2:14:46 structured list data for a better part of a decade and still reign supreme.

2:14:50 SQL was a good thing to invest your time in learning.

2:14:53 Every program I'm working with the web should know SQL to a fair degree.

2:14:58 Even if they're working with an ORM,

2:15:00 an optic relational map or as active record.

2:15:02 You still need to understand SQL.

2:15:04 What Active record does is not so much try

2:15:07 to abstract the SQL away behind a different kind of paradigm.

2:15:10 It's just making it less cumbersome to write.

2:15:12 Making it more amenable to build domain models on top of other domain models

2:15:20 in a ways that you don't have to write every damn SQL statement by hand.

2:15:23 We'll just say the active record is an ORM,

2:15:26 which is a layer that makes it intuitive

2:15:30 and human interpretable to communicate with a database.

2:15:33 Even simpler than that.

2:15:34 It turns tables into classes and rows into objects.

2:15:40 I actually think SQL is very easy to understand most of it.

2:15:42 You can write some SQL golf too.

2:15:44 That's very hard to understand.

2:15:45 But SQL at its base,

2:15:47 and much of the criticism against SQL was it was written for human consumption.

2:15:51 It's actually quite verbose,

2:15:52 especially if you're doing things like inserts over and over again.

2:15:55 It's quite verbose insert into table parentheses,

2:15:59 enumerate every column you want to insert,

2:16:02 values, parentheses, every value that fits with that column.

2:16:05 It gets tedious to write SQL by hand, but it's actually very humanly readable.

2:16:11 Active record just takes that tedious away.

2:16:15 It makes it possible to combine things

2:16:17 in a way that a humanly describable language just doesn't.

2:16:21 It composes things into methods and you can combine

2:16:24 these methods and you can build structures around them.

2:16:25 So I don't dislike SQL, I dislike a lot of things in programming.

2:16:29 I try to get rid of them.

2:16:30 SQL wasn't really one of them.

2:16:32 It was just a sense of I don't wanna write the same thing over and over again.

2:16:35 It was a, can we be a little more succinct?

2:16:38 Can we match it just slightly better to the optic orientation without trying

2:16:42 to hide away the fact that we're persisting these objects into a database.

2:16:47 That's where I think a lot of ORMs went wrong.

2:16:50 They tried to live in the pure world of objects.

2:16:54 Never to consider that those objects had to be consistent into a SQL

2:16:57 database and then they came up

2:16:59 with convoluted way of translating back and forth.

2:17:02 Active record says, do you know what?

2:17:04 Just accept it.

2:17:05 This record, this object is not gonna get saved into some NoSQL database.

2:17:09 It's not gonna be saved.

2:17:11 It's gonna be saved into SQL database.

2:17:12 So it's just structure the whole thing around that.

2:17:14 It's gonna have attributes.

2:17:16 Those attributes are gonna respond to columns in the database.

2:17:19 It's not more complicated than that stuff making it so.

2:17:21 Yeah, but I should say, so I personally love SQL because I'm

2:17:25 an algorithms person and so I love optimization.

2:17:28 I love to know how the databases actually work so I can match

2:17:32 the SQL queries and the design of the table such that there is,

2:17:38 you know, optimal, squeeze the optimal performance out of the table.

2:17:42 Okay, based on the actual way that that table is used.

2:17:46 So I mean, I think that pushes to the point

2:17:48 that like there is value in learning in understanding SQL.

2:17:53 I wonder, because I started looking

2:17:56 at active record and it looks really awesome.

2:17:59 Does that make you lazy?

2:18:00 Not you, but a person that rolls in and starts using

2:18:04 Rails you can probably get away with never really learning SQL, right?

2:18:09 As long as you wanna stay at the entry level of competence.

2:18:13 And this is actually my overarching mission

2:18:15 with Rails is to lower the barrier of entry

2:18:18 so far down that someone can start seeing

2:18:22 stuff on their browser without basically understanding anything.

2:18:26 Yeah, they can run Rail's new blog, run a couple of generators,

2:18:31 they have a whole system, they don't understand anything.

2:18:34 But it's an invitation to learn more.

2:18:36 Where I get fired up, and this ties back to the AI discussion,

2:18:40 is when that's turned into this meme,

2:18:42 that programmers no longer have to be competent.

2:18:45 I mean the AI is gonna figure it out.

2:18:47 The generators is gonna figure it out.

2:18:49 I don't need to know SQL, active record is gonna abstract it away from me.

2:18:52 No, no, no dude, hold up.

2:18:54 The path here is competence.

2:18:56 I'm trying to teach you things.

2:18:58 I understand I can't teach you everything in five minutes.

2:19:01 No one who's ever become good at anything