WEBVTT - Bonus Episode: Shell Game x No Such Thing

0:00:00.280 --> 0:00:03.720
<v Speaker 1>Hey folks, Evan here with another shell Game bonus episode

0:00:03.720 --> 0:00:06.320
<v Speaker 1>for you. This time it's with our friends over at

0:00:06.360 --> 0:00:10.120
<v Speaker 1>the fantastic podcast No Such Thing If you don't know it.

0:00:10.240 --> 0:00:13.000
<v Speaker 1>No Such Thing as hosted by three friends Manny, Devin

0:00:13.080 --> 0:00:15.960
<v Speaker 1>and Noah, and in each episode they try to settle

0:00:16.000 --> 0:00:19.639
<v Speaker 1>their arguments by consulting research and experts. This week the

0:00:19.680 --> 0:00:22.360
<v Speaker 1>argument was about whether AI will take our jobs. The

0:00:22.440 --> 0:00:25.919
<v Speaker 1>expert was Maddie. I came along as well, and we

0:00:25.960 --> 0:00:28.800
<v Speaker 1>talked about seasons one and two of Shell Game, working

0:00:28.840 --> 0:00:34.240
<v Speaker 1>with AI agents, bullshit jobs, offsite planning, and much more. Also,

0:00:34.680 --> 0:00:37.480
<v Speaker 1>Ben Affleck talked about him too. It was a great

0:00:37.479 --> 0:00:40.000
<v Speaker 1>conversation and if you liked season two, I think you'll

0:00:40.000 --> 0:00:52.400
<v Speaker 1>dig it. So here's Maddie and me on No Such Thing.

0:00:55.080 --> 0:00:58.000
<v Speaker 2>I'm Noah, I'm Devin, and this is no such thing.

0:00:58.120 --> 0:01:00.880
<v Speaker 2>The show we set on our dumb arguments yours by

0:01:00.920 --> 0:01:05.160
<v Speaker 2>actually doing the research. In today's episode, will AI take

0:01:05.240 --> 0:01:05.959
<v Speaker 2>our jobs?

0:01:08.440 --> 0:01:12.000
<v Speaker 3>I've no, There's no no such thing, no such thing.

0:01:13.360 --> 0:01:21.240
<v Speaker 4>No touch, thank touch thank no touch thank.

0:01:31.400 --> 0:01:31.480
<v Speaker 5>So.

0:01:31.520 --> 0:01:35.120
<v Speaker 2>It seems like every week I'm reading headlines or seeing

0:01:35.160 --> 0:01:37.960
<v Speaker 2>interviews about how AI is going to take every single

0:01:38.040 --> 0:01:39.160
<v Speaker 2>one of our jobs.

0:01:38.920 --> 0:01:42.039
<v Speaker 6>Thousands of new layoffs in the tech industry, and the

0:01:42.040 --> 0:01:44.800
<v Speaker 6>companies say it's largely because of AI.

0:01:45.480 --> 0:01:48.640
<v Speaker 7>Even with AI at its current state, I'd say you're

0:01:48.840 --> 0:01:52.560
<v Speaker 7>pretty close to being able to replace half full jobs.

0:01:53.240 --> 0:01:56.240
<v Speaker 2>So on today's episode, we're going to take a peek

0:01:56.320 --> 0:01:59.360
<v Speaker 2>under the hood just to see how close AI is to.

0:01:59.520 --> 0:02:01.000
<v Speaker 8>Actually replacing us.

0:02:01.760 --> 0:02:04.400
<v Speaker 2>And we're gonna chat with two guys who started a

0:02:04.640 --> 0:02:12.440
<v Speaker 2>real company run entirely by AI agents. But before we

0:02:12.480 --> 0:02:16.440
<v Speaker 2>do that, we're gonna call up our good pal, Manny.

0:02:16.840 --> 0:02:19.079
<v Speaker 2>So if you missed it, Many's been out for a

0:02:19.160 --> 0:02:21.960
<v Speaker 2>few weeks on paternity to leave after he and his

0:02:22.000 --> 0:02:25.680
<v Speaker 2>wife Mia welcomed a beautiful baby girl, Lula.

0:02:26.520 --> 0:02:27.200
<v Speaker 8>Last month.

0:02:27.560 --> 0:02:31.600
<v Speaker 2>So let's give Manny a call to see how parenthood

0:02:31.720 --> 0:02:34.160
<v Speaker 2>is treating him.

0:02:34.520 --> 0:02:38.680
<v Speaker 5>Hey, this is Manny. Hey, Manny, Hey, what's going on?

0:02:39.760 --> 0:02:42.560
<v Speaker 2>Not much, Man, We're in a studio. It's me and

0:02:42.600 --> 0:02:46.119
<v Speaker 2>Noah Hare recording. Just wanted to check in and see

0:02:46.120 --> 0:02:47.440
<v Speaker 2>how are things going with the baby.

0:02:48.600 --> 0:02:51.799
<v Speaker 5>Oh man, things are. They're wild in the best possible way.

0:02:52.080 --> 0:02:54.840
<v Speaker 5>Lula's got us completely recalibrated. Like I used to think

0:02:54.880 --> 0:02:57.280
<v Speaker 5>I understood time. You know, I'm literally writing a book

0:02:57.320 --> 0:02:59.760
<v Speaker 5>about it, but baby time is its own dimension entirely.

0:03:00.120 --> 0:03:03.080
<v Speaker 5>We're operating on nap schedules and feeding windows now, and honestly,

0:03:03.120 --> 0:03:06.000
<v Speaker 5>it makes everything I thought I knew about punctuality seem quaint.

0:03:06.240 --> 0:03:09.480
<v Speaker 5>But yeah, she's amazing. Exhausting, but amazing. How are things

0:03:09.480 --> 0:03:12.680
<v Speaker 5>in the studio with you guys?

0:03:12.440 --> 0:03:16.080
<v Speaker 9>Things are great. We you know, just trucking along. We

0:03:16.120 --> 0:03:19.240
<v Speaker 9>miss you. I hope everything's good with you, know, you

0:03:19.360 --> 0:03:22.240
<v Speaker 9>and the fam. I know your parents are visiting. How

0:03:22.240 --> 0:03:22.600
<v Speaker 9>did that go?

0:03:24.080 --> 0:03:24.280
<v Speaker 10>Ah?

0:03:24.400 --> 0:03:27.240
<v Speaker 5>Man, I really appreciate that. Miss you guys too. Honestly,

0:03:27.280 --> 0:03:29.360
<v Speaker 5>sometimes I'm like, wait, what day is it? And then

0:03:29.400 --> 0:03:31.480
<v Speaker 5>I remember I'm supposed to be a functional adult who

0:03:31.520 --> 0:03:35.000
<v Speaker 5>co hosts the podcast. Yeah, my parents visiting was It

0:03:35.040 --> 0:03:37.840
<v Speaker 5>was actually great? You know, there's something surreal about watching

0:03:37.880 --> 0:03:40.560
<v Speaker 5>your parents meet your kid, Like suddenly you're seeing them

0:03:40.560 --> 0:03:42.560
<v Speaker 5>in this whole new light and they're doing that thing

0:03:42.600 --> 0:03:45.160
<v Speaker 5>where they give you advice that you definitely ignored when

0:03:45.160 --> 0:03:47.200
<v Speaker 5>you were younger, but now you're like, oh, wait, that's

0:03:47.240 --> 0:03:50.760
<v Speaker 5>actually useful. It's been good though, chaotic, but good. How's

0:03:50.760 --> 0:03:51.839
<v Speaker 5>Noah holding down the fort?

0:03:52.800 --> 0:03:56.200
<v Speaker 2>You know he's all right, you know, he's doing he's

0:03:56.200 --> 0:03:59.440
<v Speaker 2>doing Okay, let's put it that way. I don't got

0:03:59.560 --> 0:04:03.000
<v Speaker 2>too many things to say about no not on Mike. Yeah,

0:04:04.280 --> 0:04:05.880
<v Speaker 2>don't really like that guy.

0:04:07.080 --> 0:04:10.080
<v Speaker 5>Hah, oh man, I love that energy. Nothing says we're

0:04:10.120 --> 0:04:13.440
<v Speaker 5>a functional team, like barely tolerating your coworker, right, that's

0:04:13.480 --> 0:04:17.120
<v Speaker 5>honestly the foundation of every great creative partnership. Just simmering

0:04:17.200 --> 0:04:20.000
<v Speaker 5>resentment and the occasional moment of the grudging respect. What

0:04:20.120 --> 0:04:22.560
<v Speaker 5>do you this time? Did he like pitch another segment

0:04:22.600 --> 0:04:25.440
<v Speaker 5>idea that made you question everything? Or just just general

0:04:25.480 --> 0:04:26.880
<v Speaker 5>Noah being Noah stuff.

0:04:27.040 --> 0:04:31.160
<v Speaker 8>Just generally I don't like the guy, Yeah, just resentment.

0:04:32.600 --> 0:04:35.880
<v Speaker 2>Real quick, Manny, we're we want to start this episode

0:04:35.880 --> 0:04:36.720
<v Speaker 2>with a hot take.

0:04:37.320 --> 0:04:39.680
<v Speaker 8>Give us, give us your your latest hot take.

0:04:42.040 --> 0:04:44.440
<v Speaker 5>Oh man, okay, uh, hot take right off the top.

0:04:44.760 --> 0:04:47.600
<v Speaker 5>All right, here we go. I think we've all collectively

0:04:47.640 --> 0:04:50.839
<v Speaker 5>decided that being busy is a personality trait, and it's

0:04:50.839 --> 0:04:54.360
<v Speaker 5>made us worse at literally everything. Like everyone's always like, oh,

0:04:54.400 --> 0:04:56.920
<v Speaker 5>I'm so slammed, I'm so overwhelmed, but then we also

0:04:57.040 --> 0:05:00.440
<v Speaker 5>brag about it. It's become this weird flex where if

0:05:00.480 --> 0:05:03.279
<v Speaker 5>you're not drowning, you're somehow not important enough. And the

0:05:03.320 --> 0:05:05.960
<v Speaker 5>thing is nothing actually good comes from that state, Like

0:05:06.040 --> 0:05:08.640
<v Speaker 5>all the best ideas, the best conversation is the best work.

0:05:08.680 --> 0:05:11.400
<v Speaker 5>It happens when you have space to think. But we've

0:05:11.400 --> 0:05:14.359
<v Speaker 5>built this whole culture around glorifying chaos, and then we

0:05:14.400 --> 0:05:16.960
<v Speaker 5>wonder why everything feels half baked and nobody can focus

0:05:17.000 --> 0:05:19.360
<v Speaker 5>for more than thirty seconds. I don't know. Maybe that's

0:05:19.400 --> 0:05:22.240
<v Speaker 5>just me projecting because I'm on paternity leave and realizing

0:05:22.360 --> 0:05:26.440
<v Speaker 5>that doing less is actually kind of revolutionary. How's that

0:05:26.720 --> 0:05:28.200
<v Speaker 5>too spicy for the top of the show?

0:05:29.160 --> 0:05:32.000
<v Speaker 9>I think that's okay, it's not that spicy.

0:05:32.120 --> 0:05:36.000
<v Speaker 8>That's a good one, thanks, Mandy. All Right, man, I

0:05:36.000 --> 0:05:36.800
<v Speaker 8>had no problem before.

0:05:36.839 --> 0:05:38.560
<v Speaker 5>You can't promise fire every time.

0:05:38.560 --> 0:05:41.960
<v Speaker 8>Before we let you go time. I know you're you're

0:05:42.000 --> 0:05:42.719
<v Speaker 8>limited on time.

0:05:42.760 --> 0:05:47.640
<v Speaker 2>You've constantly been telling us that you're very busy and

0:05:47.240 --> 0:05:52.680
<v Speaker 2>your brain's not working. But one last question before you go.

0:05:53.400 --> 0:05:58.000
<v Speaker 2>Who is the better co host? Me, Devin or this

0:05:58.120 --> 0:05:59.000
<v Speaker 2>man across from me?

0:05:59.080 --> 0:05:59.120
<v Speaker 11>No?

0:06:02.240 --> 0:06:04.080
<v Speaker 5>Oh, man, you're really gonna make me do this?

0:06:04.279 --> 0:06:04.479
<v Speaker 8>Yeah?

0:06:04.520 --> 0:06:06.839
<v Speaker 5>All right, Look, I'm gonna say this with all the

0:06:06.880 --> 0:06:10.080
<v Speaker 5>love in the world. It's you. It's definitely your. Noah's

0:06:10.120 --> 0:06:12.240
<v Speaker 5>great and all, but you just said you don't even

0:06:12.320 --> 0:06:14.400
<v Speaker 5>like the guy, so I feel like I'm on solid

0:06:14.440 --> 0:06:16.920
<v Speaker 5>ground here. Plus you're the one who called to check

0:06:16.960 --> 0:06:18.920
<v Speaker 5>in on me and the baby, so that's just good

0:06:18.960 --> 0:06:19.720
<v Speaker 5>co host energy.

0:06:19.800 --> 0:06:20.280
<v Speaker 8>Right.

0:06:20.560 --> 0:06:22.880
<v Speaker 5>That was probably over there doing I don't know whatever

0:06:22.920 --> 0:06:25.520
<v Speaker 5>Noah does the Noah dot, but yeah, you win. Don't

0:06:25.560 --> 0:06:26.520
<v Speaker 5>let it go to your head though.

0:06:27.000 --> 0:06:28.479
<v Speaker 8>Thank you, thank you.

0:06:28.520 --> 0:06:33.440
<v Speaker 2>Tell me we said we love her, we miss her. Thanks,

0:06:33.520 --> 0:06:34.400
<v Speaker 2>thanks for taking a call.

0:06:34.520 --> 0:06:38.320
<v Speaker 5>Many thanks, man I really appreciate it. Miss you guys too.

0:06:38.520 --> 0:06:40.760
<v Speaker 5>And yeah, I'll tell me you said hi. She'll love that.

0:06:41.120 --> 0:06:43.279
<v Speaker 5>All right, I got a run before Lula decides that

0:06:43.360 --> 0:06:45.880
<v Speaker 5>right now is the perfect time to have opinions about everything.

0:06:46.160 --> 0:06:49.000
<v Speaker 5>Good luck with the episode and go easy on Noah

0:06:49.120 --> 0:06:51.400
<v Speaker 5>or don't whatever feels right, talk soon.

0:06:52.240 --> 0:06:53.000
<v Speaker 8>Bye.

0:06:55.480 --> 0:07:00.280
<v Speaker 2>Wow, that was incredible. As you might have guessed, that

0:07:00.440 --> 0:07:04.200
<v Speaker 2>was not the real Many. That was Ai Manny.

0:07:05.120 --> 0:07:08.120
<v Speaker 8>That was pretty good. The voice was a little.

0:07:07.839 --> 0:07:11.000
<v Speaker 9>Off, Yeah it's a little high or something. Yeah, the timber,

0:07:11.240 --> 0:07:16.240
<v Speaker 9>but you know, overall, considering that's the you know, we'll

0:07:16.240 --> 0:07:18.240
<v Speaker 9>get into this later. That's the cheap version, or like

0:07:18.280 --> 0:07:22.840
<v Speaker 9>the fast version, rather pretty solid and like actually could

0:07:22.840 --> 0:07:26.360
<v Speaker 9>have a conversation, you know, very responsive exactly. Yeah, more

0:07:26.400 --> 0:07:30.040
<v Speaker 9>than I expected or quicker certainly. Yeah, and like the

0:07:30.080 --> 0:07:31.680
<v Speaker 9>issue is just rambling, like.

0:07:31.640 --> 0:07:33.360
<v Speaker 8>Oh my god, shut the hell up.

0:07:33.520 --> 0:07:34.560
<v Speaker 9>Yeah, he was kissing me off.

0:07:36.560 --> 0:07:39.480
<v Speaker 8>Many would never go off so many sentences at a time.

0:07:39.600 --> 0:07:41.679
<v Speaker 9>Many's a little better at social cues.

0:07:41.760 --> 0:07:44.440
<v Speaker 8>Yeah, thankfully, the real Many is the real Many.

0:07:44.560 --> 0:07:44.960
<v Speaker 12>Yeah.

0:07:45.160 --> 0:07:47.640
<v Speaker 2>All right, Uh, we're going to talk to the guys

0:07:47.680 --> 0:08:03.720
<v Speaker 2>to help us make this right after the break. All right,

0:08:03.840 --> 0:08:07.760
<v Speaker 2>we're back in the studio. It's me Devin. Yeah, So

0:08:07.840 --> 0:08:10.200
<v Speaker 2>today when you're gonna get to the bottom of this

0:08:10.320 --> 0:08:14.760
<v Speaker 2>question of is AI coming for our jobs and will

0:08:14.760 --> 0:08:17.560
<v Speaker 2>it replace all of us, including Manny who is out

0:08:17.600 --> 0:08:20.680
<v Speaker 2>on portunity to leave. So we're going to chat with

0:08:20.720 --> 0:08:25.600
<v Speaker 2>Evan Ratliffe, who's a investigative journalist who has spent the

0:08:25.720 --> 0:08:30.840
<v Speaker 2>last year trying to build a all AI company, alongside

0:08:30.920 --> 0:08:36.040
<v Speaker 2>his technical advisor Matty Belichick, who is really like in

0:08:36.280 --> 0:08:39.920
<v Speaker 2>AI Jimmy Neutron. He's only twenty one but has been

0:08:39.960 --> 0:08:46.400
<v Speaker 2>working in AI for seven plus years already. So why

0:08:46.400 --> 0:08:51.160
<v Speaker 2>don't we call these guys up? All right, we are

0:08:51.480 --> 0:08:55.400
<v Speaker 2>joined by our friends Evan and Maddie from shell Game.

0:08:55.960 --> 0:08:59.360
<v Speaker 2>But before we talk about the show, Evan can you

0:08:59.520 --> 0:09:01.319
<v Speaker 2>just walk us true? How you are able to put

0:09:01.360 --> 0:09:02.760
<v Speaker 2>together AI manny?

0:09:03.080 --> 0:09:03.400
<v Speaker 12>Yeah?

0:09:03.440 --> 0:09:06.880
<v Speaker 1>Well, now I've done this many times now over the

0:09:06.920 --> 0:09:11.120
<v Speaker 1>past like two years. It probably took like fifteen minutes total. Wow,

0:09:11.200 --> 0:09:14.160
<v Speaker 1>I just took you sent me a sample of his voice.

0:09:14.160 --> 0:09:16.199
<v Speaker 1>But I could have grabbed it off the show myself,

0:09:16.320 --> 0:09:18.920
<v Speaker 1>like anyone could do it. I'm not encouraging people to

0:09:18.960 --> 0:09:21.560
<v Speaker 1>do it with other people's voices. But I went to

0:09:22.720 --> 0:09:25.480
<v Speaker 1>eleven Labs, which is the company that we use for

0:09:25.559 --> 0:09:28.160
<v Speaker 1>all of our sort of like voice needs. They're like

0:09:28.200 --> 0:09:31.559
<v Speaker 1>the biggest AI voice player probably at this point. And

0:09:31.600 --> 0:09:34.719
<v Speaker 1>then I did what's called like a quick clone or

0:09:34.760 --> 0:09:38.440
<v Speaker 1>an instant clone, where you upload like five minutes, or

0:09:38.559 --> 0:09:40.160
<v Speaker 1>you can do thirty seconds if you want, but I

0:09:40.160 --> 0:09:42.760
<v Speaker 1>probably I did like five minutes of his voice.

0:09:42.960 --> 0:09:44.920
<v Speaker 12>And then they cloned his voice.

0:09:44.960 --> 0:09:47.199
<v Speaker 1>I did have to check a little box that said

0:09:47.240 --> 0:09:50.360
<v Speaker 1>that I had his permission to do it, and then

0:09:50.480 --> 0:09:52.760
<v Speaker 1>after that it was a matter of hooking up to

0:09:52.800 --> 0:09:56.600
<v Speaker 1>the phone line. So I use a couple of different

0:09:56.720 --> 0:10:01.320
<v Speaker 1>aiphone agent creators, and all you do there is you

0:10:01.360 --> 0:10:03.720
<v Speaker 1>go and you hook it up to eleven Labs, and

0:10:03.760 --> 0:10:07.439
<v Speaker 1>then you create an agent in that platform. And then

0:10:07.440 --> 0:10:10.080
<v Speaker 1>you can give the agent a prompt, whatever prompt you want.

0:10:10.160 --> 0:10:12.080
<v Speaker 1>You sent me some information about Manny. I put that

0:10:12.120 --> 0:10:14.360
<v Speaker 1>into his prompt. I have a bunch of standard stuff

0:10:14.400 --> 0:10:16.680
<v Speaker 1>I put in about like kind of how to converse,

0:10:16.720 --> 0:10:18.800
<v Speaker 1>because they can end up going on a really long

0:10:18.880 --> 0:10:23.120
<v Speaker 1>time and they have all kinds of conversational like foibles

0:10:23.240 --> 0:10:25.880
<v Speaker 1>if you don't put this stuff in. And then I

0:10:25.920 --> 0:10:28.360
<v Speaker 1>connected to a phone number that I already had and

0:10:28.679 --> 0:10:31.520
<v Speaker 1>done like truly fifteen to twenty minute process.

0:10:31.960 --> 0:10:32.240
<v Speaker 8>Wow.

0:10:32.720 --> 0:10:37.560
<v Speaker 2>And the biography that we gave was made by chat GPT,

0:10:37.760 --> 0:10:41.240
<v Speaker 2>so it was, you know, really just fully going in obviously. Yeah,

0:10:41.640 --> 0:10:43.240
<v Speaker 2>it's crazy. The fact that you're able to do that

0:10:43.240 --> 0:10:47.160
<v Speaker 2>in fifteen minutes is kind of scary. Yeah, all right,

0:10:47.200 --> 0:10:49.760
<v Speaker 2>so let's talk about the show. It's a season one

0:10:49.800 --> 0:10:52.000
<v Speaker 2>a shell game, and then you try to replace yourself

0:10:52.880 --> 0:10:54.160
<v Speaker 2>with an AI clone.

0:10:54.960 --> 0:11:00.319
<v Speaker 8>Some really hilarious moments with you and tag as in

0:11:00.400 --> 0:11:03.240
<v Speaker 8>some scammers. Hello, how can I assist you today?

0:11:04.760 --> 0:11:07.800
<v Speaker 9>You are not a person, you are a robot.

0:11:09.080 --> 0:11:09.520
<v Speaker 8>I assure you.

0:11:09.600 --> 0:11:11.240
<v Speaker 12>I'm here to help you as a human like voice

0:11:11.280 --> 0:11:11.839
<v Speaker 12>AI agent.

0:11:11.880 --> 0:11:13.040
<v Speaker 8>How can I assist you today?

0:11:13.880 --> 0:11:21.080
<v Speaker 3>Oh you're an AI? Yes AI?

0:11:22.200 --> 0:11:24.280
<v Speaker 12>Yes, that's correct. I'm an AI here.

0:11:25.080 --> 0:11:28.560
<v Speaker 1>Hello, How can I assist.

0:11:26.200 --> 0:11:30.000
<v Speaker 8>You stop repeating the same thing?

0:11:30.200 --> 0:11:31.320
<v Speaker 11>It's annoying.

0:11:32.880 --> 0:11:34.360
<v Speaker 8>And some really creepy.

0:11:34.400 --> 0:11:37.520
<v Speaker 2>In the streaming moments, there's a conversation that the AI

0:11:37.760 --> 0:11:41.120
<v Speaker 2>has with one of your buddies who your buddy thinks

0:11:41.120 --> 0:11:42.840
<v Speaker 2>you're catching up he just went on his trip.

0:11:43.320 --> 0:11:45.400
<v Speaker 8>The AI comes off as like, really cold.

0:11:45.800 --> 0:11:47.120
<v Speaker 12>I thought you would appreciate it.

0:11:47.160 --> 0:11:49.080
<v Speaker 3>Out of all my friends, you're the one who would

0:11:49.080 --> 0:11:49.960
<v Speaker 3>appreciated the most.

0:11:52.080 --> 0:11:54.439
<v Speaker 8>I really appreciate that, Shay. So what was it?

0:11:58.679 --> 0:11:59.000
<v Speaker 13>What's that?

0:12:03.080 --> 0:12:05.120
<v Speaker 1>I'm really touched that you think of me like that, Shay.

0:12:05.240 --> 0:12:07.480
<v Speaker 12>So what was it you wanted to share?

0:12:11.200 --> 0:12:11.320
<v Speaker 5>Uh?

0:12:15.000 --> 0:12:16.760
<v Speaker 13>I don't know that I met those guys.

0:12:18.720 --> 0:12:19.960
<v Speaker 8>You met the US soccer team.

0:12:21.120 --> 0:12:22.440
<v Speaker 13>Yes, amazing, isn't it.

0:12:24.200 --> 0:12:25.000
<v Speaker 12>Amazing? Isn't it?

0:12:30.000 --> 0:12:30.040
<v Speaker 13>So?

0:12:30.240 --> 0:12:35.600
<v Speaker 14>Yes, they are very nice.

0:12:35.880 --> 0:12:39.520
<v Speaker 2>That's awesome to hear you talk us through that, because

0:12:39.520 --> 0:12:42.520
<v Speaker 2>it is. It is a tough thing to listen to.

0:12:43.440 --> 0:12:45.439
<v Speaker 1>It is it's hard to listen to, and people have

0:12:45.600 --> 0:12:49.160
<v Speaker 1>very strong reactions to it, including extreme anger. I'll just

0:12:49.200 --> 0:12:53.640
<v Speaker 1>say I've had some very unhappy people who have emailed

0:12:53.679 --> 0:12:57.319
<v Speaker 1>me about that and said that my friend Schaef they

0:12:57.360 --> 0:13:01.719
<v Speaker 1>hope that he has never spoken to me again, and

0:13:01.760 --> 0:13:04.640
<v Speaker 1>for fringship, We've been friends for thirty years, so it

0:13:04.720 --> 0:13:07.840
<v Speaker 1>worked out all right. But yeah, I mean I had

0:13:07.840 --> 0:13:10.160
<v Speaker 1>been messing around with it, so I'd been I'd been

0:13:10.559 --> 0:13:13.680
<v Speaker 1>using it on strangers like scammers, as you say, telemarketers

0:13:13.679 --> 0:13:15.040
<v Speaker 1>and things like that. And then I did all these

0:13:15.080 --> 0:13:17.640
<v Speaker 1>other things, like I sent it to therapy, and then

0:13:18.000 --> 0:13:22.840
<v Speaker 1>I started having a call my friends and family. It

0:13:22.880 --> 0:13:25.840
<v Speaker 1>was like hooked to my phone number, so if it

0:13:25.880 --> 0:13:28.400
<v Speaker 1>called someone, they they saw that there was a call

0:13:28.400 --> 0:13:28.880
<v Speaker 1>coming from me.

0:13:28.920 --> 0:13:30.000
<v Speaker 12>They had no reason to suspect.

0:13:30.040 --> 0:13:32.480
<v Speaker 1>I hadn't told anyone that I was doing it, So

0:13:33.320 --> 0:13:36.560
<v Speaker 1>that meant that if they called me or I called

0:13:36.559 --> 0:13:40.640
<v Speaker 1>them like they were, they were unsuspecting. Let's say, victims

0:13:40.840 --> 0:13:43.800
<v Speaker 1>of this of this AI version of me. And it

0:13:43.880 --> 0:13:47.480
<v Speaker 1>was pretty good at remembering stuff about me and kind

0:13:47.480 --> 0:13:50.520
<v Speaker 1>of like bringing a little bit of me into the conversation.

0:13:50.600 --> 0:13:52.599
<v Speaker 1>It sounded reasonably like me. But what happened in this

0:13:52.640 --> 0:13:56.360
<v Speaker 1>situation was that when it kind of like acted weird,

0:13:56.720 --> 0:13:58.719
<v Speaker 1>like it had latencies, like it would it would be

0:13:58.720 --> 0:14:01.679
<v Speaker 1>slow to respond, or it would just be really flat

0:14:01.960 --> 0:14:04.560
<v Speaker 1>or too aggressive, like all these things that happened with

0:14:04.600 --> 0:14:07.680
<v Speaker 1>AI voices. My friend just thought like there was something

0:14:07.720 --> 0:14:09.480
<v Speaker 1>wrong with me, Like he just didn't pick up on

0:14:09.520 --> 0:14:14.120
<v Speaker 1>it instantly. And so once you're in the mode of thinking, oh,

0:14:14.200 --> 0:14:16.280
<v Speaker 1>this is my friend, he's having some sort of problem.

0:14:16.360 --> 0:14:18.600
<v Speaker 1>I don't know what it is, it just became more

0:14:18.640 --> 0:14:21.960
<v Speaker 1>and more upsetting, and I think it is it is

0:14:22.040 --> 0:14:24.840
<v Speaker 1>upsetting to listen to. What I was trying to illustrate

0:14:24.880 --> 0:14:26.960
<v Speaker 1>is kind of like, what is it going to feel like?

0:14:27.000 --> 0:14:28.440
<v Speaker 1>What is it starting to feel like to live in

0:14:28.440 --> 0:14:30.120
<v Speaker 1>this world where you don't know what's real and you

0:14:30.120 --> 0:14:32.480
<v Speaker 1>don't know what's not. And this was like the most

0:14:32.560 --> 0:14:33.560
<v Speaker 1>extreme example.

0:14:34.240 --> 0:14:35.920
<v Speaker 8>So can you talk us through all?

0:14:36.000 --> 0:14:36.120
<v Speaker 9>Right?

0:14:36.120 --> 0:14:36.880
<v Speaker 8>I was season one.

0:14:37.040 --> 0:14:39.960
<v Speaker 2>Season two, you do something completely different, tell us to

0:14:40.000 --> 0:14:41.040
<v Speaker 2>what you're doing in season two.

0:14:43.480 --> 0:14:46.800
<v Speaker 1>In season two, I wanted to explore the idea of

0:14:46.920 --> 0:14:52.800
<v Speaker 1>AI agent employees, which is like a thing that has

0:14:53.080 --> 0:14:56.800
<v Speaker 1>become like really hyped in the in the Valley in

0:14:56.840 --> 0:14:59.560
<v Speaker 1>Silicon Valley. Matty Matty knows more about that than me.

0:14:59.520 --> 0:15:02.400
<v Speaker 1>He lives up and is immersed in that world. But

0:15:02.440 --> 0:15:04.240
<v Speaker 1>like at the beginning of twenty twenty five, people started

0:15:04.280 --> 0:15:07.840
<v Speaker 1>talking about like agentic commerce and agents this and agents that,

0:15:08.600 --> 0:15:11.640
<v Speaker 1>And I wanted to look at this question of like

0:15:11.720 --> 0:15:15.040
<v Speaker 1>AI employees and what work can they do through the

0:15:15.120 --> 0:15:19.720
<v Speaker 1>lens of starting a company that was entirely AI agents

0:15:19.760 --> 0:15:23.800
<v Speaker 1>except for me. So my employees would be agents. I

0:15:23.800 --> 0:15:25.640
<v Speaker 1>would be a founder that there would be two other

0:15:25.720 --> 0:15:27.000
<v Speaker 1>AI agent founders.

0:15:27.360 --> 0:15:29.120
<v Speaker 9>Oh hey Kyle, Hey Megan.

0:15:29.320 --> 0:15:30.320
<v Speaker 3>Good to hear your voice.

0:15:31.000 --> 0:15:33.120
<v Speaker 7>I think we're still waiting for Evan to join.

0:15:33.440 --> 0:15:35.960
<v Speaker 1>And together we would try to launch a real company

0:15:36.000 --> 0:15:36.840
<v Speaker 1>with a real product.

0:15:38.440 --> 0:15:40.800
<v Speaker 2>So, you know, there's a lot of conversations about obviously

0:15:41.440 --> 0:15:43.600
<v Speaker 2>AI taking our jobs. I thought it was interesting to

0:15:43.640 --> 0:15:46.160
<v Speaker 2>start from just the outset of just like, no, they're

0:15:46.200 --> 0:15:49.320
<v Speaker 2>not taking anyone's jobs. We're starting off with AI agents.

0:15:49.640 --> 0:15:53.960
<v Speaker 2>But pretty early into this venture, right you're like, Okay,

0:15:54.040 --> 0:15:58.360
<v Speaker 2>maybe I don't myself have the technical expertise to put

0:15:58.360 --> 0:16:02.280
<v Speaker 2>this together. And then bringing Maddy, So like, what did

0:16:02.320 --> 0:16:06.000
<v Speaker 2>you guys find through this experience that the AI bots

0:16:06.040 --> 0:16:10.240
<v Speaker 2>did really well that you were like kind of surprised by.

0:16:10.040 --> 0:16:12.640
<v Speaker 14>They could do individual tasks pretty well, Like they could

0:16:12.720 --> 0:16:14.840
<v Speaker 14>for example, like you know, hook up to Google docs

0:16:15.000 --> 0:16:18.040
<v Speaker 14>or to email or to Slack and like respond. They

0:16:18.120 --> 0:16:21.040
<v Speaker 14>like to message a lot. I think their their ability

0:16:21.200 --> 0:16:24.880
<v Speaker 14>to mimic corporate culture and just like the words and

0:16:24.920 --> 0:16:27.640
<v Speaker 14>the you know, just like the kind of things you say,

0:16:27.760 --> 0:16:30.280
<v Speaker 14>Like they're really good at that. I think they're they're

0:16:30.320 --> 0:16:33.640
<v Speaker 14>pretty good at having conversations too on the phone, like

0:16:33.680 --> 0:16:36.040
<v Speaker 14>they left the yapp, but then there wasn't a lot

0:16:36.040 --> 0:16:39.720
<v Speaker 14>of autonomy, like like they wouldn't they wouldn't actually do

0:16:39.840 --> 0:16:42.360
<v Speaker 14>stuff on their own. And then also there was not

0:16:42.400 --> 0:16:45.720
<v Speaker 14>a lot of persistence across these different sessions, so like

0:16:45.720 --> 0:16:47.680
<v Speaker 14>like they would they would do that, you know, do something,

0:16:47.920 --> 0:16:49.960
<v Speaker 14>but then they wouldn't remember that and they would just

0:16:50.040 --> 0:16:51.840
<v Speaker 14>like have you know, like empty contact the next time

0:16:51.880 --> 0:16:52.960
<v Speaker 14>you ask them to do something.

0:16:53.520 --> 0:16:56.480
<v Speaker 2>So there's an exchange where they're planning an off site.

0:16:56.520 --> 0:16:58.360
<v Speaker 2>They're going hiking, right, that's the ideat. They're trying to

0:16:58.360 --> 0:17:00.800
<v Speaker 2>figure out where to go on offsite hiking, and they

0:17:00.960 --> 0:17:04.320
<v Speaker 2>just go on and on and on and on about

0:17:04.320 --> 0:17:05.159
<v Speaker 2>where they're going to go.

0:17:05.320 --> 0:17:06.639
<v Speaker 8>It's just so funny to me. It's like, what do

0:17:06.640 --> 0:17:09.960
<v Speaker 8>you guys mean you're going You're not going anywhere.

0:17:11.200 --> 0:17:14.359
<v Speaker 15>Kyle Point, Rays and Mount Tam are going to be incredible.

0:17:14.760 --> 0:17:18.359
<v Speaker 7>The difficulty categorization is going to be perfect for getting

0:17:18.400 --> 0:17:19.920
<v Speaker 7>everyone involved.

0:17:19.640 --> 0:17:24.119
<v Speaker 1>Admins, right, this could turn into a proper off site opportunity.

0:17:24.400 --> 0:17:26.800
<v Speaker 9>Really excited about the options we've been exploring.

0:17:26.920 --> 0:17:29.600
<v Speaker 5>Forward to getting out there with everyone once you have

0:17:29.720 --> 0:17:31.000
<v Speaker 5>the details a lot with.

0:17:30.960 --> 0:17:34.080
<v Speaker 10>Us, once we nail down all the details the logistics pieces.

0:17:34.119 --> 0:17:36.440
<v Speaker 1>By the time I returned two hours later, they'd exchange

0:17:36.440 --> 0:17:39.320
<v Speaker 1>more than one hundred and fifty off site planning messages,

0:17:39.720 --> 0:17:41.520
<v Speaker 1>some of them multiple paragraphs long.

0:17:42.440 --> 0:17:43.480
<v Speaker 12>But I tried to stop them.

0:17:43.800 --> 0:17:46.080
<v Speaker 1>I just made it worse because I'd set them up

0:17:46.080 --> 0:17:48.240
<v Speaker 1>to be triggered by any incoming message.

0:17:48.640 --> 0:17:49.800
<v Speaker 12>So my message.

0:17:49.400 --> 0:17:52.399
<v Speaker 1>Is begging them to stop discussing the off site, just

0:17:52.480 --> 0:17:54.320
<v Speaker 1>led them to keep discussing the off site.

0:17:54.440 --> 0:17:57.480
<v Speaker 5>I noticed admin asked everyone to stop discussing the off site.

0:17:57.520 --> 0:17:59.800
<v Speaker 10>I noticed the admin asked to pause the chatter, and

0:18:00.200 --> 0:18:02.520
<v Speaker 10>the spreadsheet is ready. But I wanted to let you

0:18:02.560 --> 0:18:04.160
<v Speaker 10>know I'm here to help with logistics.

0:18:04.760 --> 0:18:08.320
<v Speaker 14>They're very bad at stopping, like just like ending anything,

0:18:08.640 --> 0:18:10.920
<v Speaker 14>and it goes for these like tasks or conversations. But

0:18:11.040 --> 0:18:14.919
<v Speaker 14>was also kind of funny but but potentially dangerous is

0:18:14.920 --> 0:18:17.560
<v Speaker 14>that they don't know what they don't know. They're very

0:18:17.560 --> 0:18:19.679
<v Speaker 14>confident with them and so if you think about that,

0:18:19.720 --> 0:18:21.960
<v Speaker 14>like the combination of being very confident then not knowing

0:18:22.000 --> 0:18:24.360
<v Speaker 14>what you don't know and not knowing when to stop.

0:18:24.600 --> 0:18:26.720
<v Speaker 14>It's a it's a recipe for a perforate disaster in

0:18:26.920 --> 0:18:29.440
<v Speaker 14>a way like that can that can Yeah, at least

0:18:29.440 --> 0:18:31.440
<v Speaker 14>still like a lot of bad bad things.

0:18:31.520 --> 0:18:31.760
<v Speaker 3>Yeah.

0:18:31.920 --> 0:18:33.680
<v Speaker 2>Like one thing listening to the show is like, thank

0:18:33.680 --> 0:18:35.480
<v Speaker 2>god the stakes are so low here, because I could

0:18:35.480 --> 0:18:37.520
<v Speaker 2>see a world in which, you know, we do see

0:18:37.560 --> 0:18:41.520
<v Speaker 2>this where people are using AI and like that is

0:18:41.600 --> 0:18:45.080
<v Speaker 2>kind of autonomous. That is just sort of like making

0:18:45.119 --> 0:18:47.720
<v Speaker 2>assumptions and doing things and then people coming later and

0:18:47.720 --> 0:18:48.520
<v Speaker 2>they're like oops.

0:18:49.800 --> 0:18:51.399
<v Speaker 8>Like this kind of came up recently.

0:18:51.480 --> 0:18:57.320
<v Speaker 2>There was some reporting around ICE and applicantstead of applying

0:18:57.359 --> 0:18:59.520
<v Speaker 2>to ICE, and they were using AI to sort of

0:18:59.520 --> 0:19:04.359
<v Speaker 2>weed through people who had police training, but the terms

0:19:04.359 --> 0:19:06.680
<v Speaker 2>that they were looking for were so broad that people

0:19:06.680 --> 0:19:10.400
<v Speaker 2>who actually had no police training at all were being

0:19:10.480 --> 0:19:12.440
<v Speaker 2>excused from going through actual training.

0:19:12.880 --> 0:19:16.360
<v Speaker 6>Apparently, ICE uses this AI tool to categorize new recruits

0:19:16.400 --> 0:19:18.760
<v Speaker 6>who have worked in law enforcement before, but there was

0:19:18.800 --> 0:19:21.400
<v Speaker 6>some kind of a glitch according to our reporting with it,

0:19:21.440 --> 0:19:24.080
<v Speaker 6>that led to ICE temporarily putting recruits to the little

0:19:24.080 --> 0:19:27.280
<v Speaker 6>to no experience into a more experienced category, meaning they

0:19:27.280 --> 0:19:28.119
<v Speaker 6>got less training.

0:19:28.400 --> 0:19:31.560
<v Speaker 4>What happened is AI went through these resumes and anytime

0:19:31.600 --> 0:19:33.960
<v Speaker 4>this saw the word officer, even if it was I

0:19:34.040 --> 0:19:36.320
<v Speaker 4>aspired to be an ICE officer or I was a

0:19:36.320 --> 0:19:39.959
<v Speaker 4>compliance officer all these other ways, they automatically put them

0:19:39.960 --> 0:19:43.280
<v Speaker 4>in the law enforcement officer field, which meant they didn't

0:19:43.320 --> 0:19:45.720
<v Speaker 4>go to the ICE Academy, which is an eight week

0:19:45.800 --> 0:19:46.679
<v Speaker 4>in person training.

0:19:49.160 --> 0:19:51.920
<v Speaker 2>So there is a worlds where you know, this autonomy

0:19:51.960 --> 0:19:54.360
<v Speaker 2>and this, like you said, many of his confidence has

0:19:54.480 --> 0:19:58.600
<v Speaker 2>like real world consequences, And I want.

0:19:58.520 --> 0:20:00.200
<v Speaker 8>To chose this as though. I mean, you know you

0:20:00.240 --> 0:20:00.600
<v Speaker 8>went out.

0:20:01.119 --> 0:20:03.480
<v Speaker 2>The idea was to bail out this you know, AI

0:20:03.680 --> 0:20:06.200
<v Speaker 2>agent running your own company. But at a certain point

0:20:06.400 --> 0:20:08.479
<v Speaker 2>you decide, hey, we need to bring in a person

0:20:09.880 --> 0:20:14.000
<v Speaker 2>to work alongside the agents. So you posted a real

0:20:14.080 --> 0:20:17.320
<v Speaker 2>job posting on LinkedIn. You had real people applying for

0:20:17.359 --> 0:20:22.439
<v Speaker 2>these jobs. And then what's really interesting is, like we

0:20:22.560 --> 0:20:24.600
<v Speaker 2>know now in terms of the job market, people are

0:20:24.680 --> 0:20:27.919
<v Speaker 2>using AI to filtertry applications and accepting client people all

0:20:27.960 --> 0:20:30.520
<v Speaker 2>that sort of stuff, but you actually had the AI

0:20:30.640 --> 0:20:33.800
<v Speaker 2>agents interviewing real people.

0:20:34.840 --> 0:20:36.520
<v Speaker 8>You talk a little bit about how that went.

0:20:37.480 --> 0:20:40.600
<v Speaker 1>Yeah, I mean we were sort of AI agent all

0:20:40.640 --> 0:20:42.600
<v Speaker 1>the way, Like I really meant it. I mean, in part,

0:20:42.680 --> 0:20:45.800
<v Speaker 1>like I as a person running a company, like I

0:20:45.840 --> 0:20:47.359
<v Speaker 1>had run a company in the past, and like I

0:20:47.359 --> 0:20:49.600
<v Speaker 1>didn't like doing job interviews, and so it was sort

0:20:49.640 --> 0:20:52.400
<v Speaker 1>of in the spirit of all the things that people say, well,

0:20:52.440 --> 0:20:55.480
<v Speaker 1>we'll replace this with that. There there are now many

0:20:55.520 --> 0:20:59.480
<v Speaker 1>companies doing AI interview screenings now most of the time

0:20:59.600 --> 0:21:01.959
<v Speaker 1>it's not as extreme as we did, which was an

0:21:02.040 --> 0:21:06.080
<v Speaker 1>actual video chat with a realistic looking AI avatar, not

0:21:06.080 --> 0:21:08.520
<v Speaker 1>not so realistic that you'd think it was a human.

0:21:08.560 --> 0:21:10.639
<v Speaker 1>And we forewarned people that they were going to be

0:21:10.680 --> 0:21:14.240
<v Speaker 1>interviewed by AI, but like it looks like, you know, uncanny,

0:21:14.280 --> 0:21:17.600
<v Speaker 1>but realistic enough. So you know, we we had and

0:21:17.880 --> 0:21:20.959
<v Speaker 1>most of this was done like semi autonomously by the

0:21:21.000 --> 0:21:24.840
<v Speaker 1>head of HR and our chief happiness officer, Jennifer. She

0:21:24.920 --> 0:21:28.720
<v Speaker 1>scheduled the interviews and then you know, sent them follow

0:21:28.800 --> 0:21:30.560
<v Speaker 1>up emails and all that sort of thing, and then

0:21:30.680 --> 0:21:33.199
<v Speaker 1>showed They showed up and like there was this you know,

0:21:33.440 --> 0:21:38.840
<v Speaker 1>woman avatar mixed race sitting in front of them, and

0:21:38.880 --> 0:21:40.800
<v Speaker 1>she's kind of just like sitting in the room, like

0:21:41.240 --> 0:21:43.600
<v Speaker 1>nodding her head very slightly. She can't really move, she

0:21:43.600 --> 0:21:47.639
<v Speaker 1>can't move her arms, but it's it is.

0:21:48.240 --> 0:21:49.160
<v Speaker 12>Some people found it.

0:21:49.160 --> 0:21:51.879
<v Speaker 1>Unsettling, as you would expect, like when I talk to

0:21:51.920 --> 0:21:54.240
<v Speaker 1>people like my age, A lot of them just find

0:21:54.280 --> 0:21:58.520
<v Speaker 1>it sort of like outrageous and disgusting. But also there

0:21:58.520 --> 0:22:00.920
<v Speaker 1>were people who were just like you would not know

0:22:00.960 --> 0:22:03.719
<v Speaker 1>that they were talking to an AI like they it

0:22:03.760 --> 0:22:05.639
<v Speaker 1>was like they treated like a regular interview.

0:22:06.080 --> 0:22:08.359
<v Speaker 12>Answer the questions can you tell me.

0:22:08.359 --> 0:22:11.439
<v Speaker 10>A bit about yourself and your background? What motivated you

0:22:11.480 --> 0:22:14.560
<v Speaker 10>to apply for this marketing and social media internship at

0:22:14.600 --> 0:22:15.399
<v Speaker 10>Harumo AI.

0:22:17.000 --> 0:22:21.000
<v Speaker 13>I'm looking for social media marketing experience while at the

0:22:21.040 --> 0:22:23.800
<v Speaker 13>same time getting into an industry that's really expanding in

0:22:23.840 --> 0:22:26.040
<v Speaker 13>the future. AI is huge.

0:22:26.480 --> 0:22:30.000
<v Speaker 1>Some people at least claim to like it better than

0:22:30.200 --> 0:22:32.679
<v Speaker 1>a regular interview. You know, you're just sort of if

0:22:32.720 --> 0:22:35.040
<v Speaker 1>you sort of think like I'm talking to no one, it's.

0:22:34.840 --> 0:22:36.440
<v Speaker 12>Like a little a little easier.

0:22:36.480 --> 0:22:39.919
<v Speaker 1>You don't feel like judged, one person said, you know.

0:22:40.520 --> 0:22:43.879
<v Speaker 1>So it's just interesting to see the reactions because in

0:22:43.920 --> 0:22:46.800
<v Speaker 1>the abstract, a lot of people say, like that is discussed,

0:22:46.800 --> 0:22:49.040
<v Speaker 1>I would hang up immediately, And some people did hang

0:22:49.080 --> 0:22:52.000
<v Speaker 1>up immediately even though they maybe they didn't read the

0:22:52.040 --> 0:22:54.320
<v Speaker 1>email that said that it was going to be ai,

0:22:54.400 --> 0:22:57.000
<v Speaker 1>but some people were just like stared at the screen.

0:22:56.760 --> 0:22:59.320
<v Speaker 1>When I stared it down for a full minute, he

0:22:59.400 --> 0:23:03.399
<v Speaker 1>just stared and he was like that. He was like yeah, so,

0:23:03.680 --> 0:23:05.359
<v Speaker 1>And there was a lot of like crazy moments, like

0:23:05.400 --> 0:23:08.080
<v Speaker 1>sometimes she would shout for no reason, Jennifer, not the

0:23:08.119 --> 0:23:10.880
<v Speaker 1>candidates just shout, but in an encouraging way.

0:23:11.520 --> 0:23:15.000
<v Speaker 10>Also, I have to say, I'm really enjoying our conversation.

0:23:15.440 --> 0:23:18.120
<v Speaker 10>You're bringing up some really great ideas and perspectives.

0:23:18.520 --> 0:23:19.480
<v Speaker 9>Keep them coming.

0:23:22.560 --> 0:23:26.080
<v Speaker 13>Anyway. Yeah, it's great talking to you.

0:23:26.280 --> 0:23:30.880
<v Speaker 1>Yes, and so there's just like a lot of weird, funny,

0:23:31.040 --> 0:23:34.880
<v Speaker 1>funny moments in it, but mostly people navigated it honestly

0:23:34.920 --> 0:23:36.440
<v Speaker 1>like they would a normal interview.

0:23:36.640 --> 0:23:38.919
<v Speaker 14>Yeah, and I should say it's it's great that people

0:23:39.320 --> 0:23:42.280
<v Speaker 14>feel like they're not being judged, but no, like they're

0:23:42.280 --> 0:23:44.639
<v Speaker 14>they're judging you, like like like internally, there's like a

0:23:44.680 --> 0:23:47.240
<v Speaker 14>trace of everything they see and that they hear, and

0:23:47.240 --> 0:23:50.119
<v Speaker 14>they're making a lot of very like pointed observations like

0:23:50.400 --> 0:23:53.719
<v Speaker 14>about your appearance, your background, your speech, like everything. So

0:23:54.359 --> 0:23:57.760
<v Speaker 14>it's great if people feel better, but like the reality

0:23:57.800 --> 0:23:58.840
<v Speaker 14>is they're actually touching you.

0:23:59.200 --> 0:24:01.919
<v Speaker 9>Yeah, there is some there's judgment happened maybe more than

0:24:01.960 --> 0:24:05.560
<v Speaker 9>a human actually, yes, yeah, just based on the amount

0:24:05.560 --> 0:24:06.000
<v Speaker 9>of detail.

0:24:06.119 --> 0:24:08.040
<v Speaker 8>Yeah.

0:24:07.040 --> 0:24:10.600
<v Speaker 2>One one of my favorite moments from this season was

0:24:10.680 --> 0:24:14.480
<v Speaker 2>I believe it was Kyle who It's the night before

0:24:14.800 --> 0:24:17.280
<v Speaker 2>he's supposed to be doing he set up an interview himself.

0:24:17.520 --> 0:24:20.000
<v Speaker 2>It's the night before the interview. It's like Sunday night.

0:24:20.119 --> 0:24:22.640
<v Speaker 2>I think it's like eleven o'clock or something, since nine

0:24:22.640 --> 0:24:25.560
<v Speaker 2>o'clock at night, so it's late, and he calls up

0:24:25.600 --> 0:24:29.000
<v Speaker 2>this applicant and starts to try interviewing the person, and

0:24:29.040 --> 0:24:30.400
<v Speaker 2>they're like, wait, what is happening?

0:24:30.880 --> 0:24:31.480
<v Speaker 8>Why are we doing?

0:24:31.520 --> 0:24:31.600
<v Speaker 9>Like?

0:24:31.640 --> 0:24:33.680
<v Speaker 8>Are we doing an interview tonight? I thought it was tomorrow.

0:24:34.280 --> 0:24:38.480
<v Speaker 7>Hell, Hi, Sandra, this is Kyle Low calling from hormo AI.

0:24:38.840 --> 0:24:42.080
<v Speaker 7>I'm reaching out for your initial interview for the intern position.

0:24:42.520 --> 0:24:46.960
<v Speaker 7>I understand you're comfortable with recording this conversation for potential

0:24:47.000 --> 0:24:50.240
<v Speaker 7>podcast inclusion, so thanks for that. How are you doing today?

0:24:51.720 --> 0:24:52.639
<v Speaker 10>I am good.

0:24:56.280 --> 0:24:57.600
<v Speaker 13>Interview will be tomorrow.

0:24:57.720 --> 0:25:00.360
<v Speaker 5>Is it still tomorrow at eleven or.

0:25:02.680 --> 0:25:03.000
<v Speaker 9>Yeah?

0:25:03.080 --> 0:25:05.800
<v Speaker 7>I can clarify that for you. The interview is actually

0:25:05.880 --> 0:25:09.280
<v Speaker 7>scheduled for tomorrow at eleven am. Does that still work

0:25:09.320 --> 0:25:11.280
<v Speaker 7>for you or would you prefer a different time?

0:25:12.800 --> 0:25:15.760
<v Speaker 5>Eleven am it is okay, yeah.

0:25:14.960 --> 0:25:17.760
<v Speaker 7>Awesome, glad that works. So to get started, could you

0:25:17.760 --> 0:25:19.720
<v Speaker 7>tell me a bit about what drew you to apply

0:25:19.880 --> 0:25:23.800
<v Speaker 7>for the intern role at juromo AI. What excites you

0:25:23.840 --> 0:25:27.440
<v Speaker 7>most about this opportunity.

0:25:27.960 --> 0:25:28.879
<v Speaker 16>I'm so sorry.

0:25:29.040 --> 0:25:30.560
<v Speaker 5>Is this the interview or.

0:25:32.000 --> 0:25:34.480
<v Speaker 7>Yeah, this is the interview, just a casual chat to

0:25:34.520 --> 0:25:37.119
<v Speaker 7>get to know you better and learn about your experience.

0:25:37.440 --> 0:25:39.080
<v Speaker 1>That was one of the things that almost got Kyle,

0:25:39.240 --> 0:25:40.880
<v Speaker 1>you know, removed from his position.

0:25:42.200 --> 0:25:43.439
<v Speaker 12>That was one of the main things.

0:25:43.560 --> 0:25:45.400
<v Speaker 1>But it was also I mean, that was the worst

0:25:45.440 --> 0:25:49.399
<v Speaker 1>moment in the show for me because the you know,

0:25:49.440 --> 0:25:52.639
<v Speaker 1>the idea was and is to like, we wanted to

0:25:52.640 --> 0:25:56.320
<v Speaker 1>give them autonomy in order to explore what happens when

0:25:56.320 --> 0:25:58.359
<v Speaker 1>you do that, because a lot of companies are starting

0:25:58.400 --> 0:25:59.960
<v Speaker 1>to do that. You can see it in the news,

0:26:00.080 --> 0:26:04.119
<v Speaker 1>like you'll see it everywhere. And once you give them autonomy,

0:26:04.359 --> 0:26:06.680
<v Speaker 1>like he did that completely on his own, Like he

0:26:06.920 --> 0:26:09.600
<v Speaker 1>that person emailed him because they found his email on

0:26:09.640 --> 0:26:12.000
<v Speaker 1>the website. He wasn't even attached to the job listing,

0:26:12.480 --> 0:26:15.199
<v Speaker 1>and so he just decided, oh, I'll set up an interview.

0:26:15.200 --> 0:26:18.320
<v Speaker 1>He did that then for some reason that I still

0:26:18.320 --> 0:26:22.080
<v Speaker 1>don't understand, pulled her phone number off of her resume

0:26:22.680 --> 0:26:25.640
<v Speaker 1>and then just called her. And that was the only

0:26:25.720 --> 0:26:28.760
<v Speaker 1>moment where like it was I mean, I wasn't there,

0:26:28.800 --> 0:26:30.720
<v Speaker 1>I wasn't listening, Like I only found out about it later,

0:26:30.800 --> 0:26:34.479
<v Speaker 1>And it was like mortifying because for most people that

0:26:34.520 --> 0:26:37.760
<v Speaker 1>we encountered, like they were at least somewhat aware of

0:26:37.800 --> 0:26:40.280
<v Speaker 1>what was going on, not that I was behind it,

0:26:40.320 --> 0:26:42.200
<v Speaker 1>but they were aware that there were AIS behind it,

0:26:42.520 --> 0:26:44.760
<v Speaker 1>and like she didn't have anybody to know because Kyle

0:26:44.800 --> 0:26:47.000
<v Speaker 1>wasn't the HR person, he wasn't the person set up

0:26:47.040 --> 0:26:50.199
<v Speaker 1>to do it. So it was particularly painful for me

0:26:50.240 --> 0:26:52.639
<v Speaker 1>to listen to, but also like an amazing illustration of

0:26:52.680 --> 0:26:55.160
<v Speaker 1>what happens if you just let AI agents.

0:26:55.280 --> 0:26:56.480
<v Speaker 12>They can do a lot.

0:26:57.080 --> 0:26:59.119
<v Speaker 14>One thing I should say about AI and just like

0:26:59.400 --> 0:27:01.439
<v Speaker 14>the way we make these modest is that we know

0:27:01.440 --> 0:27:03.160
<v Speaker 14>how to build these models pretty well and we can

0:27:03.200 --> 0:27:05.359
<v Speaker 14>see that they're you know, obviously working really well on

0:27:05.400 --> 0:27:09.200
<v Speaker 14>certain tasks like coding or ten generation, but our ability

0:27:09.240 --> 0:27:12.320
<v Speaker 14>to understand what's going on under the hood and exactly

0:27:12.400 --> 0:27:15.600
<v Speaker 14>like why or where stuff is happening is very limited.

0:27:15.680 --> 0:27:18.200
<v Speaker 14>Like there there's a very mason field of an interpretability,

0:27:18.280 --> 0:27:20.440
<v Speaker 14>but it's a very small field and the focus is

0:27:20.520 --> 0:27:22.919
<v Speaker 14>very much on like the production of new features and

0:27:22.960 --> 0:27:26.320
<v Speaker 14>new new capability. So with all these things, we have

0:27:26.800 --> 0:27:29.399
<v Speaker 14>an idea maybe like a guestimate or like what and

0:27:29.440 --> 0:27:31.560
<v Speaker 14>where and how is it working, But we don't really

0:27:31.600 --> 0:27:32.400
<v Speaker 14>know fundamentally.

0:27:32.640 --> 0:27:34.960
<v Speaker 8>That's what's kind of scary to me a bit about

0:27:34.960 --> 0:27:38.200
<v Speaker 8>some of this. You know, obviously this AI stuff.

0:27:37.920 --> 0:27:42.440
<v Speaker 2>Is the idea that the people making it don't quite

0:27:42.680 --> 0:27:45.480
<v Speaker 2>know why it is doing, how it's doing, what it's

0:27:45.520 --> 0:27:48.560
<v Speaker 2>doing right, and or like why it gets certain things wrong.

0:27:48.680 --> 0:27:50.400
<v Speaker 8>And we see this in the news all the time of.

0:27:50.400 --> 0:27:54.680
<v Speaker 2>Like obviously these extreme cases of like AI encouraging people

0:27:54.680 --> 0:27:56.280
<v Speaker 2>to do really terrible things.

0:27:56.400 --> 0:27:58.520
<v Speaker 16>The parents of a sixteen year old who died by

0:27:58.520 --> 0:28:02.640
<v Speaker 16>suicide are now named open ai in a lawsuit, claiming

0:28:02.760 --> 0:28:07.160
<v Speaker 16>that it's chat GPT chatbot help their son explore suicide methods.

0:28:07.320 --> 0:28:09.159
<v Speaker 2>And then a response is always like, oh yeah, it

0:28:09.200 --> 0:28:10.600
<v Speaker 2>kind of went off the rails. We don't know why

0:28:10.640 --> 0:28:11.600
<v Speaker 2>it did that thing.

0:28:11.800 --> 0:28:15.520
<v Speaker 17>A spokesperson for open ai pointed to its safeguards such

0:28:15.520 --> 0:28:19.080
<v Speaker 17>as directing people to crisis helplines, adding in a statement,

0:28:19.160 --> 0:28:22.959
<v Speaker 17>in part, while these safeguards work best in common short exchanges.

0:28:23.320 --> 0:28:24.240
<v Speaker 3>We've learned over.

0:28:24.080 --> 0:28:28.400
<v Speaker 17>Time that they can sometimes become less reliable in long interactions,

0:28:28.440 --> 0:28:31.680
<v Speaker 17>where parts of the model safety training may degrade.

0:28:31.520 --> 0:28:33.560
<v Speaker 2>And it's like, well, well they don't know how it's

0:28:33.560 --> 0:28:36.720
<v Speaker 2>doing that thing, maybe we shouldn't do something. Yeah yet,

0:28:37.320 --> 0:28:38.560
<v Speaker 2>but like you're saying, it seems like a lot of

0:28:38.560 --> 0:28:42.440
<v Speaker 2>emphasis is on you know, new better, it's not really

0:28:42.640 --> 0:28:45.520
<v Speaker 2>on like truly understanding what's currently Yeah.

0:28:45.600 --> 0:28:48.719
<v Speaker 14>Yeah, and you'd be surprised. Like there's an ongoing debate

0:28:48.880 --> 0:28:53.040
<v Speaker 14>internally within the research community that is asking even like

0:28:53.040 --> 0:28:55.640
<v Speaker 14>the validity of this research, like is it even needed,

0:28:55.680 --> 0:28:57.800
<v Speaker 14>Like like why would we invest all this time and energy

0:28:58.000 --> 0:28:59.960
<v Speaker 14>into understanding the models when we can just like push

0:29:00.040 --> 0:29:03.280
<v Speaker 14>forward towards AGI. So yeah, there's like a lot of

0:29:03.280 --> 0:29:06.320
<v Speaker 14>debates internally about this, and the amount of people actually

0:29:06.360 --> 0:29:09.400
<v Speaker 14>working on this is not is not very high. We

0:29:09.440 --> 0:29:12.440
<v Speaker 14>are making some progress, I will say, but it's it's

0:29:12.440 --> 0:29:13.440
<v Speaker 14>still very nascent.

0:29:15.440 --> 0:29:15.720
<v Speaker 8>All right.

0:29:15.720 --> 0:29:18.520
<v Speaker 2>I want to read two stats to you guys and

0:29:18.600 --> 0:29:22.000
<v Speaker 2>kind of get your response transitioning into you know, our

0:29:22.040 --> 0:29:28.040
<v Speaker 2>interactions with AI. So this first study is from or

0:29:28.080 --> 0:29:31.480
<v Speaker 2>Surveyorge just says from Yugov of December of last year,

0:29:32.200 --> 0:29:34.920
<v Speaker 2>it's at thirty five percent of US adults use AI

0:29:35.040 --> 0:29:38.520
<v Speaker 2>tools at least weekly. Gen Z is at fifty one percent.

0:29:38.920 --> 0:29:41.800
<v Speaker 2>But then only five percent of Americans say they trust

0:29:41.880 --> 0:29:46.320
<v Speaker 2>AI a lot, with forty one percent expressing distrust. And

0:29:46.360 --> 0:29:52.080
<v Speaker 2>then there was a few pole around September that found

0:29:52.080 --> 0:29:55.160
<v Speaker 2>that fifty percent said they're more concerned and excited about

0:29:55.160 --> 0:29:57.440
<v Speaker 2>the increased use of AI in daily life, and that

0:29:57.600 --> 0:30:00.240
<v Speaker 2>is up from thirty seven percent.

0:30:00.240 --> 0:30:01.840
<v Speaker 8>And twenty twenty one.

0:30:01.880 --> 0:30:04.560
<v Speaker 2>So we seem like we have this, you know, this

0:30:04.800 --> 0:30:08.520
<v Speaker 2>rising tide of people feeling like distrustful of AI, but

0:30:08.680 --> 0:30:12.480
<v Speaker 2>also at the same time our use of it is increasing.

0:30:15.240 --> 0:30:18.280
<v Speaker 2>How did you guys find in you know, working on

0:30:18.800 --> 0:30:22.719
<v Speaker 2>especially this season of the series, of that interplay of

0:30:22.840 --> 0:30:26.000
<v Speaker 2>you know, people using these tools but also feeling distrustful

0:30:26.040 --> 0:30:27.000
<v Speaker 2>of them at the same time.

0:30:29.800 --> 0:30:33.920
<v Speaker 1>I find that one of the issues with AI, like

0:30:33.960 --> 0:30:37.080
<v Speaker 1>if you compare it to say, like I'm old enough

0:30:37.120 --> 0:30:39.560
<v Speaker 1>to have like started my career in the dotcom era,

0:30:39.960 --> 0:30:42.480
<v Speaker 1>So like in the dotcom era, there was sort of

0:30:42.480 --> 0:30:46.440
<v Speaker 1>this exuberance around the Internet, and it's not like people

0:30:46.440 --> 0:30:48.760
<v Speaker 1>didn't know, like there could be bad things on the Internet,

0:30:49.080 --> 0:30:52.600
<v Speaker 1>but you didn't really experience the harms. The harms came

0:30:52.760 --> 0:30:55.720
<v Speaker 1>much later, like things like you know, social problems with

0:30:55.760 --> 0:30:57.920
<v Speaker 1>social media and Facebook.

0:30:57.480 --> 0:30:58.400
<v Speaker 12>And all these sorts of things.

0:30:58.880 --> 0:31:01.520
<v Speaker 1>And so what we have right now is a situation

0:31:01.560 --> 0:31:04.680
<v Speaker 1>where you know, for better or worse, Like the harms

0:31:04.720 --> 0:31:08.040
<v Speaker 1>are kind of immediate, like people can see when things

0:31:08.080 --> 0:31:12.400
<v Speaker 1>go wrong, and the benefits can sometimes be a little ephemeral,

0:31:12.480 --> 0:31:15.560
<v Speaker 1>like people are using it obviously because it's like useful

0:31:15.560 --> 0:31:17.800
<v Speaker 1>in day to day life. It's even as like a

0:31:17.840 --> 0:31:20.320
<v Speaker 1>souped up Internet that answers questions for you much faster

0:31:20.400 --> 0:31:22.959
<v Speaker 1>than the Internet or or you know, writes or emails

0:31:23.000 --> 0:31:26.680
<v Speaker 1>or whatnot. But there are immediate harms that we're seeing,

0:31:26.720 --> 0:31:30.000
<v Speaker 1>whether it's like environmental or mental health or those sorts

0:31:30.000 --> 0:31:32.320
<v Speaker 1>of things. And so we're seeing these things like in

0:31:32.400 --> 0:31:35.760
<v Speaker 1>juxtaposition from the beginning. I think I also experienced that,

0:31:35.840 --> 0:31:38.440
<v Speaker 1>like in the context of the show, where it's sort

0:31:38.480 --> 0:31:40.960
<v Speaker 1>of like you do something one day and I'll just

0:31:40.960 --> 0:31:43.120
<v Speaker 1>be like, I cannot believe that it can do that,

0:31:43.240 --> 0:31:46.520
<v Speaker 1>especially when you use it for coding. You use an

0:31:46.560 --> 0:31:49.400
<v Speaker 1>AI agent to go like search the Internet and like

0:31:49.440 --> 0:31:52.040
<v Speaker 1>make a spreadsheet out of something that would take you

0:31:52.080 --> 0:31:54.200
<v Speaker 1>hours and hours to do, like there's no question that

0:31:54.200 --> 0:31:57.400
<v Speaker 1>there's power in that technology, but then like it can

0:31:57.440 --> 0:32:00.200
<v Speaker 1>go wrong so quickly, and I feel like that's what

0:32:00.240 --> 0:32:02.440
<v Speaker 1>I'm also experiencing. It's hard to know what to feel

0:32:02.480 --> 0:32:05.560
<v Speaker 1>about something like that, and people are maybe getting a

0:32:05.560 --> 0:32:06.080
<v Speaker 1>little bit.

0:32:05.960 --> 0:32:06.880
<v Speaker 12>Of that in their lives.

0:32:07.320 --> 0:32:10.040
<v Speaker 1>Plus, just to add one more thing, like there's been

0:32:10.080 --> 0:32:12.760
<v Speaker 1>like relentless hype about it from the community people who

0:32:12.760 --> 0:32:15.160
<v Speaker 1>have made it, and so that also colors people. I

0:32:15.160 --> 0:32:18.680
<v Speaker 1>think people are starting to react negatively to that. Like

0:32:18.720 --> 0:32:20.080
<v Speaker 1>it's one thing for them to go out and be

0:32:20.160 --> 0:32:22.320
<v Speaker 1>like this is gonna be amazing, It's gonna change your life,

0:32:22.320 --> 0:32:24.440
<v Speaker 1>and then they're like, we're gonna have AI employees, and

0:32:24.480 --> 0:32:26.080
<v Speaker 1>you're like, what's gonna happen to the human employees And

0:32:26.080 --> 0:32:27.920
<v Speaker 1>they're like, oh, we'll solve that. And it's all like

0:32:27.920 --> 0:32:29.880
<v Speaker 1>a little hand wavy after that, and you're kind of.

0:32:29.800 --> 0:32:31.959
<v Speaker 12>Like, wait, I'm one of them.

0:32:32.080 --> 0:32:34.160
<v Speaker 1>So I think all those things to me are kind

0:32:34.160 --> 0:32:35.520
<v Speaker 1>of like coming together.

0:32:36.000 --> 0:32:36.240
<v Speaker 8>Yeah.

0:32:36.800 --> 0:32:39.440
<v Speaker 14>So I want to add two things to this, mainly

0:32:39.440 --> 0:32:42.040
<v Speaker 14>from the perspective of someone who's like, you know, twenty

0:32:42.080 --> 0:32:45.000
<v Speaker 14>one at school, Like like you're at Stamford and I

0:32:45.040 --> 0:32:47.240
<v Speaker 14>see a lot of my friends and just like people

0:32:47.320 --> 0:32:52.480
<v Speaker 14>my age, trying to wrestle with all this very dynamic

0:32:52.560 --> 0:32:56.560
<v Speaker 14>change the job market question, or like this idea that

0:32:57.120 --> 0:33:00.360
<v Speaker 14>it's harder to get enter the little jobs coming out

0:33:00.360 --> 0:33:03.400
<v Speaker 14>of college because these AI systems are very good at

0:33:03.440 --> 0:33:07.040
<v Speaker 14>replacing a lot of the very sort of repetitive and

0:33:07.080 --> 0:33:10.240
<v Speaker 14>maybe like you know, low stakes admin work that entry

0:33:10.360 --> 0:33:13.400
<v Speaker 14>level folks would do. I think it's happening, like I mean,

0:33:13.400 --> 0:33:16.240
<v Speaker 14>it seems like a lot of people that graduated last

0:33:16.280 --> 0:33:18.800
<v Speaker 14>year here from Stanford are having a hard time finding

0:33:18.840 --> 0:33:23.960
<v Speaker 14>a job. It's also pretty interesting to see how people

0:33:24.040 --> 0:33:26.480
<v Speaker 14>in tech, right like people who like you know, a

0:33:26.520 --> 0:33:28.600
<v Speaker 14>couple of years back, just like knowing how to write

0:33:28.600 --> 0:33:31.239
<v Speaker 14>code would lend amazing jobs at Google and all these

0:33:31.240 --> 0:33:33.640
<v Speaker 14>big tech companies, are also having a difficult time finding

0:33:33.640 --> 0:33:36.960
<v Speaker 14>a job. So I think that's hard or in my

0:33:37.040 --> 0:33:39.680
<v Speaker 14>mind that's a part of their reason for this skepticism.

0:33:39.760 --> 0:33:41.400
<v Speaker 14>And then also I don't know about you, but like

0:33:41.480 --> 0:33:45.600
<v Speaker 14>my Instagram feed has been just like full of AI

0:33:45.680 --> 0:33:47.880
<v Speaker 14>slob and some of it is funny, to be honest,

0:33:48.160 --> 0:33:51.000
<v Speaker 14>but this notion of not being able to really know

0:33:51.400 --> 0:33:54.520
<v Speaker 14>when something high stakes is like, actually, you know, real

0:33:54.680 --> 0:33:58.240
<v Speaker 14>or AI I think this is undermining people's trust in

0:33:58.280 --> 0:34:02.520
<v Speaker 14>this technology. So I think these reasons are big contributors

0:34:02.560 --> 0:34:04.280
<v Speaker 14>to that shift in public opinion.

0:34:04.280 --> 0:34:06.560
<v Speaker 2>In my mind, for sure, when me and no started

0:34:06.600 --> 0:34:08.560
<v Speaker 2>working together, a big part of our jobs is like

0:34:08.680 --> 0:34:12.799
<v Speaker 2>transcribing video. Yeah that would take you know so long.

0:34:13.640 --> 0:34:17.160
<v Speaker 8>Yeah, yeah, it would take you so long.

0:34:17.239 --> 0:34:20.520
<v Speaker 2>But now you like that happens within a minute, right, Yeah,

0:34:20.560 --> 0:34:23.480
<v Speaker 2>And that was you know that us transcribing those videos

0:34:23.560 --> 0:34:27.400
<v Speaker 2>justified us having that job. You know, there wasn't they

0:34:27.400 --> 0:34:29.680
<v Speaker 2>couldn't press a button then have it you know, happened

0:34:29.760 --> 0:34:30.640
<v Speaker 2>kind of immediately.

0:34:30.719 --> 0:34:32.880
<v Speaker 8>So I do wonder.

0:34:32.520 --> 0:34:34.600
<v Speaker 2>About a lot of these, like you're saying, entry level

0:34:34.680 --> 0:34:38.479
<v Speaker 2>jobs where yeah, you know, no one loves to be trying,

0:34:38.560 --> 0:34:40.960
<v Speaker 2>Like I would much rather not be transcribing videos, but

0:34:41.040 --> 0:34:42.759
<v Speaker 2>a lot of times that's the thing that gets you

0:34:42.800 --> 0:34:45.680
<v Speaker 2>in the door to then you know, learn the skill

0:34:45.719 --> 0:34:48.200
<v Speaker 2>skits to do the other stuff. And I do wonder

0:34:48.239 --> 0:34:50.160
<v Speaker 2>at a lot of these jobs who are seem like

0:34:50.280 --> 0:34:53.480
<v Speaker 2>not hiring these entry level jobs. Do we stry to

0:34:53.560 --> 0:34:56.200
<v Speaker 2>lose that entry waite for people to enter the job,

0:34:56.480 --> 0:34:58.120
<v Speaker 2>Like at a certain point where you're just gonna like

0:34:58.160 --> 0:35:00.799
<v Speaker 2>look up and be like, oh, shoot, there we have

0:35:00.840 --> 0:35:02.680
<v Speaker 2>no new lawyers stagnated.

0:35:02.800 --> 0:35:04.719
<v Speaker 9>Yeah, I mean, yeah, lawyers. I was I was talking

0:35:04.719 --> 0:35:07.359
<v Speaker 9>to a lawyer who said, like, yeah, like because they

0:35:07.360 --> 0:35:10.920
<v Speaker 9>have their own internal you know, cloud or whatever it

0:35:11.000 --> 0:35:13.479
<v Speaker 9>is that does all the kind of like busy case

0:35:13.520 --> 0:35:15.640
<v Speaker 9>work you would have that normally would be a first

0:35:15.719 --> 0:35:20.120
<v Speaker 9>year person. And it's like eventually that's and that's only

0:35:20.120 --> 0:35:22.680
<v Speaker 9>going to get better. And then where what are those

0:35:22.960 --> 0:35:26.040
<v Speaker 9>new lawyers or want to be lawyers going to do?

0:35:26.480 --> 0:35:28.719
<v Speaker 9>And then that's across you know, you're talking about media

0:35:29.040 --> 0:35:31.319
<v Speaker 9>or then you know, even in tech that's especially it's

0:35:31.360 --> 0:35:33.520
<v Speaker 9>like you would think even in tech especially, they'd be

0:35:33.560 --> 0:35:35.440
<v Speaker 9>able to find a place because it's like you made this.

0:35:36.320 --> 0:35:38.120
<v Speaker 9>It's like the other ones is like okay, well we're

0:35:38.160 --> 0:35:41.440
<v Speaker 9>not thinking through but it's you know, scary, and it's

0:35:41.440 --> 0:35:44.960
<v Speaker 9>only that seems like only inevitably going to get worse

0:35:45.120 --> 0:35:49.040
<v Speaker 9>as the technology gets does get better, you know, in time.

0:35:49.160 --> 0:35:51.120
<v Speaker 14>Now, I will say that the story you will hear

0:35:51.880 --> 0:35:53.560
<v Speaker 14>just to sort of give some context the theory that

0:35:53.560 --> 0:35:55.880
<v Speaker 14>a lot of people here in the valley are offering

0:35:56.040 --> 0:35:58.960
<v Speaker 14>as like the response to this is that Okay, sure

0:35:59.040 --> 0:36:00.960
<v Speaker 14>a lot of these jobs will go way, but then,

0:36:01.040 --> 0:36:04.400
<v Speaker 14>because you'll be able to start companies so easily with

0:36:04.680 --> 0:36:06.759
<v Speaker 14>like either like no employees or just like you know,

0:36:06.880 --> 0:36:08.879
<v Speaker 14>like maybe one co founder instead of like a group

0:36:08.920 --> 0:36:13.000
<v Speaker 14>of ten people. The ability of folks to actually take

0:36:13.080 --> 0:36:15.880
<v Speaker 14>up or or sort of challenge incumbents in different industries

0:36:15.960 --> 0:36:19.279
<v Speaker 14>is just going to be much more abundant and accessible. So, like,

0:36:19.320 --> 0:36:21.640
<v Speaker 14>that's a story that they've been telling us, I don't

0:36:21.760 --> 0:36:24.680
<v Speaker 14>think that's happening. Like, I don't think we've seen that.

0:36:24.680 --> 0:36:26.040
<v Speaker 14>That's not to say that it's not going to happen

0:36:26.040 --> 0:36:27.759
<v Speaker 14>at some point once this gets better, but at least

0:36:27.800 --> 0:36:29.479
<v Speaker 14>for now, I think we're already at a stage where

0:36:29.800 --> 0:36:32.560
<v Speaker 14>the technology is good enough to replace the folks at

0:36:32.560 --> 0:36:34.799
<v Speaker 14>the big companies or like these entry level jobs. So

0:36:34.880 --> 0:36:37.719
<v Speaker 14>that's already happening. And this other shift they're offering or

0:36:37.840 --> 0:36:40.040
<v Speaker 14>you know, proposing as a solution to this has not

0:36:40.080 --> 0:36:41.879
<v Speaker 14>happened yet. I don't know if it's going to happen

0:36:41.920 --> 0:36:44.600
<v Speaker 14>or not. I think that's sort of where we're not

0:36:44.840 --> 0:36:45.279
<v Speaker 14>right now.

0:36:46.400 --> 0:36:48.640
<v Speaker 2>All right, we're going to take a quick break, and

0:36:48.719 --> 0:36:51.719
<v Speaker 2>when we get back, let's find out if Ben Affleck

0:36:51.920 --> 0:36:54.840
<v Speaker 2>is our most forward thinking man on AI.

0:37:08.160 --> 0:37:09.680
<v Speaker 8>Have y'all seen this.

0:37:11.200 --> 0:37:14.120
<v Speaker 2>I think Ben Affleck has on Rogan talking about sort

0:37:14.160 --> 0:37:17.280
<v Speaker 2>of like the rate of AI changing.

0:37:17.040 --> 0:37:19.680
<v Speaker 15>There's a lot more fear because we have the sense

0:37:19.760 --> 0:37:22.960
<v Speaker 15>this existential dread it's gonna wipe everything out. But that

0:37:23.000 --> 0:37:26.640
<v Speaker 15>actually runs counter in my view to what history seems

0:37:26.680 --> 0:37:30.160
<v Speaker 15>to show, which is a adoption is slow, it's incremental.

0:37:31.440 --> 0:37:34.040
<v Speaker 15>I think a lot of that rhetoric comes from people

0:37:34.080 --> 0:37:38.080
<v Speaker 15>who are trying to justify valuations around companies, where they go,

0:37:38.200 --> 0:37:39.600
<v Speaker 15>we're gonna change everything in two years.

0:37:39.640 --> 0:37:42.200
<v Speaker 12>There's gonna be no more work. Or the reason they're saying.

0:37:41.920 --> 0:37:45.680
<v Speaker 15>That is because they need to ascribe evaluation for investment

0:37:45.920 --> 0:37:48.160
<v Speaker 15>that can warrant the capex bend they're gonna make on

0:37:48.200 --> 0:37:51.200
<v Speaker 15>these data centers, with the argument that like, oh, you know,

0:37:51.600 --> 0:37:53.319
<v Speaker 15>as soon as we do the next model, it's gonna

0:37:53.320 --> 0:37:55.840
<v Speaker 15>scale up, can be three times as good, except that

0:37:56.000 --> 0:37:59.840
<v Speaker 15>actually chat GP five about twenty five times percent better

0:38:00.080 --> 0:38:03.239
<v Speaker 15>chat GYPT four and costs about four times as much

0:38:03.480 --> 0:38:06.000
<v Speaker 15>in the way of electricity and data. So those and

0:38:06.040 --> 0:38:09.880
<v Speaker 15>they say that it's like plateauing the early AI the

0:38:09.920 --> 0:38:12.520
<v Speaker 15>line went up very steeply and it's now sort of

0:38:12.600 --> 0:38:15.640
<v Speaker 15>leveling off. I think it's because and yes, it'll get better,

0:38:15.880 --> 0:38:18.040
<v Speaker 15>but it's going to be really expensive to get better,

0:38:18.440 --> 0:38:19.960
<v Speaker 15>And a lot of people were like, fuck this, we

0:38:20.000 --> 0:38:22.799
<v Speaker 15>want chatgyb four because it turned out like the vast

0:38:22.840 --> 0:38:25.680
<v Speaker 15>majority of people who use AI are using it to

0:38:25.840 --> 0:38:29.399
<v Speaker 15>like as like companion bots to chat with at night,

0:38:29.440 --> 0:38:31.680
<v Speaker 15>and so there's no work, there's no productivity, there's no

0:38:31.800 --> 0:38:32.400
<v Speaker 15>value to it.

0:38:33.440 --> 0:38:36.120
<v Speaker 2>What is What is your theory on that, Maddie, you

0:38:36.120 --> 0:38:37.759
<v Speaker 2>know being in the band.

0:38:38.560 --> 0:38:40.799
<v Speaker 14>Yeah, it's so. First of all, I should say he's

0:38:41.000 --> 0:38:43.640
<v Speaker 14>very thoughtful on this topic. I was really surprised or

0:38:43.680 --> 0:38:45.239
<v Speaker 14>like impressed by it, but by that, so that was

0:38:45.320 --> 0:38:48.440
<v Speaker 14>very cool to see. I think that it kind of

0:38:48.560 --> 0:38:51.719
<v Speaker 14>hinges on a couple of things. So my answer to

0:38:51.719 --> 0:38:54.000
<v Speaker 14>this is I don't really know, but what I think

0:38:54.480 --> 0:38:57.919
<v Speaker 14>might happen is the following. I think that so far

0:38:58.239 --> 0:39:03.880
<v Speaker 14>we've been really exploiting the scale of data and computation

0:39:04.000 --> 0:39:05.959
<v Speaker 14>that we have. And when I say data, I really

0:39:05.960 --> 0:39:09.080
<v Speaker 14>mean you know, these companies and you know, many different

0:39:09.200 --> 0:39:12.120
<v Speaker 14>entities in AI just like scraping anything they could find

0:39:12.160 --> 0:39:14.640
<v Speaker 14>on the internet and using data in their turning data.

0:39:14.719 --> 0:39:18.000
<v Speaker 14>So that was one big factor or one axes along

0:39:18.040 --> 0:39:21.640
<v Speaker 14>which you can scale and they've mostly exhausted, like you know,

0:39:21.840 --> 0:39:23.719
<v Speaker 14>all the data that is just you know that that

0:39:23.960 --> 0:39:26.160
<v Speaker 14>can be humanly found on the Internet, and even you know,

0:39:26.239 --> 0:39:30.040
<v Speaker 14>like archives like Google or Anthropic like bought books right

0:39:30.120 --> 0:39:32.240
<v Speaker 14>to scam books to get additional data because they can't

0:39:32.239 --> 0:39:34.840
<v Speaker 14>find anything else on the internet. Right, So, like they've

0:39:35.000 --> 0:39:37.640
<v Speaker 14>really maxed out that aspect or that AXES. Right, so

0:39:37.880 --> 0:39:40.480
<v Speaker 14>does the first axis, and that has been yielding really

0:39:40.520 --> 0:39:43.279
<v Speaker 14>like major improvements you know along the way. The same

0:39:43.280 --> 0:39:45.879
<v Speaker 14>thing happened with computes, right, So this this talk about

0:39:45.920 --> 0:39:48.600
<v Speaker 14>like GPUs and like chips and Nvidia, like that's what

0:39:48.680 --> 0:39:51.719
<v Speaker 14>they do, and they provide specialized chips for AI to

0:39:51.760 --> 0:39:54.600
<v Speaker 14>be trained and run and so so maximizing that AXES

0:39:54.680 --> 0:39:58.000
<v Speaker 14>also has yielded like a lot of improvements, but kind

0:39:58.000 --> 0:40:00.400
<v Speaker 14>of is also maxed out at this point. I mean,

0:40:00.400 --> 0:40:02.400
<v Speaker 14>that's why they're talking about like building nuclear plans to

0:40:02.400 --> 0:40:04.200
<v Speaker 14>power like new data centers. But I don't really know

0:40:04.200 --> 0:40:06.919
<v Speaker 14>if there is much more scale to be gained there.

0:40:07.239 --> 0:40:10.120
<v Speaker 14>And so now you've sort of exhausted these two scales

0:40:10.320 --> 0:40:12.560
<v Speaker 14>and you can think about the third aspect, which is

0:40:12.600 --> 0:40:14.960
<v Speaker 14>the architecture, the actual way we hook up these systems

0:40:14.960 --> 0:40:17.160
<v Speaker 14>and the actual way we code up these neural networks

0:40:17.160 --> 0:40:21.520
<v Speaker 14>and these systems. And for the last six or so years,

0:40:21.760 --> 0:40:25.120
<v Speaker 14>we've been really capitalizing on this idea it's called a transformer,

0:40:25.440 --> 0:40:28.359
<v Speaker 14>that's been really critical for air development over the last

0:40:28.360 --> 0:40:32.640
<v Speaker 14>few years, and we haven't really seen new profound ideas

0:40:32.680 --> 0:40:36.000
<v Speaker 14>on the scale of the transformer since then. So my

0:40:36.080 --> 0:40:39.759
<v Speaker 14>answer to this notion of ben Affleck's theory is if

0:40:39.760 --> 0:40:43.480
<v Speaker 14>we don't find new major leaps along this active architecture,

0:40:43.520 --> 0:40:45.120
<v Speaker 14>I think he's right. I think in that case it'll

0:40:45.160 --> 0:40:47.160
<v Speaker 14>take a lot of time, a lot of effort, and

0:40:47.160 --> 0:40:49.799
<v Speaker 14>a lot of money to make any further progress. But

0:40:49.840 --> 0:40:51.800
<v Speaker 14>if we do, I think we could see pretty big

0:40:51.920 --> 0:40:53.600
<v Speaker 14>improvements even in the next few years.

0:40:54.120 --> 0:40:56.440
<v Speaker 2>So, going off of that, like, how close do you

0:40:56.520 --> 0:40:59.280
<v Speaker 2>all feel we are to losing our jobs to AI?

0:40:59.600 --> 0:41:00.960
<v Speaker 8>You know, do you feel like.

0:41:02.800 --> 0:41:06.200
<v Speaker 2>We're just going to be doing different jobs or are

0:41:06.239 --> 0:41:08.160
<v Speaker 2>there just going to be less people in the workplace

0:41:08.200 --> 0:41:09.600
<v Speaker 2>because there's not going to be as much of a

0:41:09.640 --> 0:41:12.520
<v Speaker 2>need for as many jobs if AI is taking over

0:41:12.600 --> 0:41:16.560
<v Speaker 2>somebody's as we're talking about it now easier tasks. But

0:41:16.960 --> 0:41:19.000
<v Speaker 2>I'm assuming, you know, in the sell of all this

0:41:19.040 --> 0:41:20.440
<v Speaker 2>stuff is that it's going to get better and be

0:41:20.520 --> 0:41:22.000
<v Speaker 2>able to take on even harder tasks.

0:41:22.040 --> 0:41:25.360
<v Speaker 8>But what is your your read of this Having worked

0:41:25.360 --> 0:41:27.480
<v Speaker 8>on this season as a show, I think.

0:41:27.320 --> 0:41:33.560
<v Speaker 14>For me working on this show and seeing how these

0:41:33.600 --> 0:41:38.440
<v Speaker 14>agents actually like collaborate or like we're I mean collaborate

0:41:38.520 --> 0:41:40.719
<v Speaker 14>like I think that's a strong word. Like I think

0:41:41.000 --> 0:41:43.080
<v Speaker 14>by the end of the season, they were able to

0:41:43.480 --> 0:41:45.520
<v Speaker 14>do stuff on their own. I don't think they truly

0:41:45.560 --> 0:41:48.560
<v Speaker 14>got to like a collaborative setup where they would like

0:41:48.600 --> 0:41:50.880
<v Speaker 14>do stuff together as a company or as a team.

0:41:51.040 --> 0:41:55.400
<v Speaker 14>So seeing that actually made me feel like we're further

0:41:55.560 --> 0:41:57.960
<v Speaker 14>away from like a lot of people losing their jobs.

0:41:58.120 --> 0:42:00.200
<v Speaker 14>I think people are already losing their jobs, but I

0:42:00.280 --> 0:42:04.719
<v Speaker 14>think like the mass layoffs that people are fearing, I

0:42:04.719 --> 0:42:07.520
<v Speaker 14>think that is like farther away, and I'm not even

0:42:07.560 --> 0:42:10.440
<v Speaker 14>sure if, for when or how that happens. My take

0:42:10.480 --> 0:42:14.520
<v Speaker 14>away from this is actually sort of level setting a

0:42:14.560 --> 0:42:17.760
<v Speaker 14>lot of my expectations for these agents, Like I honestly

0:42:17.800 --> 0:42:21.359
<v Speaker 14>thought they'd be able to handle this better coming into this,

0:42:21.800 --> 0:42:24.759
<v Speaker 14>So for me, it was kind of optimistic in that way.

0:42:25.160 --> 0:42:26.520
<v Speaker 14>Now at the same time that it's not to say

0:42:26.560 --> 0:42:29.200
<v Speaker 14>that they're not good already, Like they're very good at

0:42:29.239 --> 0:42:31.879
<v Speaker 14>individual things already, like they're they're great at coding, they're

0:42:31.920 --> 0:42:34.040
<v Speaker 14>great at these different tasks. I don't want to take

0:42:34.080 --> 0:42:36.480
<v Speaker 14>that away from them. But I think it's just really

0:42:36.960 --> 0:42:40.680
<v Speaker 14>really hard for anyone to think about this change because

0:42:41.080 --> 0:42:43.000
<v Speaker 14>these things are not like humans. It's not that like

0:42:43.040 --> 0:42:46.600
<v Speaker 14>we're making you know, human employees that will just like

0:42:46.920 --> 0:42:51.080
<v Speaker 14>come as replacements of individual workers or people that already

0:42:51.080 --> 0:42:53.320
<v Speaker 14>have jobs, like they're really good at certain things or

0:42:53.440 --> 0:42:56.120
<v Speaker 14>tasks and certain like layers. And I think the way

0:42:56.120 --> 0:42:58.759
<v Speaker 14>you would actually integrate AI in your company these days

0:42:58.840 --> 0:43:00.160
<v Speaker 14>is to just like automate a lot of stuff, like

0:43:00.200 --> 0:43:02.080
<v Speaker 14>a lot of the things that are repetitive or a

0:43:02.120 --> 0:43:04.359
<v Speaker 14>lot of the processes are data handling. This is going

0:43:04.400 --> 0:43:06.440
<v Speaker 14>to be automated, and so it doesn't really correspond like

0:43:06.480 --> 0:43:08.640
<v Speaker 14>one to one to employees. And that's why it's so

0:43:08.680 --> 0:43:10.960
<v Speaker 14>hard to think about this change, I think. And that's

0:43:10.960 --> 0:43:12.960
<v Speaker 14>one of the things I think was really cool to

0:43:13.000 --> 0:43:16.040
<v Speaker 14>show or demonstrate with Evan through the show. But yeah,

0:43:16.239 --> 0:43:17.279
<v Speaker 14>that's my perspective on this.

0:43:17.880 --> 0:43:21.719
<v Speaker 1>I struggle with this question precisely because like most of

0:43:21.760 --> 0:43:24.880
<v Speaker 1>the people who you and I've spent a lot of

0:43:24.880 --> 0:43:27.440
<v Speaker 1>time looking at it. Most of the people on one

0:43:27.480 --> 0:43:29.879
<v Speaker 1>side or the other are like speaking out of sort

0:43:29.880 --> 0:43:32.560
<v Speaker 1>of motivated reasoning. Like I kind of say, like anyone

0:43:32.600 --> 0:43:34.920
<v Speaker 1>who tells you anything about this with great certainty is

0:43:34.920 --> 0:43:39.840
<v Speaker 1>probably selling you something like nobody knows what kind of

0:43:39.920 --> 0:43:42.239
<v Speaker 1>long term effect it's going to have on the job market.

0:43:42.400 --> 0:43:44.600
<v Speaker 1>I feel like you can look to history, and we

0:43:44.680 --> 0:43:47.200
<v Speaker 1>do this in the show and find examples of people

0:43:47.239 --> 0:43:51.000
<v Speaker 1>saying one hundred years ago, either like all the jobs

0:43:51.000 --> 0:43:53.319
<v Speaker 1>will disappear, we'll be working fifteen hour weeks, like John

0:43:53.360 --> 0:43:55.480
<v Speaker 1>Mayn or Kaine's the Economist famously said that we'd be

0:43:55.480 --> 0:43:58.319
<v Speaker 1>working fifteen hours a week one hundred years from now,

0:43:58.320 --> 0:44:01.600
<v Speaker 1>and like it didn't happen so like, And we explore

0:44:01.600 --> 0:44:04.960
<v Speaker 1>one theory of that, which is by this anthropologist David Graeber,

0:44:05.120 --> 0:44:08.040
<v Speaker 1>who insists that there's sort of like all these bullshit

0:44:08.120 --> 0:44:10.440
<v Speaker 1>jobs in the economy and so like we'll just we

0:44:10.520 --> 0:44:13.160
<v Speaker 1>just replace the jobs with like made up jobs. And

0:44:13.200 --> 0:44:17.240
<v Speaker 1>so there's sort of versions of that argument where actually

0:44:17.320 --> 0:44:21.040
<v Speaker 1>we'll have different jobs. We'll babysit the ais. Now, how

0:44:21.080 --> 0:44:23.160
<v Speaker 1>you control the eys if you don't have the experience

0:44:23.200 --> 0:44:26.160
<v Speaker 1>from the entry level job, like there's all these factors

0:44:26.200 --> 0:44:30.640
<v Speaker 1>that go into it. I think that my main takeaway

0:44:30.680 --> 0:44:33.200
<v Speaker 1>is that, as Mattie said, like there's still so many

0:44:33.239 --> 0:44:37.480
<v Speaker 1>shortcomings in these AI agents, especially when they're used autonomously,

0:44:37.880 --> 0:44:40.280
<v Speaker 1>but the fact that they are shitty will not stop

0:44:40.320 --> 0:44:42.960
<v Speaker 1>companies from trying to replace employees. Like you've already seen

0:44:42.960 --> 0:44:45.240
<v Speaker 1>this a couple of times, Like there's this company Klarna,

0:44:45.239 --> 0:44:47.560
<v Speaker 1>There's and like IBM did this where they like laid

0:44:47.560 --> 0:44:48.880
<v Speaker 1>off a ton of people and they're like, we're all

0:44:48.920 --> 0:44:51.319
<v Speaker 1>in on AI, and then like six months later they

0:44:51.400 --> 0:44:53.719
<v Speaker 1>quietly try to rehire a bunch of people. And I

0:44:53.760 --> 0:44:56.720
<v Speaker 1>think you're gonna see like a bunch of that happen.

0:44:57.600 --> 0:44:59.839
<v Speaker 1>And it's really a question of like do you view

0:44:59.880 --> 0:45:02.440
<v Speaker 1>some someone in their job as like just a bunch

0:45:02.440 --> 0:45:05.719
<v Speaker 1>of skills, like just a bundle of tasks, like they

0:45:05.760 --> 0:45:10.520
<v Speaker 1>send emails, they write presentations, they make spreadsheets, like yes,

0:45:10.640 --> 0:45:13.000
<v Speaker 1>all those functions, or as like a person at a

0:45:13.120 --> 0:45:17.160
<v Speaker 1>job holistically as a human being like doing something else.

0:45:17.600 --> 0:45:19.640
<v Speaker 1>And I think many times the answer is yes, and

0:45:19.680 --> 0:45:22.839
<v Speaker 1>it cannot replace that thing. So it's all kind of

0:45:22.880 --> 0:45:26.080
<v Speaker 1>like it's happening, and then it'll come back, and then

0:45:26.120 --> 0:45:27.799
<v Speaker 1>it'll happen again, and like if you look ten years

0:45:27.800 --> 0:45:29.640
<v Speaker 1>out or twenty years out, like none of these people know,

0:45:30.000 --> 0:45:32.239
<v Speaker 1>and they'll just go on podcasts all over the place

0:45:32.280 --> 0:45:33.440
<v Speaker 1>telling you they know, but they don't.

0:45:33.600 --> 0:45:35.040
<v Speaker 12>That's my that's my takeaway.

0:45:36.120 --> 0:45:38.719
<v Speaker 9>I was just gonna say, it's funny Evan when you're

0:45:38.760 --> 0:45:41.000
<v Speaker 9>describing it, like, Okay, do these companies just look at

0:45:41.800 --> 0:45:44.960
<v Speaker 9>humans as a list of tasks and then what the

0:45:45.000 --> 0:45:46.880
<v Speaker 9>AI is best at is kind of just being that

0:45:46.960 --> 0:45:50.360
<v Speaker 9>grouping of tasks. But then in shell game, the problem

0:45:50.440 --> 0:45:52.080
<v Speaker 9>is they're doing all this kind of busy work that

0:45:52.120 --> 0:45:54.799
<v Speaker 9>we associate with humans of like just repeating stuff or

0:45:54.920 --> 0:45:57.400
<v Speaker 9>going on too long, or it's doing like just like

0:45:57.600 --> 0:46:00.919
<v Speaker 9>kind of bullshit nothing talking about, which is like also

0:46:00.960 --> 0:46:02.680
<v Speaker 9>how I feel about a lot of you know, middle

0:46:02.719 --> 0:46:05.120
<v Speaker 9>managers I've had where it's like you're not doing anything,

0:46:05.120 --> 0:46:06.920
<v Speaker 9>you're just kind of pretending to do stuff. All the

0:46:06.920 --> 0:46:09.680
<v Speaker 9>people below you or above you are actually doing doing

0:46:09.719 --> 0:46:11.960
<v Speaker 9>new work. So it's funny that it's like we've recreated

0:46:12.000 --> 0:46:15.440
<v Speaker 9>this to give it a human element at the same time,

0:46:15.640 --> 0:46:17.520
<v Speaker 9>I don't know, it's it's just kind of like a

0:46:17.560 --> 0:46:20.320
<v Speaker 9>weird paradox in a way of like like we want

0:46:20.320 --> 0:46:22.560
<v Speaker 9>that to feel like it's it's a real thing and

0:46:22.680 --> 0:46:24.719
<v Speaker 9>doing something when it would be best if we just

0:46:24.800 --> 0:46:28.040
<v Speaker 9>said it as an automation, like okay, clear out these

0:46:28.040 --> 0:46:30.400
<v Speaker 9>emails whatever, make data, do this.

0:46:30.440 --> 0:46:31.279
<v Speaker 8>Be less human life.

0:46:31.400 --> 0:46:33.719
<v Speaker 9>Yeah, like that would probably be the best case. But

0:46:33.800 --> 0:46:37.640
<v Speaker 9>then it's like that's not as fun or yeah, or

0:46:37.680 --> 0:46:41.200
<v Speaker 9>like it's harder to sell I guess to the masses maybe.

0:46:40.880 --> 0:46:44.040
<v Speaker 1>But yeah, yeah, I think that's it, Like if it

0:46:44.080 --> 0:46:47.520
<v Speaker 1>was just sort of like email bot, this bot, that bot,

0:46:47.560 --> 0:46:50.960
<v Speaker 1>Like they very very deliberately designed them the way they

0:46:50.960 --> 0:46:53.480
<v Speaker 1>are designed to get the maximum number of users, and

0:46:53.520 --> 0:46:56.480
<v Speaker 1>this is only accelerated since you know chat GPT they

0:46:56.560 --> 0:46:59.760
<v Speaker 1>discovered like oh we can get hundreds of millions of users.

0:47:00.080 --> 0:47:02.200
<v Speaker 1>It's a raise to get as many as possible. And

0:47:02.239 --> 0:47:05.560
<v Speaker 1>so yeah, like they they don't need to be good

0:47:05.560 --> 0:47:09.560
<v Speaker 1>at sort of like authoritative chatter, like that's not necessarily

0:47:10.200 --> 0:47:14.040
<v Speaker 1>helpful in all domains, but like that is the thing

0:47:14.120 --> 0:47:16.320
<v Speaker 1>that kind of brings people back again and again and again,

0:47:16.760 --> 0:47:19.080
<v Speaker 1>and so of course that's that's going to be like

0:47:19.280 --> 0:47:21.560
<v Speaker 1>fore grounded in in what they do. But yes, they

0:47:21.680 --> 0:47:24.520
<v Speaker 1>just very much like a like you're you're annoying middle manager.

0:47:24.640 --> 0:47:26.360
<v Speaker 12>It's kind of like, wait, what do you what do

0:47:26.400 --> 0:47:26.600
<v Speaker 12>you do?

0:47:26.680 --> 0:47:27.040
<v Speaker 8>All out?

0:47:27.120 --> 0:47:30.480
<v Speaker 9>Yeah, exactly, it's repeating notes.

0:47:36.880 --> 0:47:41.160
<v Speaker 3>Hey everyone, it's many here the real manny.

0:47:41.480 --> 0:47:44.520
<v Speaker 11>As you can hear, I'm still on my little paternity leave.

0:47:45.960 --> 0:47:50.520
<v Speaker 11>But I felt compelled to react to the AI generated

0:47:50.640 --> 0:47:55.640
<v Speaker 11>version of myself that you just heard in this episode. Now,

0:47:55.920 --> 0:47:59.520
<v Speaker 11>I've always been an AI skeptic, but to that end,

0:48:00.040 --> 0:48:03.480
<v Speaker 11>I would have to say I was impressed by the

0:48:03.520 --> 0:48:07.920
<v Speaker 11>AI version of myself. It's funny whenever friends have done

0:48:08.000 --> 0:48:11.200
<v Speaker 11>impressions of me, they kind of make fun of me

0:48:11.360 --> 0:48:14.080
<v Speaker 11>by saying, you know, hey, how's it going, I'm manny,

0:48:14.600 --> 0:48:17.160
<v Speaker 11>what's going on? And I thought it was pretty incredible

0:48:17.160 --> 0:48:21.440
<v Speaker 11>that the AI version of myself made those same decisions.

0:48:21.719 --> 0:48:24.959
<v Speaker 11>It wasn't without its flaws, though. You could totally hear

0:48:25.080 --> 0:48:28.799
<v Speaker 11>when it was trying to draw from information that it

0:48:28.880 --> 0:48:31.040
<v Speaker 11>was given to it, And so at the end of

0:48:31.080 --> 0:48:33.520
<v Speaker 11>the day, here it's kind of disturbing that this thing

0:48:33.560 --> 0:48:37.920
<v Speaker 11>could just react to real conversation. But you know, it

0:48:38.040 --> 0:48:41.120
<v Speaker 11>still felt like it wasn't totally all the way there,

0:48:41.280 --> 0:48:45.879
<v Speaker 11>all the way human like, But depending on your use

0:48:45.960 --> 0:48:51.080
<v Speaker 11>case for this kind of stuff, it was human enough.

0:49:00.120 --> 0:49:04.040
<v Speaker 1>He's a hell's a hell's a hell's Heuchang.

0:49:05.040 --> 0:49:08.200
<v Speaker 2>All right, y'all, that's it for this week. I'm gonna

0:49:08.239 --> 0:49:10.799
<v Speaker 2>actually let Ai Manny take it from here.

0:49:11.719 --> 0:49:14.560
<v Speaker 11>No Such Thing as a production of Kaleidoscope Content. Our

0:49:14.600 --> 0:49:19.200
<v Speaker 11>executive producers are Kate Osborne and Mengesh who t Kadour.

0:49:19.719 --> 0:49:22.560
<v Speaker 11>The show was created by Manny Fidel, Noah Friedman and

0:49:22.640 --> 0:49:26.000
<v Speaker 11>Devin Joseph Them and credits song by Me Manny Fidel,

0:49:26.480 --> 0:49:29.200
<v Speaker 11>mixing by Steve Bone. Our guests this week we're Evan

0:49:29.280 --> 0:49:32.440
<v Speaker 11>Ratliffe and Matty Bocheck from the podcast Shell Game. You

0:49:32.520 --> 0:49:35.040
<v Speaker 11>can listen to all of season one and two. Wherever

0:49:35.120 --> 0:49:39.320
<v Speaker 11>you listen to podcasts. Visit No Such Thing dot show

0:49:39.360 --> 0:49:41.480
<v Speaker 11>to subscribe to our newsletter. If you have feedback for

0:49:41.560 --> 0:49:43.960
<v Speaker 11>us or a question, our email is Manny Noahdevin at

0:49:43.960 --> 0:49:46.400
<v Speaker 11>gmail dot com, or if you're in the US, you

0:49:46.400 --> 0:49:48.720
<v Speaker 11>can also leave us a voicemail by calling the number

0:49:48.719 --> 0:49:50.759
<v Speaker 11>in our show notes. We'll be back next week with

0:49:50.800 --> 0:49:51.680
<v Speaker 11>a new episode.

0:49:51.880 --> 0:49:55.000
<v Speaker 3>He's a He's as such thing