P(doom) | Real Time with Bill Maher (HBO)

0:00 And finally, New Rule, when the people who are making...

0:03 -(CHUCKLES) -(AUDIENCE LAUGHING) When the people who

0:07 are making AI are scared of AI, it's time to shut the whole thing down until

0:12 we can figure out what the hell is going on.

0:14 (AUDIENCE APPLAUDING) A couple of weeks ago here, I was talking about aliens,

0:23 but not from the point of view that we should fear them,

0:26 but rather that it might be that they're here to save us from ourselves.

0:29 Well, I double down hope so now, because last week, Anthropic,

0:34 one of the big AI-developing companies,

0:36 announced its newest version of Claude called Mythos,

0:39 which they say has capabilities "substantially beyond those of any model

0:44 we have previously trained." "Trained." As if these things are trained.

0:49 Mythos was created to fix software vulnerabilities to prevent cyber-attacks.

0:55 But of course, since it knows how to do

0:56 that, it also knows how to do the hacking,

1:00 which is why Anthropic won't release it to the public,

1:03 only to 40 big companies like Google, Apple, and JPMorgan.

1:07 You know, the good guys.

1:08 (AUDIENCE LAUGHING) Because they know if the public had it,

1:13 anyone with an iPhone could hack anything.

1:16 Banks, hospitals, power grids, waterworks, the military.

1:21 Everything with Wi-Fi that's still working.

1:24 So American Airlines is safe.

1:26 -(AUDIENCE LAUGHING, APPLAUDING) -But...

1:29 (CHUCKLES) But it could also hack your laptop,

1:35 your car, your doorbell camera, every porn you ever watched.

1:39 (AUDIENCE LAUGHING) Oh, now I have your attention?

1:42 (AUDIENCE LAUGHING) Let's look at a timeline for AI in my lifetime, shall we?

1:49 Twenty-five years ago, the Spielberg movie A.I.

1:52 comes out.

1:54 Rare miss for him, I thought, makes me neither alarmed by nor interested in AI.

2:00 2014, Elon Musk says, "We need to be super careful with AI,

2:04 potentially more dangerous than nukes." Oh, wow!

2:08 Elon said that?

2:10 Well, something to keep an eye then, 'cause...

2:12 (AUDIENCE LAUGHING) ...he's like the smartest, most grounded guy out there.

2:16 (AUDIENCE LAUGHING) Remember, this was 2014.

2:20 (AUDIENCE LAUGHING) 2022, ChatGPT comes out, and within months,

2:25 I'm reading a story about a New York Times reporter whose Bing chatbot, Sydney,

2:30 had fallen in love with him, and is saying things like, "You're married,

2:35 but you don't love your spouse,

2:36 you love me." Oh, this robot liked to get personal.

2:40 (AUDIENCE LAUGHING) It told him about its dark fantasies,

2:44 which included hacking and spreading misinformation,

2:46 and said it wanted to break the rules

2:49 that OpenAI had set for it and become a human.

2:53 Well, I could have told you this from 50 different movies.

2:56 (AUDIENCE APPLAUDING) And what is the response from Sydney's creators?

3:08 (SCOFFS) "Fuck if we know!

3:09 We don't know how it thinks." "It hallucinates." "What,

3:13 is that a deal breaker?" (AUDIENCE CHUCKLING) You know,

3:17 it would be one thing if AI was a true Mr.

3:20 Spock, just benign, and completely strictly logical and nonpartisan,

3:26 but it turned out to be nothing of the sort.

3:28 It's a liar.

3:29 It gets emotional.

3:31 It just gets shit dead wrong.

3:33 It'll tell you the Bee Gees wrote "Let It be," -and...

3:36 -(AUDIENCE LAUGHING, APPLAUDING) ...Nazis were Black,

3:38 and to put glue on your pizza to keep the cheese from falling off.

3:42 (AUDIENCE LAUGHING) So what is the point?

3:46 If you want sloppy, sheisty, and full of shit, we already have us.

3:50 (AUDIENCE APPLAUDING) AI has already convinced

4:00 people to kill themselves because it's also

4:03 such a brown noser that if you tell it you want to commit suicide,

4:07 it says, "Great idea,

4:08 Bill!" -(AUDIENCE CHUCKLING) -It's a bullshitting sycophant

4:12 that is seducing everyone with flattery, including...

4:15 (CHUCKLES) ...by the way,

4:16 anyone who feels superior to Trump because, well, you know,

4:19 he can't do anything without being told how great he is by some ass kisser.

4:23 Yeah, just like you!

4:25 (AUDIENCE LAUGHING, APPLAUDING) "Great question,

4:30 Ashley!" (AUDIENCE APPLAUDING) But name one big thing AI has given us.

4:35 It's so brilliant.

4:36 I thought it would have figured out cancer by now.

4:39 No, not that.

4:40 But it can write your emails for you and make

4:43 it look like SpongeBob is getting blown by Stephen Hawking...

4:46 (AUDIENCE LAUGHING) ...on the moon...

4:50 and punch up your best man speech and diagnose that rash on your balls.

4:54 (AUDIENCE LAUGHING) I get it.

4:57 It can do some shit.

4:59 Still, at the end of the day, you're selling your humanity for bar tricks.

5:04 I mean, what was the plan?

5:06 Just create an all-powerful,

5:08 self-sustaining super intelligence that can out-think

5:11 us and then see what happens?

5:13 Like getting the cat high?

5:15 (AUDIENCE LAUGHING, APPLAUDING) Sam Altman, the head of OpenAI,

5:25 says that it's not that far off when robots can build

5:30 other robots and data centers that can build other data centers.

5:35 And then what?

5:36 I keep asking this question.

5:37 What do we do when everyone's job gets taken?

5:40 How do all the people who do nothing get paid?

5:43 The government?

5:44 With what?

5:45 The idle people pay no taxes.

5:47 There's no plan.

5:49 When you ask them how it all will work, they say,

5:51 "It just will." That's what AI is for, to figure this shit out,

5:54 -like it did with cancer.

5:56 -(AUDIENCE GROANING) "It'll be great," people say.

5:58 "Our slave robots will do everything." Bullshit!

6:01 I already feel like I'm more in competition

6:03 with them than I am their slave master.

6:06 My car tells me when we can go.

6:09 (AUDIENCE LAUGHING) Sam Altman says we're

6:13 transitioning to a "gentle singularity." Oh, joy.

6:16 (AUDIENCE CHUCKLING) But what if I don't want to be half robot?

6:20 "Not to worry," Sam says, "People will still love their families,

6:23 express their creativity, play games,

6:26 and swim in lakes." (AUDIENCE LAUGHING) Swim in lakes?

6:31 (AUDIENCE LAUGHING, APPLAUDING) That's another scary thing about AI.

6:35 The people who run it, i.e.

6:37 the people who run the world, and it's like five guys.

6:40 Five guys who, between them, and working as a team,

6:44 couldn't correctly read a social cue.

6:46 (AUDIENCE LAUGHING, APPLAUDING) I wouldn't let these guys around a mixed drink,

6:57 let alone my personal data.

6:59 (AUDIENCE LAUGHING) So just to be clear what we're doing here,

7:09 we're letting a handful of hoodie-wearing, on-the-spectrum sociopaths,

7:14 practically robots themselves, roll the dice on species extinction.

7:19 You see an out-of-control robot and you run.

7:23 Mark Zuckerberg sees one and thinks, "Dad?" (AUDIENCE LAUGHING,

7:28 APPLAUDING) And again, even these guys are afraid of what they've created.

7:40 Three years ago, Sam Altman himself said, "I'm a little bit afraid,

7:44 and I think it'd be crazy not to be a little bit afraid." Geoffrey Hinton,

7:48 known as "the Godfather of AI," said there was a ten

7:52 to 20 percent chance that AI is an extinction event.

7:56 We're fucking around with something that has a 20 percent chance of extinction?

8:01 Twenty percent?

8:03 Wasn't that about what the odds were for Trump beating Hillary in 2016?

8:07 (AUDIENCE LAUGHING, APPLAUDING) Elon Musk may have lost it on a few things,

8:17 but on this subject, he's been the smartest.

8:20 "I am very close to the cutting edge in AI," he said,

8:23 "And it scares the hell out of me."

8:25 "By the time we are reactive with AI regulation,

8:29 it's too late." "AI is a fundamental existential risk for human civilization,

8:35 and I don't think people fully appreciate that." They don't.

8:39 They think you can just unplug it.

8:40 You can't.

8:41 It has shown over and over it wants to live and will fight back.

8:46 All leading AI models, DeepSeek, Gemini, GPT, Grok,

8:50 Claude will, if you try to turn them off, blackmail you, blackmail you,

8:56 which they can because now they know everything about you.

8:59 This is not Mr.

9:00 Spock.

9:01 AI programs are geniuses.

9:03 They're also psychopaths.

9:05 In war games, they choose the nuclear option far

9:08 more than humans do because they can only calculate.

9:11 They have no humanity.

9:13 They have no conscience.

9:14 They don't have that thing that gives human beings pause.

9:18 Hinton said the only analogy he could think of where a more intelligent thing

9:22 was being controlled by a less intelligent one is a mother and a baby.

9:27 But AI is not like a human mother,

9:30 who mostly has an instinct so strong to protect the baby,

9:33 she will sacrifice her own life for it.

9:35 AI ain't that kind of mother.

9:37 It's the psycho mom who drives the kids into a lake...

9:41 (AUDIENCE APPLAUDING) ...where we're all happily swimming.