Latest / What's Your Problem? / The AI that Might Take My Job (And Yours)

Transcript

The AI that Might Take My Job (And Yours) · What's Your Problem?

• 0:15
  Speaker 1: Pushkin.
• 0:22
  Speaker 2: After every interview we do for the show, we upload
• 0:24
  Speaker 2: the audio to a piece of software called descript. Descript
• 0:28
  Speaker 2: turns the audio into a transcript, and then I can
• 0:31
  Speaker 2: edit that transcript, cut out the boring parts, move sections around,
• 0:35
  Speaker 2: and when I do that, descript edits the underlying audio
• 0:40
  Speaker 2: to match. As software, Descript is pretty janky, buggy. It's
• 0:45
  Speaker 2: constantly changing in ways that can make it hard to use,
• 0:48
  Speaker 2: and sometimes it just blows stuff up. But we use
• 0:51
  Speaker 2: it anyway because Descript is an incredible.
• 0:54
  Speaker 1: Advance over what came before.
• 0:57
  Speaker 2: Before descript audio software represented audio files not as words
• 1:01
  Speaker 2: that you can read and edit, but as waveforms, as
• 1:05
  Speaker 2: squiggly lines presented.
• 1:07
  Speaker 1: On a timeline. So when does the script came along?
• 1:09
  Speaker 2: Being able to edit audio by editing words on a
• 1:12
  Speaker 2: screen was this huge advance, and it was an advance
• 1:15
  Speaker 2: made possible by artificial intelligence. Eventually, Descript expanded to allow
• 1:21
  Speaker 2: people to edit not just audio but also video, and
• 1:25
  Speaker 2: last fall, open AI, the company that makes chat GPT,
• 1:29
  Speaker 2: led a fifty million dollar.
• 1:30
  Speaker 1: Investment round in Descript.
• 1:32
  Speaker 2: It's a sign that Descript is moving out to the
• 1:35
  Speaker 2: new AI frontier, the frontier of generative AI. AI that
• 1:40
  Speaker 2: creates words and pictures. This is of immediate interest to me,
• 1:44
  Speaker 2: as in is AI gonna help me do my job?
• 1:48
  Speaker 2: Is AI gonna do my job?
• 1:51
  Speaker 1: But there is also a bigger question here what is
• 1:54
  Speaker 1: AI going to mean?
• 1:55
  Speaker 2: More broadly for people whose jobs involve writing things and
• 1:58
  Speaker 2: creating visuals, which is to say, what is AI going
• 2:02
  Speaker 2: to mean for almost all white collar workers. I'm Jacob
• 2:11
  Speaker 2: Goldstein and this is What's Your Problem, a show about
• 2:13
  Speaker 2: people trying to make technological progress. My guest today is
• 2:17
  Speaker 2: Andrew Mason, founder and CEO of descript or maybe it's
• 2:21
  Speaker 2: descript by the way, I've always said descript and I'm
• 2:25
  Speaker 2: pretty sure that's wrong, right, it's a descript like dtour.
• 2:28
  Speaker 3: Weird noncommittal on the issue.
• 2:30
  Speaker 2: Let's do the subjective version.
• 2:31
  Speaker 1: You're just one man. How do you say the name
• 2:33
  Speaker 1: of your company?
• 2:34
  Speaker 3: Yeah, I've kind of cultivated the ability to flip between
• 2:37
  Speaker 3: them as I speak.
• 2:39
  Speaker 1: You're killing me.
• 2:40
  Speaker 3: The world still needs a little mystery.
• 2:42
  Speaker 1: Okay, how about this? Say your name and your job.
• 2:45
  Speaker 3: My name is Andrew Mason. I work at descript. That's
• 2:49
  Speaker 3: dscript dscript.
• 2:53
  Speaker 1: Well played.
• 2:55
  Speaker 2: Earlier in his career. Andrew Mason was the co founder
• 2:58
  Speaker 2: of groupunk. He took the company public and then got
• 3:01
  Speaker 2: fired after its stock fell by something like seventy five percent.
• 3:05
  Speaker 2: After that, he started a company called Dtour or maybe
• 3:08
  Speaker 2: it's I don't know. The company made these highly produced
• 3:12
  Speaker 2: audio walking tours that you could listen to on your phone.
• 3:15
  Speaker 2: In that job, Andrew saw the challenges of working with
• 3:18
  Speaker 2: the old waveform.
• 3:19
  Speaker 1: Based audio editing software.
• 3:21
  Speaker 2: At the same time, AI generated transcripts were getting better
• 3:25
  Speaker 2: and cheaper, and new technology was making it possible to
• 3:28
  Speaker 2: automatically match a transcript to an audio file. Andrew looked
• 3:32
  Speaker 2: at those two developments and thought, we should make an
• 3:35
  Speaker 2: audio editor that works like a word processor, which he
• 3:38
  Speaker 2: admits was a distraction from what he was supposed to
• 3:41
  Speaker 2: be doing, which was making walking tours.
• 3:44
  Speaker 3: If I'm being honest, it was a bit of an indulgence.
• 3:47
  Speaker 3: It just felt like an incredibly cool problem to work on.
• 3:51
  Speaker 3: I went to school for music technology and worked in
• 3:53
  Speaker 3: a recording studio after I graduated, and just always loved tools,
• 3:58
  Speaker 3: and audio visual tools in particular. It was just so
• 4:02
  Speaker 3: fun to start thinking about this puzzle.
• 4:05
  Speaker 1: Uh huh.
• 4:06
  Speaker 3: So we told ourselves it was kind of way of diversifying,
• 4:10
  Speaker 3: but that's just like a ridiculous way for a product
• 4:13
  Speaker 3: that at like a startup that just hasn't even found
• 4:16
  Speaker 3: product market fit in their core product to be thinking
• 4:19
  Speaker 3: about the world. You know, all of the advice textbooks
• 4:22
  Speaker 3: will tell you not to do that, and it's probably
• 4:24
  Speaker 3: generally good advice, but it was just it was just irresistible.
• 4:28
  Speaker 1: You know.
• 4:28
  Speaker 2: So I am a fan of descript. I started using
• 4:31
  Speaker 2: it around when it came out several years ago. Certainly
• 4:35
  Speaker 2: I think it's great. It is kind of janky, and
• 4:39
  Speaker 2: it's always kind of janky, right, And my guess is
• 4:43
  Speaker 2: jankie meaning like a little bit unstable, things don't quite work.
• 4:46
  Speaker 2: It's always telling you to restart.
• 4:48
  Speaker 3: By the way, if I'm not sure, if I have
• 4:50
  Speaker 3: your side, so I may ask you to send me
• 4:52
  Speaker 3: this entire portion of the interview so I can share
• 4:54
  Speaker 3: it with the team.
• 4:55
  Speaker 2: So the thing is, like, I wonder, why is it
• 4:58
  Speaker 2: always kind of janky? Why is it never just like
• 5:00
  Speaker 2: stable and it works? And my guess is it's because
• 5:02
  Speaker 2: you're pushing forward really fast, right, You're trying to make
• 5:05
  Speaker 2: it better and better and better, and presumably there is
• 5:08
  Speaker 2: some trace off, right, like the faster you try and
• 5:11
  Speaker 2: push it forward, the more janky it's gonna be. You could,
• 5:14
  Speaker 2: I'm sure, just perfect the way it was four years ago,
• 5:17
  Speaker 2: but then it would never get better, but it would
• 5:18
  Speaker 2: be stable, right, And so this is like a big
• 5:21
  Speaker 2: whatever startup founder type question, like, is that some balance
• 5:25
  Speaker 2: you're always trying to figure out how fast do we
• 5:27
  Speaker 2: iterate versus how much do we try and make it
• 5:29
  Speaker 2: just stable and work.
• 5:31
  Speaker 3: Yeah, that's an astute observation, not the fact that it's janky.
• 5:36
  Speaker 3: That doesn't take a genius.
• 5:38
  Speaker 2: Respectful but respectfully as a fan.
• 5:41
  Speaker 1: I'm telling you it's.
• 5:43
  Speaker 3: But I think like your your attempt to make sense
• 5:45
  Speaker 3: out of it, I think like a good story to
• 5:48
  Speaker 3: tell here is maybe like going back to the to
• 5:51
  Speaker 3: the very beginning of Descript. So when we became Descript,
• 5:56
  Speaker 3: we sold off Detour to Bows and we decided to
• 6:01
  Speaker 3: just focus on building out this media word processor thing.
• 6:06
  Speaker 3: And some of the public radio producers who had worked
• 6:10
  Speaker 3: at Detour went on back into public radio and they
• 6:14
  Speaker 3: became some of the earliest customers of dscript. And what
• 6:21
  Speaker 3: we found was that they pushed it so much farther
• 6:27
  Speaker 3: than we were ready for.
• 6:30
  Speaker 1: Ah, so quickly, what do you mean by that? Like,
• 6:32
  Speaker 1: what is an example of that.
• 6:34
  Speaker 3: Yeah, I mean specifically in the case of like some
• 6:37
  Speaker 3: of these shows, it means putting together three to five
• 6:40
  Speaker 3: hour cuts of tape from many different files, with lots
• 6:45
  Speaker 3: like tons and tons of edits and notes mixed into
• 6:49
  Speaker 3: the edits, and just like stuff that we hadn't pressure
• 6:54
  Speaker 3: tested from a performance just giant.
• 6:58
  Speaker 2: The files are really big, right, Like a three hour
• 7:01
  Speaker 2: audio file is actually a giant file, right, And if
• 7:04
  Speaker 2: you're stacking up a bunch of those, so you have
• 7:05
  Speaker 2: all these giant files and you're making tons of cuts,
• 7:08
  Speaker 2: that's just like computationally intensive, that kind of thing storage intensive.
• 7:15
  Speaker 3: Yeah, it was just something that we hadn't that we
• 7:17
  Speaker 3: hadn't optimized for. It's it's an eminently solvable problem, but
• 7:20
  Speaker 3: it was something that in the earliest versions we hadn't done.
• 7:26
  Speaker 3: And so that is kind of in many ways been
• 7:28
  Speaker 3: the story of descript up to this point, where there's
• 7:32
  Speaker 3: there's been that element of it, and there were kind
• 7:35
  Speaker 3: of realities of needing to make quick progress that we
• 7:39
  Speaker 3: had to balance against stability and what we had for
• 7:46
  Speaker 3: our customers in terms of like the core product idea
• 7:50
  Speaker 3: of being able to edit by text was still for
• 7:53
  Speaker 3: them so much better than the alternative that there was
• 7:59
  Speaker 3: just a tolerance of the stability issues that honestly made
• 8:04
  Speaker 3: us sick to our stomachs that we had to put
• 8:06
  Speaker 3: people through. And it's not like wegn but it was
• 8:10
  Speaker 3: like we had to make trade offs there. So all
• 8:13
  Speaker 3: of this pushing kind of culminated with this release of
• 8:16
  Speaker 3: a pretty major overhaul that we did at the end
• 8:20
  Speaker 3: of last year and since then, since last November and
• 8:26
  Speaker 3: really like through the first half of this year, is
• 8:30
  Speaker 3: when we think we start to get to a good place.
• 8:32
  Speaker 3: Our goal is that if we're having this conversation, like
• 8:35
  Speaker 3: we're not going to be having the same conversation in
• 8:38
  Speaker 3: say July, for sure at the very latest, like the
• 8:41
  Speaker 3: conversation we'll be having with someone like you will be Wow,
• 8:44
  Speaker 3: it's gotten. It's like not an issue anymore.
• 8:46
  Speaker 2: So you say all that, but also, you just got
• 8:50
  Speaker 2: this big investment from open Ai. You got a thing
• 8:54
  Speaker 2: on descript that says sign up to try GPT four
• 8:57
  Speaker 2: with Descript, which I just signed up for and I'm
• 9:00
  Speaker 2: very curious about That doesn't sound like, oh, we've arrived
• 9:04
  Speaker 2: and now we've got our product and we've just got
• 9:06
  Speaker 2: a hone it that sounds like there's this whole giant
• 9:08
  Speaker 2: new universe of things you were about to try and
• 9:10
  Speaker 2: figure out.
• 9:11
  Speaker 3: That's true, And that's the funny thing about all of
• 9:13
  Speaker 3: this are is that at the same time that we're
• 9:16
  Speaker 3: turning to focus on quality, it's a moment where generative
• 9:21
  Speaker 3: AI has arrived at a scale and with a force
• 9:25
  Speaker 3: that no one really saw coming this quickly.
• 9:28
  Speaker 2: So so okay, I know from the beginning Descript was
• 9:31
  Speaker 2: built on top of AI, you know, the technology for
• 9:35
  Speaker 2: transcription for matching audio to text, but was Descript itself
• 9:39
  Speaker 2: an AI company.
• 9:41
  Speaker 3: So we had some really smart people on the team
• 9:44
  Speaker 3: in UH with machine learning experience, but I wouldn't say
• 9:47
  Speaker 3: in the early days we were like a company that
• 9:50
  Speaker 3: was with anybody that was doing like original AI research
• 9:54
  Speaker 3: or anything like that. We saw that as a gap
• 9:59
  Speaker 3: that we wanted to solve. And so I forget exactly
• 10:05
  Speaker 3: what year it was, it was maybe about four years
• 10:07
  Speaker 3: ago we saw this company called Liarbird. It was a
• 10:12
  Speaker 3: a company out of y Combinator with some really smart
• 10:17
  Speaker 3: PhD candidates. They had built model that would build a
• 10:21
  Speaker 3: clone of your voice based on I think about three
• 10:24
  Speaker 3: minutes or five minutes of training data. Of just talking
• 10:27
  Speaker 3: to it.
• 10:27
  Speaker 2: Let me just say, I know Liarbird is spelled l yri,
• 10:32
  Speaker 2: but I assume they're aware.
• 10:34
  Speaker 1: Of the hominem.
• 10:35
  Speaker 2: Right, this is a thing that is cloning your voice
• 10:37
  Speaker 2: so that you can make it sound like you're talking
• 10:39
  Speaker 2: even if you're not talking. And the company is called Liarbird,
• 10:43
  Speaker 2: and this is a somewhat fraud thing, right, Like, I
• 10:47
  Speaker 2: feel like they're throwing it in my face that this
• 10:49
  Speaker 2: is a sketchy product that they're developing.
• 10:54
  Speaker 1: Did it cross your mind?
• 10:56
  Speaker 3: Did it cross my mind? Is like the ethical quandary
• 10:59
  Speaker 3: that we were getting into, or like the branding implications
• 11:01
  Speaker 3: of the name.
• 11:03
  Speaker 1: More the ethical quandary.
• 11:05
  Speaker 3: Yeah, the ethical quandary absolutely entered our mind. And our
• 11:10
  Speaker 3: point of view on that, and has been our point
• 11:12
  Speaker 3: of view on these things in general, has been that
• 11:17
  Speaker 3: we don't want to be like out there paving the
• 11:22
  Speaker 3: way for any new paths to the apocalypse, so to speak.
• 11:27
  Speaker 3: We actually think, like have always felt like not really
• 11:32
  Speaker 3: sure how society was going to put the brakes on
• 11:36
  Speaker 3: this sort of thing. We just knew that we didn't
• 11:38
  Speaker 3: want to be part of it, and we tried to
• 11:40
  Speaker 3: put guardrails in place on our product. That would make
• 11:43
  Speaker 3: it easy to stay off the slippery slope. So in
• 11:46
  Speaker 3: the case of Lyyerbird, which once we bought them, we
• 11:51
  Speaker 3: integrated their technology and released it as something that we
• 11:53
  Speaker 3: call overdub. It's a way that you can clone your voice.
• 11:56
  Speaker 3: We require you to authenticate that it's actually you, and
• 11:59
  Speaker 3: we only let you clone your own voice, and that's
• 12:02
  Speaker 3: worked really well. We're now in a world where there's
• 12:05
  Speaker 3: other people that have similar models and they're not putting
• 12:08
  Speaker 3: those protections in place. And the use case that we've
• 12:11
  Speaker 3: always been the most excited about is making it possible
• 12:14
  Speaker 3: to edit your natural recordings, so going in and changing
• 12:18
  Speaker 3: an individual word, and we've built some special stuff that
• 12:22
  Speaker 3: will kind of listen to the audio on either sides
• 12:24
  Speaker 3: and make sure that it blends in. From an intonation perspective,
• 12:27
  Speaker 3: we started with the ability to delete stuff and move
• 12:30
  Speaker 3: stuff around. Now you can just type and really make
• 12:32
  Speaker 3: it feel like it's a word processor.
• 12:34
  Speaker 2: Presumably the better you get, the better the technology you
• 12:39
  Speaker 2: use to Clona Voice gets, the more words it can do. Right,
• 12:43
  Speaker 2: I mean, every week, for what's your problem? I write
• 12:46
  Speaker 2: a little introduction and then I read it. But presumably
• 12:50
  Speaker 2: at some point overdub will be good enough that no
• 12:53
  Speaker 2: one knows will know whether it's me reading it or
• 12:55
  Speaker 2: I'm just typing it right.
• 12:57
  Speaker 3: We have a new version of overdub that will release
• 13:01
  Speaker 3: in the next couple of months, and it's the first
• 13:04
  Speaker 3: time that I've heard my own voice doing a narration
• 13:08
  Speaker 3: of something that made me say, like, this sounds so
• 13:12
  Speaker 3: much like me in a way that it's not distracting
• 13:16
  Speaker 3: or the AI does not get in the way.
• 13:18
  Speaker 2: Can I try that new version now, like, not this minute,
• 13:23
  Speaker 2: but like for the show?
• 13:24
  Speaker 1: Yeah, for the show.
• 13:26
  Speaker 3: I bet we could find a way to do it.
• 13:27
  Speaker 3: It's just so you could hear it and stuff.
• 13:30
  Speaker 2: There's a universe where I say, at this moment in
• 13:33
  Speaker 2: the show, guess what today? That voice me reading the
• 13:37
  Speaker 2: intro at the top of the show that was overdubbed.
• 13:39
  Speaker 1: It wasn't really made.
• 13:40
  Speaker 3: Yeah, we tried overdubb for the voice doing the intro
• 13:45
  Speaker 3: at the top of the show.
• 13:47
  Speaker 1: And we decided it wasn't quite.
• 13:49
  Speaker 2: Good enough, but we decided it would work for this
• 13:51
  Speaker 2: part of the show.
• 13:53
  Speaker 1: What you're hearing right now, it's not really me. It's overdubbed.
• 13:58
  Speaker 2: In a minute, what overdubb and chat GPT and generative
• 14:01
  Speaker 2: AI will mean for descript and for the.
• 14:04
  Speaker 1: World and also for me.
• 14:12
  Speaker 2: Now back to the show, descript is expanding from podcasts
• 14:16
  Speaker 2: to video, and it just took a big investment from
• 14:18
  Speaker 2: open Ai, the company that makes chat GPT, and also
• 14:22
  Speaker 2: this system called Dolly that uses AI to generate images.
• 14:26
  Speaker 2: So Descript is clearly pointing toward a future where it's
• 14:30
  Speaker 2: going to be software for creating AI generated or at
• 14:33
  Speaker 2: least AI.
• 14:34
  Speaker 1: Enhanced audio and video.
• 14:36
  Speaker 2: And I asked Andrew, what does that future look like?
• 14:39
  Speaker 2: How is generative AI going to work in descript?
• 14:43
  Speaker 3: I don't think we know entirely yet. In a lot
• 14:45
  Speaker 3: of ways, it feels to me like you're letting this
• 14:48
  Speaker 3: alien into into your app. You're just giving it the
• 14:51
  Speaker 3: keys and then the interfaces. How do you find how
• 14:56
  Speaker 3: do you find a way to kind of give the
• 15:00
  Speaker 3: aliens some buttons in tier UI, give them the ability
• 15:04
  Speaker 3: to press the buttons, and then how do you talk
• 15:06
  Speaker 3: to the alien?
• 15:07
  Speaker 1: What do you mean? Like?
• 15:08
  Speaker 2: That is a striking metaphor a little scarier right. It
• 15:13
  Speaker 2: suggests a certain level of uncertainty and potential downside. It's
• 15:17
  Speaker 2: not like, oh, this is great, this is going to
• 15:18
  Speaker 2: solve a problem like, why do you say it's like
• 15:20
  Speaker 2: letting an alien.
• 15:21
  Speaker 3: In as opposed to letting a human in.
• 15:25
  Speaker 1: It's a really interesting choice of words. Tell me more
• 15:27
  Speaker 1: about it.
• 15:30
  Speaker 3: So let's start by just saying, like, very specifically, what
• 15:35
  Speaker 3: I mean. I think, when implemented, well, what this will
• 15:40
  Speaker 3: feel like is as if you had a co editor
• 15:44
  Speaker 3: in a document with you, in our case, in a
• 15:47
  Speaker 3: video or a podcast that you're working on that is smart,
• 15:52
  Speaker 3: knows how to do everything, definitely knows how to do
• 15:54
  Speaker 3: the tedios busy work, and you can kind of kind
• 15:59
  Speaker 3: of guide or direct through giving these tasks. You know,
• 16:03
  Speaker 3: it's almost like it's the production assistant or something like that,
• 16:07
  Speaker 3: and you're the director and you're able to just guide
• 16:10
  Speaker 3: it and give it feedback on how it's doing and
• 16:13
  Speaker 3: what it's doing well and what it's not doing well.
• 16:15
  Speaker 2: There's a version of it where it's like we've gotten
• 16:18
  Speaker 2: used to the graphical user interface, right, We've been trained
• 16:21
  Speaker 2: since the Magintosh computer in the mid nineteen eighties that
• 16:24
  Speaker 2: the way you interact with a computer is like there's
• 16:26
  Speaker 2: little pictures and little folders and you point.
• 16:28
  Speaker 1: And click one way or another, right, and.
• 16:31
  Speaker 2: One possibility here is the new standard interface is chat.
• 16:35
  Speaker 2: You just type in like whatever, please trim all the
• 16:39
  Speaker 2: ums from this file, or even please turn this thirty
• 16:42
  Speaker 2: minute interview into a twenty minute interview, and the way
• 16:44
  Speaker 2: that makes it most interesting, right, and you just type
• 16:47
  Speaker 2: that in and it happens.
• 16:48
  Speaker 1: I mean that's a version of what I hear you
• 16:50
  Speaker 1: saying there.
• 16:50
  Speaker 3: I think some people believe that that chat or a
• 16:53
  Speaker 3: text field will become the primary interface for making things.
• 16:58
  Speaker 3: I think of it more as like it's the primary
• 17:01
  Speaker 3: interface for interacting with the alien, and then you and
• 17:04
  Speaker 3: the alien are still going to be working, like have
• 17:07
  Speaker 3: other buttons that they can press. You still, sometimes you
• 17:11
  Speaker 3: just want to take the thing in your hands and
• 17:13
  Speaker 3: do it yourself.
• 17:14
  Speaker 2: The alien metaphor, I mean there's a real like do
• 17:17
  Speaker 2: we welcome our alien overlord's question? When you choose that metaphor,
• 17:21
  Speaker 2: it makes.
• 17:21
  Speaker 3: Me I mean, maybe it feels that way.
• 17:24
  Speaker 2: It doesn't it it doesn't make me feel better.
• 17:26
  Speaker 3: I'll say that I think it feels the way that
• 17:30
  Speaker 3: an alien arrival would probably feel, where you know, maybe
• 17:33
  Speaker 3: you shake its hand and immediately it has something in
• 17:37
  Speaker 3: its skin that cures your cancer, and you feel hopeful,
• 17:42
  Speaker 3: but you also want to know what they're up.
• 17:45
  Speaker 2: To and yeah, and cure your cancer is definitely the
• 17:48
  Speaker 2: happy version.
• 17:49
  Speaker 1: Not usually in the alien movie.
• 17:51
  Speaker 2: What happens, but I guess that could happen.
• 17:53
  Speaker 3: Well, there's the good there's a good part, right, But
• 17:55
  Speaker 3: you never really know, I think is the point. And
• 17:57
  Speaker 3: I think we're all living in this kind of like
• 18:00
  Speaker 3: pushing forward in this mystery, kind of kind of stuck
• 18:03
  Speaker 3: between awe and terror.
• 18:06
  Speaker 2: You sound more ambivalent than I might have thought. Why
• 18:10
  Speaker 2: is that because you just took a giant investment from
• 18:12
  Speaker 2: open AI.
• 18:14
  Speaker 3: I think like at moments like this, you have a
• 18:17
  Speaker 3: choice between either renunciation and just like stopping and out
• 18:23
  Speaker 3: of from a place of fear. Which maybe that's right,
• 18:27
  Speaker 3: you know, maybe fulfillment and happiness everything we have for
• 18:32
  Speaker 3: that is is already here, and we should focus our
• 18:37
  Speaker 3: energies on making peace with our inevitable death.
• 18:41
  Speaker 1: In any case, we should do that.
• 18:42
  Speaker 3: But go on the other way to think of it
• 18:47
  Speaker 3: is to just forge ahead and realize that the potential
• 18:52
  Speaker 3: of what's on the other end of this might make
• 18:54
  Speaker 3: us feel in retrospect like we were just in the
• 18:59
  Speaker 3: earliest possible innings of our of the human experiment. So
• 19:06
  Speaker 3: you know, I feel like we're all going to die
• 19:09
  Speaker 3: one way or another, might as well forge ahead. It's
• 19:12
  Speaker 3: not ambivalence, but it's more just being clear eyed about
• 19:15
  Speaker 3: the fact that not trying to pretend that there's parts
• 19:18
  Speaker 3: of it that don't seem scary.
• 19:21
  Speaker 2: I mean, one of the things that's really striking to
• 19:24
  Speaker 2: me with AI, and that seems quite different from other
• 19:30
  Speaker 2: technologies in the past, is the people who are working
• 19:34
  Speaker 2: on it, the people who really understand it, seem more
• 19:38
  Speaker 2: scared than everybody else.
• 19:41
  Speaker 3: I'm not a first time founder. I went through the
• 19:44
  Speaker 3: experience of being a young person building building group on
• 19:51
  Speaker 3: telling myself a story about how it was going to
• 19:54
  Speaker 3: revolutionize local commerce and all the good stuff, and it
• 19:59
  Speaker 3: just didn't turn out that way. And I think we've
• 20:01
  Speaker 3: seen a generation of tech companies that just like didn't
• 20:06
  Speaker 3: turn out the way that the the super Rose colored
• 20:10
  Speaker 3: Glasses mission statement would have suggested. And I think we're
• 20:15
  Speaker 3: just trying to be we just have that experience, that
• 20:19
  Speaker 3: recent experience at top of mind, and are trying to
• 20:24
  Speaker 3: think about it in a way that has guardrails around
• 20:27
  Speaker 3: around repeating that history and just make sure we're really
• 20:30
  Speaker 3: proud of what we build. Does that make sense?
• 20:33
  Speaker 1: It makes sense.
• 20:34
  Speaker 3: Am I Am I going to regret saying all this?
• 20:37
  Speaker 1: I don't think so.
• 20:37
  Speaker 2: You haven't said anything like incriminating as far as I
• 20:40
  Speaker 2: can tell. You know, I heard somebody saying the other day, like,
• 20:43
  Speaker 2: it's an interesting question to ask somebody like, what was
• 20:46
  Speaker 2: the first thing you asked GPT chet GPT to do?
• 20:50
  Speaker 2: And the first thing I asked chet GPT to do
• 20:53
  Speaker 2: was write an episode of Planet Money podcast I used
• 20:57
  Speaker 2: to host, of which there are you know, a thousand
• 20:59
  Speaker 2: transcripts on the internet. Write an episode of Planet Money
• 21:01
  Speaker 2: about whether the FED is going to raise interest rates
• 21:04
  Speaker 2: by twenty five basis points or leave them unchanged, right,
• 21:07
  Speaker 2: And it wrote something that was pretty good, like not
• 21:10
  Speaker 2: a whole show, it's not there now, but at the
• 21:13
  Speaker 2: rate of current improvement, you could definitely imagine it writing
• 21:16
  Speaker 2: that episode pretty well in whatever a year or two
• 21:20
  Speaker 2: years or some amount of time when I will still
• 21:23
  Speaker 2: want to be gainfully employed. And like I do wonder
• 21:26
  Speaker 2: on this one, is there a day slash? How far
• 21:29
  Speaker 2: are we from.
• 21:30
  Speaker 1: The day when generative a I can just make a
• 21:34
  Speaker 1: podcast without me?
• 21:36
  Speaker 3: How does that make you feel?
• 21:39
  Speaker 2: I mean somewhat afraid, also like interested in figuring out
• 21:43
  Speaker 2: how to use it, right, Like it feels like a steamroller.
• 21:46
  Speaker 2: It's like, oh, maybe I should go get in that steamroller.
• 21:49
  Speaker 2: If my choices are get in the steamroller or get
• 21:51
  Speaker 2: run over by it.
• 21:53
  Speaker 3: Yeah, I think, like before I comment on it, I
• 21:56
  Speaker 3: think it's important that people understand, Like it's very true that,
• 22:03
  Speaker 3: like it's easy to think that I'll have a bullshitty
• 22:07
  Speaker 3: answer to a question like this because I work at
• 22:10
  Speaker 3: a tech company that's working on a lot of this stuff.
• 22:12
  Speaker 3: But you have to remember that, like, if that's true,
• 22:16
  Speaker 3: we're out of jobs as soon as like a human
• 22:19
  Speaker 3: is no longer in the loop. That's really bad for us.
• 22:24
  Speaker 3: Like does that make sense to you buy that.
• 22:26
  Speaker 1: At some margin?
• 22:28
  Speaker 2: Right, there's a long way between all the people who
• 22:30
  Speaker 2: are doing it now and zero people. There's a lot
• 22:33
  Speaker 2: of intermediate cases between the way it is now and
• 22:36
  Speaker 2: like a fully AI generated podcast, right, and like we're
• 22:40
  Speaker 2: already starting down the road, right, getting AI to write
• 22:43
  Speaker 2: show notes or something that's basically has happened now. And
• 22:48
  Speaker 2: you know, like I know the history of technology and
• 22:51
  Speaker 2: the labor market pretty well, you know, from the Industrial
• 22:54
  Speaker 2: Revolution on.
• 22:55
  Speaker 1: I'm pro.
• 22:58
  Speaker 2: Technological innovation. I believe in productivity gains and efficiency gains.
• 23:03
  Speaker 2: I'm also aware that there are instances when highly skilled
• 23:06
  Speaker 2: crafts people are displaced by technology. Right, that is definitely
• 23:09
  Speaker 2: a thing that happens. And I recognize that the pie
• 23:11
  Speaker 2: gets bigger and everybody's better off than the long run,
• 23:13
  Speaker 2: But like, I just want to not get pinched, right,
• 23:16
  Speaker 2: I just want to be you know, you don't want
• 23:18
  Speaker 2: to be the one.
• 23:19
  Speaker 1: I don't want to be the one. And you know.
• 23:21
  Speaker 2: I'm not out on using it. It's getting really good,
• 23:26
  Speaker 2: really fast. It's doing a lot of the things that
• 23:28
  Speaker 2: I can do.
• 23:29
  Speaker 3: There's one other thing I wanted to say, just about
• 23:31
  Speaker 3: the fear for your job thing, which is something we
• 23:35
  Speaker 3: say around here a lot, is that you should struggle
• 23:38
  Speaker 3: with your story and not your tools. That's almost like
• 23:42
  Speaker 3: a guiding light for us, is we want to take
• 23:45
  Speaker 3: all of the cognitive friction away from using the tools.
• 23:48
  Speaker 3: The funny thing about all of these things is like
• 23:51
  Speaker 3: there's a brief moment in time where you feel like
• 23:54
  Speaker 3: you have superpowers, but then everybody has them, and humans
• 23:59
  Speaker 3: once again become the differentiator. And we really think to
• 24:02
  Speaker 3: make like making great stuff is always going to be
• 24:05
  Speaker 3: a thing, and great is always going to be determined
• 24:09
  Speaker 3: by the human that's in the loop.
• 24:11
  Speaker 2: I mean, you know, there's this story about chess, right,
• 24:14
  Speaker 2: a computer chess program beat a person a long time ago,
• 24:18
  Speaker 2: decades ago now. But then after that people pointed out
• 24:22
  Speaker 2: the fact optimistically from my point of view, that a
• 24:25
  Speaker 2: computer plus a person could still beat any computer. Right,
• 24:30
  Speaker 2: a person working with a computer was better than the
• 24:32
  Speaker 2: best computer in the world. And that was like the
• 24:34
  Speaker 2: metaphor for like, yes, if we work with machines, we
• 24:37
  Speaker 2: can be better. That is no longer true now the
• 24:40
  Speaker 2: computer's kept getting better, and now people can't make them better.
• 24:43
  Speaker 2: Even a person plus a computer cannot beat a computer.
• 24:47
  Speaker 2: And I know that chess is less complex than the
• 24:49
  Speaker 2: real world, and so perhaps still a reason for optimism.
• 24:52
  Speaker 2: I certainly think I'm clever and good at making podcasts
• 24:56
  Speaker 2: and hope that I can do that. I hope that
• 24:58
  Speaker 2: I can work with AI to make something better than
• 25:01
  Speaker 2: ANYII or more like me or something.
• 25:05
  Speaker 3: It's it might not be true, though, but here's the
• 25:08
  Speaker 3: amazing thing. People are still playing chess. Right. It's like true,
• 25:13
  Speaker 3: there's some separation. Some separation happens where the machines become
• 25:18
  Speaker 3: so good and we just say, okay, you you machines,
• 25:20
  Speaker 3: you go off and do your thing, and we're going
• 25:23
  Speaker 3: to be here kind of reveling in our humanity with
• 25:28
  Speaker 3: each other. I think what we'll see is there's there's
• 25:31
  Speaker 3: going to be a certain category of content that's really
• 25:34
  Speaker 3: just about like the transmission of bits of information from
• 25:37
  Speaker 3: your brain to my brain, and that's all that it's about.
• 25:41
  Speaker 1: That.
• 25:42
  Speaker 3: Maybe we do one day see humans taken out of
• 25:44
  Speaker 3: the loop, but I really do believe there will always
• 25:47
  Speaker 3: be space for like at the core great content, storytelling,
• 25:53
  Speaker 3: whatever you call it, it's it's about feeling connected to
• 25:57
  Speaker 3: the humans and other people. And as soon as machines
• 26:01
  Speaker 3: play to have too heavy a hand, it's just not
• 26:04
  Speaker 3: interesting anymore.
• 26:08
  Speaker 2: We'll be back in a minute with the Lightning Round,
• 26:10
  Speaker 2: which includes a message from Andrew to.
• 26:13
  Speaker 1: His future self. That's the end of the ads. Now
• 26:24
  Speaker 1: we're going back to the show.
• 26:25
  Speaker 2: Okay, so this is the Lightning Round, now you ready.
• 26:28
  Speaker 2: It's just a bunch of questions. Do you use generitive
• 26:31
  Speaker 2: AI in your life outside of work?
• 26:34
  Speaker 1: Now?
• 26:35
  Speaker 3: You know what's interesting. I did something this morning where
• 26:38
  Speaker 3: I was actually like, I don't I don't even care
• 26:40
  Speaker 3: if it's wrong. I don't even care if it's.
• 26:44
  Speaker 2: Like the test of a theory is not is it correct?
• 26:46
  Speaker 1: But is it interesting?
• 26:48
  Speaker 3: Yeah? Exactly. I was asking it about I think, like
• 26:52
  Speaker 3: my son got hit in the head with a baseball,
• 26:54
  Speaker 3: and I was trying to I really should care about this. Actually,
• 26:57
  Speaker 3: you should.
• 26:58
  Speaker 1: Not ask chat GPT anything significant.
• 27:00
  Speaker 2: About that, not to give your parents advice.
• 27:04
  Speaker 3: It's stuff like that, like I've I've pretty quickly been.
• 27:08
  Speaker 2: Able to like that you should not be asking for
• 27:10
  Speaker 2: medical advice about your child.
• 27:15
  Speaker 3: I know. But when I say stuff like that, like
• 27:18
  Speaker 3: I would have googled it and probably just done what
• 27:20
  Speaker 3: I was going to do anyway. So it was almost
• 27:22
  Speaker 3: just a curiosity. He was fine, He didn't.
• 27:25
  Speaker 2: Need to go see a doctor, not according to JGBT. No,
• 27:31
  Speaker 2: I'm curious about your time working in a recording studio, right,
• 27:36
  Speaker 2: You worked in a recording studio where musicians came in
• 27:39
  Speaker 2: and recorded. Did you see there any like moments of
• 27:43
  Speaker 2: musical genius?
• 27:44
  Speaker 1: Is there one?
• 27:44
  Speaker 2: In particular?
• 27:46
  Speaker 3: I worked for this guy named Steve Albini, who is
• 27:51
  Speaker 3: a pretty well known engineer producer that was in some
• 27:57
  Speaker 3: popular kind of punk rock bands in the in the
• 28:00
  Speaker 3: eighties and currently and definitely saw some cool bands. But
• 28:05
  Speaker 3: I think also I really feel like I learned a
• 28:08
  Speaker 3: ton from watching him work. He's so talented, so articulate,
• 28:15
  Speaker 3: so smart in many ways, like an example of what
• 28:18
  Speaker 3: I aspired to be at the time, and so seeing
• 28:23
  Speaker 3: that output, but then also seeing him every day and
• 28:26
  Speaker 3: how hard he worked, it was a real like, oh,
• 28:30
  Speaker 3: this is how it happens kind of moment for me,
• 28:34
  Speaker 3: and it kind of inspired me. It inspired within me
• 28:39
  Speaker 3: a kind of work ethic that I'm not sure I
• 28:41
  Speaker 3: would have gotten to otherwise.
• 28:44
  Speaker 2: What's the best deal you ever got from group on?
• 28:53
  Speaker 3: Man? You know, it's so funny because, like obviously I
• 28:56
  Speaker 3: was asked. I used to be asked that question all
• 28:58
  Speaker 3: the time. I think it was a sensory deprivation tank.
• 29:02
  Speaker 3: They had a sensory deprivation tank center in somewhere in Chicago.
• 29:08
  Speaker 3: Had never tried. It was really cool.
• 29:11
  Speaker 2: This is a descript question. Now, how will you know
• 29:14
  Speaker 2: when it's time to do something else?
• 29:15
  Speaker 1: But leave? Dude?
• 29:18
  Speaker 3: I don't know if I want to say this on
• 29:19
  Speaker 3: a podcast, because if I do decide to take the
• 29:22
  Speaker 3: company public, it'll come back to haunt me. But I
• 29:25
  Speaker 3: almost want to say it specifically for that reason, Andrew,
• 29:29
  Speaker 3: I'm talking to future Andrew right now. You do not
• 29:31
  Speaker 3: want to be a public company CEO again, Okay, hire
• 29:35
  Speaker 3: someone else to do that. I know you're talking yourself
• 29:38
  Speaker 3: into it and saying it's going to be different time.
• 29:40
  Speaker 3: It's okay, but you hate it. It's the things that
• 29:45
  Speaker 3: those people are good at is and are interested in
• 29:48
  Speaker 3: is different than you go do something else.
• 29:53
  Speaker 1: Amazing.
• 29:54
  Speaker 2: I've never had someone leave themselves at time. Tast a
• 29:57
  Speaker 2: lot of podcasts before.
• 30:02
  Speaker 1: I'm going to send that to you. If you go public,
• 30:04
  Speaker 1: I'm to have you back on the show and I'm
• 30:05
  Speaker 1: going to play it to you.
• 30:09
  Speaker 2: Thank you, Thank you for being so generous with your time.
• 30:11
  Speaker 2: I appreciate your candor and I'm grateful for that.
• 30:14
  Speaker 3: I appreciate that I had. I had fun too. You're
• 30:17
  Speaker 3: good at your job in the sense that, like uh you,
• 30:21
  Speaker 3: you bring it out in me.
• 30:22
  Speaker 1: I'm better than a machine for now. It's gonna that's
• 30:26
  Speaker 1: my model, better than a machine for now.
• 30:34
  Speaker 2: Andrew Mason is the founder and CEO of Descript. Today's
• 30:38
  Speaker 2: show was edited by Sarah Nix, produced by Edith Russolo, and.
• 30:42
  Speaker 1: Engineered by Amanda k Wong. I'm Jacob Goldstein.
• 30:45
  Speaker 2: We'll be back next week with another episode of What's
• 30:48
  Speaker 2: Your Problem? And here, finally is the top of today's show.
• 30:52
  Speaker 2: The intro to the show as read if that's what
• 30:55
  Speaker 2: you'd call it, as generated by overdub descripts AI, powered
• 31:01
  Speaker 2: voice whatever emulator. After every interview we do for the show,
• 31:06
  Speaker 2: we upload the audio to a piece of soft were
• 31:09
  Speaker 2: called the script. Descript turns the audio into a transcript,
• 31:13
  Speaker 2: and then I can edit the transcript, cut out the
• 31:16
  Speaker 2: boring parts, move sections around, and when I do that,
• 31:19
  Speaker 2: descript edits the underlying audio to match. As software, descript
• 31:24
  Speaker 2: is pretty janky, it's buggy, it's constantly changing in ways
• 31:27
  Speaker 2: that can make it hard to use, and sometimes it
• 31:30
  Speaker 2: just blows stuff up. But we use it anyway because
• 31:34
  Speaker 2: descript is an incredible advance over what came before. Before
• 31:38
  Speaker 2: descript audio software represented audio files not as words, but
• 31:42
  Speaker 2: as waveforms, squiggly lines presented on a timeline. So when
• 31:46
  Speaker 2: descript came along, being able to edit audio by editing
• 31:50
  Speaker 2: words on a screen was a huge advance, and it
• 31:52
  Speaker 2: was an advance made possible by artificial intelligence.