Addicted to Prime Numbers (full interview) - Numberphile

0:05 I am an organic chemist.

0:06 I'm more or less a scientist who's like eager to find new things.

0:11 And um when I when I heard from the BOINC,

0:15 the Berkeley Open Infrastructure Network

0:17 for distributed computing, like 15 years ago,

0:21 I was immediately addicted to finding prime

0:23 numbers because by the time they um yeah,

0:29 still had like low low effort goals and all

0:33 this or like like um numbers you can find.

0:36 Uh yeah, then I I got interested in in in these numbers,

0:41 but started to realize the the records are not for me because

0:47 I don't have I don't have like a lot of computing power.

0:52 And um yeah, so I concentrated on on the smaller numbers.

0:57 Small in the in the sense of of records.

0:59 I mean, you just have to do 10,000 digits.

1:01 Do you have many uh scalps?

1:03 Have you got many primes under your belt?

1:05 Like how many of how many have you been responsible for?

1:08 Uh I had like two other records

1:12 from the in the official database for prime numbers,

1:15 uh which was a twin prime, which was top 10 by the time.

1:19 And I still have I just had a look looked up today,

1:22 still have a uh triple prime uh top five.

1:26 So, that's that's something.

1:29 Yeah.

1:30 So, you're a chemist by day,

1:31 but prime numbers just kind of um tickle your brain, do they?

1:35 Uh it's it's it's a it's a seasonal addiction.

1:38 So, once or twice a year,

1:39 I get like super addicted to like finding a prime number.

1:43 And um I don't know why.

1:46 It's it's just the the the kick of it to to find something new, I guess.

1:53 And uh yeah, with with the with these prime kind of prime numbers.

1:57 There is no software to sieve for them,

1:59 so um I try with my limited programming skills

2:04 as as uh Matt um says it all the time, it's like terrible Python code.

2:10 In In my case, it's terrible C code,

2:12 so because I need the speed to to for sieving.

2:15 And um yeah, so I try to to program my my own

2:19 sieves and uh check the prime numbers or check for prime numbers.

2:24 What computer did you use to actually do the business?

2:28 Um I used a a mini PC, which is basically laptop hardware,

2:33 so eight cores, 16 threads, and uh yeah,

2:37 started in the morning to crunch on the the candidates

2:41 till I I got home back from work.

2:44 Then the leftover candidates I uh checked

2:48 again for for the reversible prime um property.

2:52 So uh yeah, it took me I don't know,

2:55 like 2 weeks or so to to find the the record prime.

2:59 Um it was honestly it was sooner than expected.

3:02 I thought it was going on for weeks that I have to to sieve for primes,

3:08 check check the candidates, and yeah.

3:11 When you got that first batch of candidates that you

3:13 were then going to reverse to to find the the winner, how big was that list?

3:18 How many were on that list on the short list?

3:21 Uh on the short list um I can't remember.

3:23 It was uh roughly about 10,000 candidates from a batch of I think 20 million.

3:30 20 million was done in about a work day, which is just okay for me,

3:34 so I can't run it overnight because the computer's in my my bedroom.

3:39 And um yeah, so the the main work was done over the day when I was not home,

3:44 and the the leftover 10,000 candidates were done

3:47 during having a cup of coffee, so Right.

3:51 So, when you found your your winner, your new number, how did that play out?

3:55 Were you at work and you came home and it was on the computer

3:58 or how did you Tell me about the moment the moment you found it.

4:02 Actually, I was a bit stumped that it happened so soon.

4:05 I was pretty lucky finding them.

4:08 Uh finding number and um Yeah,

4:11 I was doing the the business as usual like the the other week before.

4:16 Reversing the numbers, start starting the script again.

4:19 I was like, yeah, okay, it's going to be running.

4:20 No no number will be prime and suddenly there was

4:23 like one candidate left and I was like, okay, that's cool.

4:27 It was also a bit underwhelming because like, huh, now it's done.

4:31 What to do next, right?

4:32 Did you Did you think you'd made a mistake?

4:34 Did you think this can't be true or like

4:37 Um oh, I I double-checked and triple-checked.

4:39 I I printed out the numbers like in in normal form and reverse form.

4:44 Then I wrote um a script to like check them digit by digit just to see if they

4:50 are like um the same and the reversed and not

4:54 reversed or like if you reverse the one candidate, it will give the other.

4:59 And uh yeah, I double triple-checked and then I was like, yeah, okay,

5:02 that's that's really really a reversible prime and then I knew

5:06 the main work is still to come because with these numbers,

5:09 they are probable prime.

5:11 Um so um you really have to prove them prime by um factoring or so.

5:18 And uh yeah, the factorization was done last year.

5:23 Uh it took me 2 and 1/2 days per number,

5:26 constant computing time on on this eight eight-core CPU.

5:31 Uh but yeah, both numbers proved to be proved to be prime and the um

5:36 results are uploaded on a on a central

5:38 database for for these kind of checks, yeah.

5:41 So, who were the first people you tell when you when you do this?

5:44 Is Is a community you tell?

5:46 Do you upload somewhere?

5:48 Do you email Matt?

5:49 Well, um I thought about emailing Matt right at the beginning.

5:54 I was like, "Yeah, okay, I will I will do it someday." Right?

5:57 First First person I I talked to was my girlfriend.

6:00 She was like, "Cool." Yeah,

6:02 she Does she really think it's cool or is she into this stuff?

6:05 Um she She's not into this stuff.

6:07 She She thinks it's it's pretty cool that I found something,

6:11 but um she really can't um yeah,

6:14 see the the nerdy value of of these kind of numbers.

6:19 And uh the community, well, um there's there there are forums,

6:23 like the Mersenne Forum, for example, which all the experts hang out.

6:28 Um I didn't tell them because it's it's was just my my small project.

6:32 I uploaded to to the database, as I said.

6:36 Uh and a person came back to me and was like, "Hey,

6:38 are you Were you the guy who proven this this number

6:40 prime because it's also in the top 400 of like,

6:45 size-wise, numbers that were proven prime by this um method." I was like,

6:50 "Yeah, that it was me." He was like, "Okay,

6:53 cool." Matt explained to me the the how you

6:56 used this technique of using the ones that had

6:58 lots of zeros and a one cuz it gave you like something to get your teeth into.

7:03 Yeah.

7:04 And I As I understood it,

7:05 I then said to Matt, "Okay, so was this low-hanging fruit?

7:09 Was this, you know, was he going for something

7:11 that was e- easier than what he could have gone

7:14 for, which was a clever thing to do?" Would

7:16 you describe this as like low-hanging fruit or Uh honestly,

7:19 I I don't know whether it's Well, it's easier than another method.

7:23 That's was the only um form, I would say,

7:27 of this number that I could uh grasp with my limited math knowledge.

7:33 Um and there was the the previous record was was of the same form,

7:37 and I kind of saw the um saw the pattern.

7:40 I was like, "Yeah, makes sense to construct a number like this.

7:44 And um so I I used the same form as the previous record owner and yeah,

7:50 worked worked out for me.

7:51 And again, everything more difficult would be

7:55 a huge challenge for me to program.

7:58 What next, Stefan?

7:59 Are you going to go for another one?

8:00 Are you going to Are you going to stay with these e-merps, these 000001 e-merps,

8:05 or are you going to go for I thought

8:07 briefly about the possibility of having as a twin prime,

8:11 but that this doesn't work out.

8:13 I thought about the pandigital primes.

8:16 I thought you also made a video about it.

8:18 These are prime numbers with all the numbers in it at least once.

8:23 But this is like I'm still thinking

8:25 about how to generate the candidates for this.

8:28 So currently I'm I'm satisfied with my math uh uh urge to to define something.

8:37 So um No no current no current project.

8:42 So when you found the the new biggest one,

8:45 the biggest version of it, did you do a quick check for twinness?

8:49 Did you Did you add two and subtract two to check if it was a twin?

8:53 Uh you can't do this with these numbers

8:54 because uh it would end Would it Could it?

9:01 I I I did not check the four for twinness.

9:03 So because by the time I had the the record I

9:05 didn't think about the having twin prime a twin prime twin project.

9:10 But um yeah, I I don't think I can realize it programming-wise.

9:16 So yeah.

9:18 Someone's going to do it now, you realize.

9:20 Yeah, of of course, and I really hope about

9:22 the commenters to to say what I did wrong,

9:25 what I did badly, and how easy it was to to find a bigger one.

9:30 But that's the thing,

9:31 my record is just four digits longer than the previous one,

9:34 and it's it's an encouragement for someone uh someone else out there

9:38 to to break my record and get a new um reversible prime.

9:44 Do you prefer calling them reversibles or emirp's?

9:47 I don't know.

9:48 Um I just stick with the wording reversible.

9:52 But yeah.

9:54 Well, it's strictly speaking yes,

9:56 they're also the palindromic primes in the term of reversible.

10:00 And one last thing, have this this number that's kind of, you know,

10:04 you found as your little discovery, your little gold nugget.

10:07 Have you done anything with it?

10:08 Have you had it like printed out and put it in a frame or on a t-shirt or have

10:13 you done anything to to bond with the number

10:16 or does it just exist in in a computer chip?

10:19 That that's currently just existing in a text file,

10:23 but print printing it out is is actually a fun idea.

10:27 Um I have to check how how many pages this this is, like 10,000 digits.

10:32 It's not that as many as Matt did, so yeah.

10:35 I'm I'm I'm just a chemist.

10:37 I'm I'm doing math stuff poorly, so yeah.

10:41 Let me ask you this then.

10:43 Do you watch my chemistry videos?

10:45 Of course, I do.

10:46 Oh, Like like periodic videos is uh I think

10:50 I watched periodic videos before I watched number file.

10:54 You make me so happy, Stefan.

10:55 Thank you.

10:55 Yeah, always always especially the um

10:58 the very early episodes about just every element.

11:01 Yeah.

11:02 I love it.

11:03 We're still doing them.

11:04 We're still doing We're updating them.

11:06 I've got another big element coming in the next few weeks.

11:08 Yeah.

11:09 Oh, that's nice.

11:09 Yeah.

11:09 Yeah.

11:10 Yeah.

11:10 Stay tuned.

11:11 And we're we're updating all the element ones cuz some

11:13 of them weren't very good at the start, so Yeah.

11:16 Yeah.

11:16 Yeah.

11:16 Yeah, it's it's perfect.

11:17 I mean, the experiments you do or the lab technician does.

11:21 Yeah, that's that's warms my heart.

11:23 It's it's the thing I I love to see.

11:26 What what kind of chemistry are you working on?

11:27 Are you going to give us a Can you tell us anything that you're doing?

11:29 Like give us a little tiny sniff.

11:32 We can't tell you much about that.

11:34 Sorry.

11:35 Um Ooh, that sounds even more mysterious than your prime number work.

11:39 [laughter] Well Uh and another record prime I found

11:42 my girlfriend wants to wrote me a nice card.

11:45 And uh first before that she asked me,

11:48 "What was your record prime number again?" I was like,

11:51 "Weird question." I sent it to her.

11:53 Um and she wanted to to uh make me a card about I don't know something.

11:59 But Yeah.

12:01 As as she as she gave me a card, she was like,

12:03 "I kind of realized this is a lot of digits and I

12:05 can't like put these digits on the card." I was like

12:08 Okay, so she wanted to put it onto like

12:09 a nice card for you and then she Oh, hang on.

12:12 and I think this was like 30,000 digits or so.

12:14 And uh Yeah, I can't can't write them out.

12:18 Yeah, she sounds like me.

12:19 Every time I find out about these numbers, I'm like, "Oh,

12:21 I can't wait to put them on screen or something." And then I'm like,

12:23 "Oh, no." Yeah, 50-minute scroll.

12:26 Yeah.

12:29 [laughter] Show me the computer that did it.

12:32 Is that little one there?

12:33 Is that under the Yeah, it it's it's a mini computer with like uh eight cores.

12:39 That's it.

12:40 That's where the world's largest reversible prime was found.

12:45 In there.

12:46 Yeah.

12:53 They're going to let me touch it with my hands.

12:56 Watch carefully.

12:57 Only time in my life I'm going to wash my hands in diamonds.

13:05 Feels just like sand.

13:07 It's quite strange.

13:09 It feels a bit like metal.