E15: Will GPT-4 Cause Economic Transformation?

Nathan Labenz and Erik Torenberg delve into the upcoming economic transformation and the future of work in light of the threshold crossed by GPT-4. We're hiring across the board at Turpentine and for Erik's personal team on other projects he's incubating. He's hiring a Chief of Staff, EA, Head of Special Projects, Investment Associate, and more. For a list of JDs, check out: eriktorenberg.com. RECOMMENDED PODCAST: The HR industry is at a crossroads. What will it take to construct the next generation of incredible businesses – and where can people leaders have the most business impact? Hosts Nolan Church and Kelli Dragovich have been through it all, the highs and the lows – IPOs, layoffs, executive turnover, board meetings, culture changes, and more. With a lineup of industry vets and experts, Nolan and Kelli break down the nitty-gritty details, trade offs, and dynamics of constructing high performing companies. Through unfiltered conversations that can only happen between seasoned practitioners, Kelli and Nolan dive deep into the kind of leadership-level strategy that often happens behind closed doors. Check out the first episode with the architect of Netflix’s culture deck Patty McCord. https://link.chtbl.com/hrheretics LINKS REFERENCED IN THE EPISODE: Microsoft research paper: https://arxiv.org/abs/2303.12712 TIMESTAMPS: (0:00) Preview (1:56) Unpacking "There will be economic transformation," Nathan’s reflection after red teaming GPT-4 (4:55) A cross-section of prompts Nathan ran through GPT-4 (17:00) Sponsor: Omneky (18:00) Future of business and a new style of Taylorism of knowledge work (28:00) The “shrinking category of work” and a new definition of work (36:00) A future of “Muggles” and the 1000x developer (51:15) Are we facing a bifurcated economy and regulations from special interest groups? (58:30) Upcoming morale shift of white-collar workers and predictions TWITTER: @CogRev_Podcast @labenz (Nathan) @eriktorenberg (Erik) SPONSORS: Shopify is the global commerce platform that helps you sell at every stage of your business. Shopify powers 10% of ALL eCommerce in the US. And Shopify's the global force behind Allbirds, Rothy's, and Brooklinen, and 1,000,000s of other entrepreneurs across 175 countries. From their all-in-one e-commerce platform, to their in-person POS system – wherever and whatever you're selling, Shopify's got you covered. With free Shopify Magic, sell more with less effort by whipping up captivating content that converts – from blog posts to product descriptions using AI. Sign up for $1/month trial period: https://shopify.com/cognitive Thank you Omneky for sponsoring The Cognitive Revolution. Omneky is an omnichannel creative generation platform that lets you launch hundreds of thousands of ad iterations that actually work, customized across all platforms, with a click of a button. Omneky combines generative AI and real-time advertising data. Mention "Cog Rev" for 10% off. This show is produced by Turpentine: a network of podcasts, newsletters, and more, covering technology, business, and culture — all from the perspective of industry insiders and experts. We’re launching new shows every week, and we’re looking for industry-leading sponsors — if you think that might be you and your company, email us at [email protected]. More show notes and reading material released in our Substack: https://cognitiverevolution.substack.com/

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Nathan Labenz?

  And then we've also seen the Microsoft research paper in which they basically said that GPT-4 is appropriately considered an early AGI. Can a language model do the entirety of this job? You will almost always conclude, no. But that's going to be 5% of cases, I would guess, in the early going. And then, you know, before too long, it'll be 1% of cases, and then before long, 0.1% of cases. Everybody who's ever wanted to take piano lessons and didn't feel like they have time might have time, you know? And the AI might give you the piano lesson, by the way, as well. But there's not necessarily money waiting for you to, like, perform at the end of that.

• 0:43Nathan Labenz?

  I know this thing can answer my medical questions, but it's not allowed to. And it's not allowed to because of who? The top comments that I saw were all women saying, well, maybe at least ChatGPT will listen to me when I go in and talk about my problems. So. Hello and welcome to The Cognitive Revolution, where we interview visionary researchers, entrepreneurs, and builders working on the frontier of artificial intelligence. Each week, we'll explore their revolutionary ideas, and together, we'll build a picture of how AI technology will transform work, life, and society in the coming years. I'm Nathan Labenz, joined by my co-host, Erik Torenberg.

• 1:27Erik Torenberg?

  Before we dive into The Cognitive Revolution, I want to tell you about my new interview show, Upstream. Upstream is where I go deeper with some of the world's most interesting thinkers to map the constellation of ideas that matter. On the first season of Upstream, you'll hear from Marc Andreessen, David Sacks, Balaji, Ezra Klein, Joe Lonsdale, and more. Make sure to subscribe and check out the first episode with A16Z's Marc Andreessen. The link is in the description. Let's get into it. So in our last GPT-4 solo episode, you, we talked about how you told the OpenAI team that there will be economic transformation and you are confident in that. Why don't you describe or unpack what you mean by that?

• 2:13Nathan Labenz?

  Yeah, sure. I mean, I think there's a lot to be still figured out in terms of the details. But maybe just to recap real quick what I did and sort of what my takeaways were from it. I just set out to get the AI, which, you know, is now the model we now know is GPT-4, to play the role of a lot of different professional advisors that we interact with, you know, in daily life and that we, you know, pay real money for and see real value in and that people work hard, you know, to get to these positions where they're qualified to play these roles. So the obvious, you know, kind of first one would be a doctor, right? You have a concern, you want to talk to a doctor about it.

• 2:55Nathan Labenz?

  In today's world, you have to set up an appointment, you have to, you know, drive in there, you have to wait in the waiting room, you know, it ends up often taking a half a day to do that. A lot of people feel like even when they go to all that trouble, the doctor doesn't necessarily listen to them, you know, as well as they might like. But that's kind of the state of the art. Setting up a dialogue with an AI doctor was trivially easy for me to do. I use just, you know, simple kind of role casting, as it's called in the prompt engineering world. Tell the doctor or tell the AI that it is a doctor, that it's going to, you know, interact with a patient, set up a dialogue, and then just chat back and forth. And I really found it to be extremely easy

• 3:38Nathan Labenz?

  to essentially replicate the interaction that I had had with a human doctor. Now, I'm not a doctor. Evaluation on this stuff is really hard. And so I kept my experience pretty few in number and focused on experiences that I had had, you know, real, fortunately, minor for me, medical issues or, you know, things we'd been concerned about with our three-year-old, you know, that we might have otherwise called a pediatrician about. Just within that two months, we did get to a point where There was one incident where we didn't call the pediatrician where we otherwise would have because we got some advice from the AI and we deemed it like good enough, reliable enough to kind of assuage our concerns. And you know, we went on with our day without

• 4:22Nathan Labenz?

  even actually calling the pediatrician, right? So just in that two-month window, even knowing that it's, you know, an alpha, knowing, you know, all the difficulties around evaluation, we personally got to a point where there was a clear moment of substitution where, you know, real question that we had, the AI was able to answer satisfactorily enough. We didn't have to call the doctor. We didn't have to go in. Now, it wasn't, you know, a big problem. Turned out it wasn't really a problem at all. So that's maybe an easy case. But, you know, that was a call that otherwise would have been made, you know, and it wasn't as a result of our access to GPT-4. So you kind of go down the line, you

• 5:02Nathan Labenz?

  know, of a lot of different sorts of professional roles. And I think I told the story last time of the dentist, you know, where I had the weird thing on my tooth and it gave me this feedback that's like, it sounds like your dentist did something non-standard that, you know, is not in line with best practices or standard of care. And sure enough, you know, that was definitely true. Tried it in an immigration law situation, did some kind of far out things like playing with a three-year-old, giving tech support to my grandmother, asked the AI at one point to be a mediator between two neighbors that had a dispute over a fence. I played the role of both neighbors, you know, that were kind of at odds with each other. And the AI's job was to bring us together and try to find some common ground. Did a pretty admirable job on that. Interestingly, I sent

• 5:46Nathan Labenz?

  it to a friend of mine who is a lawyer now and asked what he thought about it. And he said, well, you know, from a legal perspective, like this is a pretty open and shut case. Like the one neighbor's fence is on the other neighbor's property and, you know, the neighbor whose property it is can say what happens and that's that. But I thought it was actually kind of revealing in a way too, right? That it's like,

• 6:08Nathan Labenz?

  Maybe we didn't actually want a lawyer for that situation, you know? What we wanted, what we needed was the mediator, not the legal perspective. Because we're going to continue, you know, this is all fictional, right? But we're going to continue living next to each other as neighbors. And, you know, it really helps if we can resolve this in a non-legalistic way. So like, what role you ask the AI to play can make a world of difference, I think, in terms of, you know, how effective it's going to be for you. Other experiments that I did were setting up a personal, virtual personal trainer, you know, set up a group chat type of dynamic and then brought in the AI to be like the coach. So, you know, we've got me who was in like, okay, shape. We've got my wife who's pregnant. You know, we've got my brother in there who is probably in the best shape of all of us, but frequently injured, you know?

• 6:52Nathan Labenz?

  And it's doling out specific advice to each of us. Like Nathan, you should, you know, see if you can do five more next time. And, you know, Amy, well, you're pregnant. You've just got to take it easy. Like, anything you can do is a bonus. You know, you just got to focus on staying well, getting to the end of this thing. You know, very kind of friendly. Honestly, if you, you know, obviously people talked about the Turing test for a long time. If you double-blinded something like this and ran a study of which of these is the real, you know, virtual, you know, trainer by text versus the AI, you're not going to distinguish those, right? So, giving gifts, you know, coming up with gift ideas, generating grocery lists, you know, saying, hey, I want to cook this.

• 7:37Nathan Labenz?

  You know, can you pull me up a recipe? Translate that to a shopping list. We can put that in the cart. You know, it goes, it just goes on and on, right? Called a virtual car repair place and told them what's going on with my old car that had a problem with it. I would say better service from the AI in the role of answering the phone at the garage than I would expect to get calling a real garage, right? I mean, patience and just kind of the willingness to listen to you and really sit there and, you know, get to the point where you're happy with the conversation is a huge strength. For an AI system compared to, you know, the guy who's at the garage who's like not getting paid to talk to you on the phone, right?

• 8:22Nathan Labenz?

  He's got cars there that he needs to fix. He's probably got, you know, grease on his hands, whatever, as he's trying to talk to you. You know, we've all had that that experience where the guy's just like, look, bring it in if you want to bring it in. I can't really diagnose this over the phone, you know, and you're kind of frustrated and he doesn't really want to deal with that. And so, you know, if the AI doesn't care, it will sit there and answer your questions all day long. You know, it did in that moment remind me that, hey, you know, really, you should bring it in and we're here for you, whatever. This is all without any fine tuning, right? All I'm doing here is setting up. Your job is to answer the phone at a car repair shop and help people, you know, diagnose what's going on, give them a sense of what to expect. You know, and then obviously ultimately they're going to come in and we'll get it fixed.

• 9:06Nathan Labenz?

  Hardware store associate was another one that I tried where I have an old house. And we've got these old, you know, light fixtures in the basement, probably put up in the eighties.

• 9:16Nathan Labenz?

  The lights are, you know, old halogen lights, they're energy hogs, whatever. So looking at potentially upgrading those. You know, it's a very idiosyncratic thing, right? The these particular lights with these particular kind of wirings. What can we do about that? You know, do I have to rip the whole thing out or can I just replace it? It suggested a specific light bulb, which I was then able to search for, find that that light bulb did exist, that it is the right light bulb, you know, to kind of convert these old fixtures into a modern, you know, low energy lighting. Ordered them, put them in and they fit, you know, so it's like, wow, that's amazing, right?

• 9:57Nathan Labenz?

  Just for me setting up my situation exactly as I would do if I walked into a hardware store, I, you know, start with the story. Got an old house. I think the guy put it in, you know, same exact dynamic. I was looking for a solar panel, just to have a little bit of energy generation backup in this scenario that my power goes out because we had a big power outage during this time of red teaming. And, you know, I wanted to be able to charge a cellphone. So I had to figure out, well, how much solar panel do I need to charge a cellphone? Well, how much energy does a cellphone take? So go through that whole conversation, you know, get a consult on what kind of power generation, you know, is enough to charge a couple of cellphones. Like, what would it take if I wanted to run a refrigerator off of it?

• 10:41Nathan Labenz?

  What would it take if I wanted to run an air conditioner off it? Spoiler alert, like, you don't want to run your air conditioner off a retail solar panel. But, you know, the cellphones take very little power. I was actually really kind of amazed to learn how little energy it takes to to charge and run a cellphone. Menu planning, catering, you know, my wife puts on these these huge events, thousand people. It's typically an all vegan menu in the community that she serves. And she's had a lot of problems getting, and they do them all over, right? So it's different cities and like, you know, often there's a new catering company that she's working with. A lot of challenges in getting a good vegan menu out of, you know, kind of a typical catering

• 11:25Nathan Labenz?

  company that doesn't often do full vegan menus. I sent her, I sent her an example of, here's a, you know, a menu setup for your, you know, three, four day event. She said, she said, GPT catering is amazing. It's better than all of our caterers at setting up the menu. So it goes on and on. There were a few that I tried that were definite fails at the time. And I would have to go back and verify, like, to what degree has this been patched? But the math ability was still pretty limited. It could do like your SAT problems, your kind of high school level, you know, story problems pretty consistently

• 12:06Nathan Labenz?

  and pretty well. But if you went up a level beyond that, you know, it started to be kind of iffy, like college math. Yes, I could do some calculus and stuff, but you would hit the limits. And then especially as a teacher, there were challenges if you tried, if you got confused as a student, then a lot of times it might also get confused and kind of, you know, think that what you said was right, or, you know, we've seen these sorts of things with Bing, where like it gets very out of sync in terms of what day it is. And I definitely experienced some of that kind of stuff with math, with chemistry as well, like balancing, you know, chemical equations. It could do basic equation balancing, but if I posed as a student and

• 12:49Nathan Labenz?

  said, you know, here's what I think it is, then I found that it was pretty easy to confuse it still. So in the end, I probably did, you know, a dozen of these different roles in some depth. And, you know, they were kind of a pretty broad cross section of the jobs in society that range from, you know, literal M D or lawyer, you know, advanced degree, high salaries, you know, high prestige professions, to, you know, things where there is a lot of kind of idiosyncratic knowledge, but it's not necessarily so high status, like a hardware store associate, or, you know, somebody taking a call at a, at an auto mechanic shop. And

• 13:31Nathan Labenz?

  just across the board, you know, it seemed like it did very, very well, with just a couple of exceptions that I mentioned. So this has all been really borne out now by publications that have come out as well, right? We've, we've seen the GPTs or GPTs paper, which is, you know, the second GPT there is general purpose technology. And then we've also seen the Microsoft research paper in which they basically said that GPT four is appropriately considered an early AGI. And I think they, you know, they've characterized it at quite a bit of length there, too. So I would definitely recommend both of those papers for, you know, just further characterization of what the model can and can't do. All of that is just

• 14:16Nathan Labenz?

  kind of raw material, right? Like this is what we saw, this is what has been observed now.

• 14:22Nathan Labenz?

  How does that lead to economic transformation? Like, on the one hand, it's pretty obvious. You know, if you just kind of, at least for me, if you just kind of ask your gut, like, okay, I now have an AI that can handle a first conversation with a doctor, you know, first conversation with a lawyer in, you know, and go 10 rounds and like have some real depth and substance to it. Does that feel like it would be transformative? Like, yes, to me, it definitely seems pretty obvious. But then you still have the question of like, how? You know, one of the lines in the GPTs or GPTs paper that is really true is that if you go around and look at jobs as they exist today,

• 15:06Nathan Labenz?

  and you ask, can a language model do the entirety of this job? Then you will almost always conclude, no, that it can't. And that could be for multiple reasons, right? There could be too much context required. It could be that there's a physical nature to some of the work or all of the work that, that, that obviously a language model, you know, without a robot to control is not going to be able to do. So, you know, it's almost all jobs you'd look at as they are bundled today and say, yeah, a language model can't do that.

• 15:45Nathan Labenz?

  From that, though, I infer that what we're about to see is a huge unbundling of jobs into tasks. And I think that that is basically the same lens that the GPTs or GPTs paper takes, looking at what are the tasks that constitute a given job? How prevalent are they? You know, what, what is the task mix for any given role? And then which of those tasks can a language model either do outright? Or accelerate, I think they said like, you know, cut the time by at least half that it would take to, to do that task.

• 16:26Nathan Labenz?

  And there they find that a lot of the tasks are doable. So they kind of categorize jobs by how exposed they are to language model impact, where they define exposed as what percentage of your time is spent on tasks that can either be, again, done or like greatly accelerated with language models. And there they find that about half of all jobs have about half of all tasks exposed to language models.

• 17:04Erik Torenberg?

  Omnikay uses generative AI to enable you to launch hundreds of thousands of ad iterations that actually work, customized across all platforms with a click of a button. I believe in Omnikay so much that I invested in it, and I recommend you use it too. Use Cogrev to get a 10% discount.

• 17:22Nathan Labenz?

  So there's still a lot of kind of wiggle room there to figure out exactly what happens and exactly how things play out. But I think we're going to see, and I think it's likely going to start with major corporations because they are the ones that are going to be buying the most advanced systems. They're the ones that are going to be seeing huge dollar sign savings opportunities in fine tuning the most advanced systems for reliability, to do things exactly the way that they want them to be done or need them to be done in a corporate context. And so I expect that you're going to see just tons of first large businesses, and then it will probably kind of trickle down to smaller businesses, looking at,

• 18:07Nathan Labenz?

  what is it that we do around here? Like if we actually, and you're kind of see a new, you know, new kind of tailorism for knowledge work. So instead of, you know, we've seen this once before, right? With the transformation of kind of artisanal manufacturing to assembly line style manufacturing. Like if you go back before Henry Ford and before the assembly line and look at how people made cars, they were making cars before there was an assembly line, but.

• 18:39Nathan Labenz?

  The tolerance, the machining was not so precise. The tolerances were all much looser. People kind of had to make things fit. You know, in a very sort of artisanal, hands-on, like, runtime, problem-solving sort of way to actually get a car out of a process that would work. And so every one that they made, you know, tended to be a little bit different. And you'd have kind of, you know, less reliability, less power, you know, just you couldn't, you could not get the same kind of precision. Well, now you go into a factory and you see a high level of automation. You see extremely tight tolerances, you know, in the most like advanced,

• 19:21Nathan Labenz?

  you know, go to a semiconductor fab and like a speck of dust, you know, can throw the whole thing off. And so it's been really kind of studied, broken down. What are the tasks? How precisely can we do them? And that's, you know, totally transformed manufacturing. I think a very similar phenomenon is coming to knowledge work where people are going to say, okay, let's look at this job. Let's figure out which parts of it can be done by a language model. And then let's train a language model to do it. Now, sometimes they might even just be able to do it off the shelf. But a lot of times there is going to be some special training required to really make sure that like,

• 20:03Nathan Labenz?

  you know, you're getting the, especially the reliability of the output that you want. So, you know, people will really spend the time and put some elbow grease in to ensure that the language models are reliably doing the tasks. And, you know, hopefully people will take care that, you know, they're robust to weird scenarios and all that kind of stuff. But honestly, you know, you only need to get to a certain threshold of reliability before it makes sense to say, all right, look, you know, we pay people X dollars an hour to take phone calls. And if we have a language model do that instead, then we probably see at least a tenfold, you know, reduction in cost, 90% reduction in cost.

• 20:48Nathan Labenz?

  And often maybe like a 95, 98, 99% reduction in cost relative to what the human would have cost. In the doctor scenario, you know, if you figure whatever average appointment is 100 bucks, that's probably conservative in the US. You can have that 45-minute conversation. For basically a dollar at current prices.

• 21:12Nathan Labenz?

  If you're talking about a call center or customer service sort of thing, you know, maybe that difference is not 100 to one, maybe it is more like 10 to one, but it seems like many things are going to be kind of 10 to 100 to one cost ratio. And that's going to be very enticing, especially when you consider the fact that it's also going to enable 24/7 access and especially as tooling and you know, systems, databases, memory types of things get built out, then you also won't have the problem that you have so often today where you call and the person like doesn't know who you are and you know, they you didn't talk to them last time. So they have, you know, no memory of this and the notes in the system often aren't that great. You know, all that stuff is going to kind of get smoothed over as well.

• 21:58Nathan Labenz?

  And you end up, you know, pulling these things apart, right? So what, what parts of your job can be done by a language model ultimately likely do get sent that way, the parts that can't, you know, then they maybe get rebumbled into another job. For a while, if you take the call center type of situation, like escalation may be the thing, you know, that is the kind of most, the most common thing that the language model can't handle, right? Like if you've ever said, you know, is there anybody else I can talk to? Can I talk to a manager? Like, I'm not happy about this. You know, then that type of thing maybe is in the first generation, kind of where the language model

• 22:39Nathan Labenz?

  responsibility ends and it, you know, it kicks up to a human who maybe has more discretion, more context, you know, whatever. But that's going to be, you know, 5% of cases, I would guess in the early going and then, you know, before too long, it'll be 1% of cases and then before long, you know, it might even be 0.1% of cases. So, you know, I think people get tripped up a lot of times when they think about this future and they're like, well, there's no way that the AI can do everything. You know, we'll always need humans. And I think that that is.

• 23:15Nathan Labenz?

  True, but doesn't quite mean what, or at least it's true with the current technology, right? I'm kind of bracketing like no further, you know, major upgrades. Like I'm not, I'm not trying to analyze GPT five, six, seven here by any means. But even with just the current technology, you know, we'll always need humans. Yes. But what the humans are doing is probably going to shift a lot. There's going to be a flurry of activity, analyzing jobs, pulling things apart, figuring out what tasks can be delegated to AI, figuring out, you know, the training set and the, you know, the reinforcement that can get good performance out of them, and then figuring out how to wire that into a system, how to handle escalations

• 23:59Nathan Labenz?

  when they're needed, you know, how to handle whatever other corner cases. And then the things that are not doable, you know, those maybe get rebundled into other jobs. So you think about going to a doctor's office, right? Like today, you might have the same person answer the phone, schedule your appointment, and then also like take your weight and your blood pressure when you come in. The scheduling, you know, I don't, I think it will be months at most before there is a competitive market in, you know, automated scheduling systems for doctors that you can just call and talk to. And now, you know, everybody, everybody

• 24:40Nathan Labenz?

  wins, right? Like the patient gets better service, you can call anytime. The doctor's office is going to save money on that, you know, in theory, that gets passed through to the customer. All great. Well, what about that person who was sitting there, taking the calls and, you know, and checking people in when they get there? Well, it probably depends, you know, it depends on a lot of details, but they don't have to take the calls anymore. Is there still enough of a job left to?

• 25:09Nathan Labenz?

  You know, merit somebody to to do the weight and the blood pressure. You know, maybe the doctor picks that up in certain offices. Maybe if it's big enough, you know, you had three receptionists before that were also doing that. And now that goes down to one. The details, the context and all that kind of stuff are really obviously going to matter a lot. But people are very good at figuring out, you know, certainly this is like what entrepreneurs do, right? It's like identify opportunity to use new technology to solve problems in, you know, dramatically more efficient ways. And I think that's going to happen extremely quickly and will kind of be everywhere. You could say like, does this mean there will be no jobs or that people will have nothing to do?

• 25:54Nathan Labenz?

  And I wouldn't jump to that conclusion. That's another thing where I think there's kind of a fallacy that's like, it's not all or nothing. You know, it's not, it's not that AI is either going to complement people co-pilot style or it's going to replace them, you know, assembly line replaced artisanal manufacturing. It's going to be both at the same time. And, you know, similarly, what are people going to do? I think is is a much more nuanced and contextual question than saying like, nothing's going to change or everything's going to change and like, nobody's going to have anything to do. The outcome is going to, in all likelihood, involve a ton of change. Uh, but the the kind of end state is not going to be like on one of these poles. It never is.

• 26:39Erik Torenberg?

  Yeah. It it reminds me, you, um, you and our friend Antonio Garcia Martinez had a little back and forth on Twitter the other week where he just posted, you know, some of the stats around how every technological innovation, where, you know, people were concerned that it was going to lead to the end of, uh, end of certain, certain jobs. It it did in some instances, you know, you copy and pasted the the farmer, um, you know, stats where, you know, the the country used to be like 90% farmers or whatever, and now it's like 3%. The technological innovation ended up creating more jobs in general because human desires and needs are are infinite. Do you think that is a unhelpful paradigm for thinking about what we're seeing here with with AI?

• 27:20Erik Torenberg?

  I is this time actually different relative to all the other previous, um, you know, technological advancements? And, you know, what what types of people will be the the farmers that will be looking, looking back on? How would you, uh, eh eh expand on on some of those ideas?

• 27:36Nathan Labenz?

  Yeah, it's, I mean, it's a really tough question. So I should, you know, say, I think I have a read on where it's going. But certainly I also expect to be surprised by plenty of things. I do think this time is different in that it's not really clear what we could shift everybody into.

• 28:01Nathan Labenz?

  And I also think there's a generational question, which is a big one. So I think that I would cite Yuval Harari and Sapiens, I believe, or his next book, After Sapiens. And it's funny, people hate on that book, but I think there's a lot of value in just trying to zoom out and tell that story from, you know, as kind of far removed as possible. Obviously, you can, you open yourself up to being criticized for missing lots of important aspects of the story when you try to do that. But I find it to be a pretty useful attempt to cover whatever, a hundred thousand, 10,000 years of history in, you know, just a few hundred pages. But I think, I would, I think it was from him that I heard just the general notion that like, we can do kind of

• 28:45Nathan Labenz?

  a couple of different general kinds of tasks, like physical labor that relies mostly on our muscles. And then like, you know, cognitive labor that relies mostly in the brain. And obviously there's some overlap there, but there's not like an obvious third place for us to go, right? So when sort of physical work got semi automated through machines and the harnessing of, you know, fossil fuel and electrical power, then it was like, okay, cool. Like, we don't have to do as much of that stuff anymore. We can go do more cognitive work and that's great, you know? Largely, people prefer it and, you know, there's certainly a lot less injuries from it and whatnot. So, you know, it's a win.

• 29:27Nathan Labenz?

  But if you zoom out to that level, like, where are we going to go beyond cognitive work? And you could maybe come up with some candidates, like, well, what about like emotional work? Is that distinct enough from cognitive work? My experiments suggest that the AIs are getting pretty good at that sort of emotional work as well.

• 29:54Nathan Labenz?

  I don't have a lot of experience with like cognitive behavioral therapy or that kind of modality, but it seems like that will be pretty readily provided by language models. You know, is there some sort of connection, the sort of realness of which is can't be recreated, like possibly in some scenarios for some people, I could see, you know, kind of emotional work becoming a category that's like distinct from cognitive work and kind of remains a sort of, you know, a place where humans are maybe not even dominant, but like preferred for, for kind of reasons.

• 30:36Nathan Labenz?

  I think you could also kind of imagine like an interesting local kind of bespoke service economy, you know, highly like individualized entertainment. People, you know, do these sort of like dinner, party murder mystery games, you know, for example, could an AI like put together a good murder mystery for a group to solve? Yes, I don't think that's like out of the range of what it could do, but I could imagine, you know, that there's sort of in a world of like abundance and, you know, all the kind of bullshit jobs and, you know, frontline customer service and all, you know, all that kind of stuff being delegated to AI, then maybe we end up in a more, you know, kind

• 31:21Nathan Labenz?

  of local network sort of economy where we're kind of doing cool, you know, fun things with and for each other. But honestly, it's hard to, you know, if you think of like work as stuff that people don't necessarily want to do, therefore they need to get paid to do and it won't get done otherwise. It seems like that's a shrinking category, you know, it, it, it seems like there's just, there's going to be a lot less work that.

• 31:48Nathan Labenz?

  Only humans can do that they won't do unless they get paid for that, you know, that won't get done otherwise. That seems like a shrinking category. So I, I do think people will have plenty to do. Like, I don't worry about myself at all. Like, if everything I do got automated, you know, I don't think I'd be bored. I'd still be interested in studying AI, for example, or, you know, I've got plenty of, I, I'm interested in history. Like, I'd love to read more history books than I currently feel like I have time for.

• 32:16Nathan Labenz?

  So I think there is tremendous potential for people to be self-actualized, to be, you know, to develop their talents, to create stuff that they find worthwhile and that others find worthwhile.

• 32:31Nathan Labenz?

  But I don't think that that's quite the same as work as we think of it today. You know, and I, I need to maybe refine this definition a little bit, but I'm kind of liking the, people wouldn't, you know, that wouldn't get done otherwise. Uh, the people, you know, don't like to do, like, you have to pay them to do it. Uh, you know, that only people can do, that there's not really a way to mechanize. Something like that does feel like a pretty good definition of kind of work as it, you know, as it exists today. And I do think that is going to be a significantly shrinking category.

• 33:06Erik Torenberg?

  Yeah, it is interesting. It does feel like the, um, the concept of like, uh, you know, muggles from Harry Potter almost like might emerge in our society and how we think about just like the different, like, classes of people and their, their capabilities. And to some extent, it kind of exists today. Like, some people can code and build tremendous things, but it, it's not as, you know, once those people are like supercharged and the other people are, you know, or many people are like basically replaceable. It does feel like that, that just gap in understanding of people is going to, to widen. Feel, feel free to challenge any of those premises, but.

• 33:46Nathan Labenz?

  I mean, we, we talked to Amjad a little bit about the 1000x developer. And that definitely seems like part of the future. It seems like some people are going to be so good with these tools, you know, these programming assistants or whatever, that they can, likely will be able to accomplish what, you know, a full team can do today. And.

• 34:12Nathan Labenz?

  I do think it's true that there's like a lot more demand for software, especially at extremely low prices relative to the software that's being created today. But again, I don't know that that's like infinite, right? I mean, we can only like, what sort of software is there like infinite demand for? I guess, you know, video games, like just software that entertains on some level would be kind of the, you know, end state of like, if there's nothing else to do, then you'll, you know, create things to explore and mess around with. But even then, we only have so much time, right? So you can kind of imagine that at some point, if you're getting to, and by at some point, I don't mean like,

• 34:56Nathan Labenz?

  you know, this is not like a hundred years away. We're talking like a handful of years away, maybe. Where you can sort of have you know, a narrative language model type system, kind of generate plots of games or, you know, customizations to you from archetypes that have been previously created. And on top of that, you can kind of spin up 3D environments, you can have kind of the language model, speak to another specialist model to generate a 3D landscape for you to explore. And you, you know, you can do all this in kind of a virtual reality or an augmented reality generated for you on the fly.

• 35:36Nathan Labenz?

  I mean, how much more software do we need, you know, beyond that? Like, if you can see a pretty clear path to enough software to entertain people on an individualized level, nonstop.

• 35:48Nathan Labenz?

  I don't know what, you know, I don't know really what more is needed than that. Like there's language models right now cannot do science. So that's kind of a, you know, I think maybe a really important threshold that we have not hit yet. We've hit a ton of thresholds over the last two years. You know, I can't really write coherent copy. It's not really even useful. It's like a marketing copywriting assistant was like two years ago. And now we're at like, closing in on expert level doctor performance. There's a lot of little thresholds that have been hit along the way. We've not yet hit the one where it's like, AI can do original science. So, you know, for now,

• 36:28Nathan Labenz?

  that remains a thing. And similarly with like hardcore engineering, you know, it's not yet to a point where, you know, it's not going to set up its own semiconductor, you know, supply chain anytime soon. Well, maybe sometime soon, but you know, again, not with this generation. This generation will not do it. Next generation, as soon as that comes, you know, we'll have to reevaluate all claims. So again, there's like, there are some things that are not in the, in the province of like what a language model can do, even with fine tuning, it's just not there yet.

• 37:04Nathan Labenz?

  But yeah, it seems like, uh, seems like a shrinking category and it seems like something that fewer and fewer people will be, you know, kind of the rare specialists that have skills that like AI just doesn't touch yet. So I don't, I don't have a percentage on that, but it feels like today, you know, how many people are really in semiconductor fabrication? Right? Like not that many. How many people are really optimizing at a low level? How iOS interfaces with, you know, the hardware? Not that many, you know, it's, it's an important job. It's extremely high skilled and it's not immediately under threat from a language model.

• 37:48Nathan Labenz?

  But just, just don't think there's that many people. The other thing I wanted to mention too, from earlier was the generational question. And I think this is really important because this is happening so fast. So, you know, there's a lot at this point, right? We've We grew up in the end of history era, not to get, not to encourage on the MOZ territory and discussion topics here, but it was,

• 38:14Nathan Labenz?

  you know, kind of assumed for a while that the economy will adjust and, you know, it's fine if, like, you know, I'm from Detroit, I'm in Detroit. It's fine if the car companies, you know, send all the jobs to China or Mexico or Vietnam or what have you, because we'll adjust, we're dynamic, everything's going to be great. And we ran that experiment for like 20 years and it was a pretty gradual process. A lot of people did not adjust in the way that the textbook economics predicted that they would, right? Like these, there's towns all over the Midwest that are not what they once were because the main employer, the main factory, whatever, is gone. There wasn't a, you know, a full recovery. There wasn't this, you know, there was some

• 38:59Nathan Labenz?

  adjustment, certainly, but a lot of the adjustment was like people left the town or some of the other adjustment was like people adjusted to lower standards, standards of living individually, you know, they got jobs that like pay less and are lower status than what they're used to. And people are not happy about that. So, and that's obviously been, you know, I think a significant force in American politics by any telling. That was a slow process compared to what I think we're going to see over the next two to five years. And if you, if you do those kinds of things on like a generational time scale, as a society, you give, you know, collectively, we would give ourselves the opportunity for, okay, well, yeah, it might suck for this person who's, you know,

• 39:43Nathan Labenz?

  in their 40s and they've got 20 years of this experience and that job is going away. And, you know, to some degree, there's like creative disruption, nothing we can do about it. But like their kids, you know, can come up and kind of adapt to a different reality and they'll be educated in a different way and they'll, you know, they'll prepare for other jobs and they'll get those jobs. So I think society, I guess what I'm trying to say there is I think society historically has been more adaptable than individuals. And a big part of that, as I understand it, is the generational change that, you know, you can educate the next generation for the opportunities that will exist and, you know, the the previous generation maybe doesn't catch up and that sucks for them, but, you know, that is kind of something that we've tolerated.

• 40:25Nathan Labenz?

  If that happens to a huge section of employed people over just a couple of years across many sectors, all at the same time, then again, I don't really know how we adjust to that, you know, and that's not to say that people would be unfulfilled or, you know, would wouldn't be incapable of finding like things to do that are self-actualizing. It's just not clear to me that there's going to be a lot of demand to pay people to do that sort of stuff, you know, you like everybody who's ever wanted to take piano lessons and didn't feel like they have time, might have time, you know, and the AI might give you the piano lesson, by the way, as well.

• 41:07Nathan Labenz?

  But there's not necessarily, you know, money waiting for you to like perform at the end of that. So you can kind of play that out across a lot of different things, you know, people have a lot of hobbies, they have a lot of interests, they have a lot of ways that they would like to entertain each other and, you know, socialize and explore. And it could be a very rich, fun, rewarding life. But it seems like the paradigm of like money flowing to those activities does not necessarily cross over, as far as I can tell.

• 41:42Erik Torenberg?

  Is there anything I haven't asked you that that I should ask you or any big questions you yourself are having that we have not yet discussed on this topic?

• 41:52Nathan Labenz?

  The impact of regulation is an interesting one, I think, that could shape a lot of how this plays out. I don't think, I would not expect that we're going to avoid the kind of changes that I'm describing. But I do think, again, like, there's a lot of details to be worked out and the timeline and exactly who can do what, under what circumstances, with what licensing and, you know, whose say so is a really big one.

• 42:25Nathan Labenz?

  Honestly, I've been a little surprised by how slow the interest groups have been in responding to the developments that we've seen. It seems like there's been more denial than like actual attempts to do something about it for the most part. But just, you know, take a couple leading, you know, American civil society organizations, the AMA and the ABA, the, you know, the Medical Association, the Bar Association. It seems to me like leaderships, leadership of those organizations is going to look at data like this and they're going to say, we've got to protect our members' interests here. How do we do that?

• 43:09Nathan Labenz?

  The obvious thing would be to say, we'll make it, you know, we'll try to make it illegal to, you know, like it's already illegal to practice medicine without a medical degree and it's already illegal to practice law without a law degree or a license. Then maybe we can say it's illegal to use a medical language model without a proper license, right? Or at a minimum, you know, it has to be under supervision of a, you know, a licensed doctor or lawyer or whatever.

• 43:40Nathan Labenz?

  As of, you know, I don't know, it seems like that should already be happening. Maybe it is starting to happen. We've heard a little bit of noise from the FDA around regulating language models and like devices, which means, you know, you, you do have to show safety and it might just be a safety standard. I would have to double check to see if there's also like a improvement on, you know, previous clinical best practice. A little bit of a difference between device regulation and drug regulation, which I don't have full mastery of. But FDA is starting to get into the game a little bit and kind of saying that this is going to fall into a device, you know, type of paradigm. Haven't heard a lot from lawyers, you know, haven't heard, I haven't heard as much as I would expect from doctors.

• 44:25Nathan Labenz?

  But I think that has to be coming. Like, I would, I would be shocked if we get out of this year without the first fights ramping up on who is going to be allowed to do what with these models, who's going to be, you know, responsible for supervising, who's going to be liable, uh, you know, if, if OpenAI is providing something directly to a user, you know, is it their responsibility to make sure that it doesn't, you know, dispense medical advice? Is it, um, you know, is there some sort of reasonable standard where they try to filter, but then if you jailbreak it as a user, that's on you, you know, the, these kind of protectionist questions are going to bump up against safety questions

• 45:06Nathan Labenz?

  in, in kind of weird ways. And everybody, of course, will want to frame their interest, you know, this is how it's always done, right? The interest group will be framed as being in the, you know, the safety, uh, public safety interest, even though a lot of times it's much more about, uh, you know, protecting the interest of the, the incumbents in a market.

• 45:30Nathan Labenz?

  So I'm not a, you know, I don't have a lot of predictions around politics. Um, I guess I would be very surprised if there aren't fights about that soon. I would be very surprised if the doctors and the lawyers don't get their way. I would, I would guess that like the higher status a profession is in society currently, the more effective it will be in creating restrictions around the use of language models to do their core stuff. You don't see the same level of, you know, protectionism or organization in like customer service. So the drive-through at McDonald's will probably be like language model powered and nobody's

• 46:13Nathan Labenz?

  going to, you know, nobody's going to stop that, right? Because there's just not that, it's not that powerful of a group there and, you know, Congress is probably just not going to take it up. But, you know, there's a lot of uncertainty there. If it does play out where there is a lot of restriction, then I think what you get is kind of a leapfrog scenario where, you know, maybe.

• 46:41Nathan Labenz?

  Maybe strategically, maybe because it's kind of the only way it can go, maybe just because like the demand is so great and it's, it seems like almost irresponsible or immoral not to. You know, maybe these things just get deployed in places where like there just aren't that many doctors, you know, so you think geez, like, you know, how many doctors per capita are there in like the Democratic Republic of the Congo? Not that many, you know, but people do have cell phones and they do have important questions and maybe they can get those answered and you know, maybe that's just kind of where this, this stuff goes first, right? And kind of works its way up the, the country ladder by like income, you know, more or less as it becomes clear that like it works and you know, it's really not in the public

• 47:26Nathan Labenz?

  interest to restrict this stuff at some point. So eventually that could tip, but you know, could we, is it conceivable that we could get stuck, you know, in like, say the United States or the West or whatever, in a kind of similar spot where we like can't build any nuclear power plants or we like, you know, can't build a second avenue subway? It's conceivable, but I really don't think so. You know, those are fundamentally fixed location. There's some interesting like political science theory around like what allows a state to pop up, you know, and there was recently a great econ talk about this, looking at state formation in locations where grain naturally grows versus where tubers naturally

• 48:11Nathan Labenz?

  grow. And the big difference between grain and tubers from a taxation standpoint is you can store the grain in a central, you know, silo or whatever, but a tuber, like potato, once taken out of the ground, it rots quickly. You can't store it long-term. So looking back in history at the, the time where states were formed, they see that there are a lot more, lot more state formation in the grain areas because there was something that you could tax. Whereas in the tuber locations, you can't really tax the tubers because they rot, so it's pointless. So, you know, they kind of, those locations maintained more decentralization, had less state formation.

• 48:55Nathan Labenz?

  I think there is something similar here where it's like the AIs are, they're not in one location.

• 49:04Nathan Labenz?

  It's really hard to like choke them off entirely. You know, you're you know where the Second Avenue subway is. So you can just, you know, stop people from doing it there. If that's what you want to do as a state. It's gonna be a lot harder, you know, just like we have. You know, there's the great firewall in China, but there's still VPNs in China. There's ways to get around that kind of stuff. I don't see how, you know, absent like very draconian, like super heavy handed measures. I don't see how people are going to prevent individuals from accessing their AI doctor. You know, especially as it might get deployed in other countries, right?

• 49:45Nathan Labenz?

  Like if it's made illegal here and OpenAI is like, well, okay, fine. You know, we're going to respect the law here, but we are going to go make a, you know, subsidiary in Kenya. And, you know, we'll have our Nairobi office and that will be the place where that, you know, gets run out of and served out of.

• 50:05Nathan Labenz?

  People are still going to be able to access that from their devices and their networks in the United States. You know, unless there again, unless there's a super heavy handed regulatory regime on that. So it's just hard for me to see how this stuff gets denied because, you know, it. There's so many little cracks it can get through. You know, people are going to be motivated too, right? I mean. Middle of the night, your kid's sick, you're thinking, you know, it's winter. You're thinking, God, do I, do I need to go or do I not need to go? You know, I really just want to talk to somebody.

• 50:41Nathan Labenz?

  And I could do it at 1% the cost, you know, of going into that thing like people will be pretty motivated to to find ways to access these services.

• 50:55Nathan Labenz?

  So I expect a lot of drama and a lot of, you know, battles in the regulatory space. But it seems like in the end, it's that will be more a question of speed than whether or not people, you know, ultimately can consult an AI doctor. It just seems very hard for that ultimately to be prevented.

• 51:16Erik Torenberg?

  I'm a, I'm a bit more negative or cynical on, on what's going to happen than you are perhaps. I think people are going to go full on panic mode. I think you already see some of the, you know, like with crypto, you know, it transcended the, the tech kind of media sphere into, into mainstream. But the, the, the real concern there was, hey, does this thing really do anything? Are there any use case? Is this just speculation? Is this just a waste of time, waste of money, et cetera? AI is similarly breaking out, broken out of the tech media sphere, but people are, yeah, some people will say, hey, this doesn't really do anything special. Even, you know, the people building it will try to downplay it.

• 51:59Erik Torenberg?

  But then I, I see, you know, my podcast feed every day, a whole number of, you know, kind of mainstream podcasts saying, hey, what does this mean? So like this has already captured the concerns of sort of a, of a media class. And as the thing just gets better and better, I think, you know, those concerns are going to get louder and louder. And we already have, say, like in education, my, my belief is that education prioritizes the teachers over the students. And sometimes they use the language of, of kind of student, you know, benefits in order to justify not allowing school choice, for example, when really it, you know, serves, it serves teachers.

• 52:40Erik Torenberg?

  That's going to be harder and harder to do as it's clear that AI presents all these opportunities. So they're going to try to use, or all these benefits, you're going to try to use the language of equity. But AI will be will be available to all, hopefully. So I think it's going to become the contrast between, you know, supporting the beneficiaries of the services versus the providers of the services. Or is it going to become harder and harder to conflate? But still, ultimately, I suspect that those stakeholders, the doctors, the teachers, the unions, the lawyers, just have so much power that it's, you know,

• 53:22Erik Torenberg?

  they're going to cause sort of greater negative influence on this than we we would want. Can they materially slow down OpenAI? I don't think so. I think what we'll have is just a continuous bifurcated economy. You know, Mark Andreessen likes to point to that chart where it says, you know, here are the industries that are that are affected by government or regulated by heavily by government. Housing, healthcare, education, cost just rises and rises. And here are the industries that are not affected by or not regulated by government. The cost decreases, decreases. I just think that that will continue to bifurcate in a way that's like unavoidable. And you can't, you know, not not see it. Anyway, those are my concerns.

• 54:04Nathan Labenz?

  I think they're going to try, you know, or a lot of interest groups will try. That seems unavoidable. But it kind of feels like, you know, the rise has been so fast. You can imagine a different world where things were much more gradual in terms of just the rate of capability improvement. And we had more time to sort of be like, yeah, this thing could sort of be your doctor, but it's a shitty doctor. And it's like, you know, it shouldn't be out there. But we kind of just crashed through that world in the last year before anybody even knew what was happening. And so we're waking up, you know, collectively to a world where it's like.

• 54:50Nathan Labenz?

  That concern is kind of already over, you know, or it's like very soon to be over. There was just a paper that that somebody sent me this morning that was like GPT four beat Medpom. And then like Medpom two, which also just came out like in the last week or two is like neck and neck still seemingly with with GPT four. But it's basically expert level on answering these, you know, standardized medical questions. It just happened so fast, you know, it's like the first people are hearing of it. It's kind of already at expert level. So I think you're right to say it's going to be increasingly difficult to do that conflation. And it's going to be kind of more and more obvious that it that some of these things are just kind of naked power grabs.

• 55:34Nathan Labenz?

  And then it does remain kind of a question of ultimately like political economy. I do think you could, you know, if you had some heavy handed laws, like you could definitely slow things down. You know, with enough FUD, you know, you could, you could probably scare off, you know, a lot of the consumer population for a bit at least. I don't even want to say a while, though, because like people already have kind of been out of shape about AI censorship. You know, how are people going to feel when it's like, I know this thing can answer my medical questions in a, you know, like pretty good, strong, reliable way, but it's not allowed to.

• 56:15Nathan Labenz?

  And it's not allowed to because of who, you know, because of the doctors, you know, convinced Congress that I shouldn't be able to get these questions answered on my own terms. I mean, we live with a lot of crazy stuff, so that could happen. But I kind of think those curves are going to bend, you know, that I would guess that I would expect convergence. You know, I think costs will ultimately drop even in those government sectors. And for me, it's like a question less of can that line be maintained and more like, you know, how fast or slow might the retreat ultimately be? And by the retreat, I just mean like ultimately kind of accepting that more and more can be done by AI.

• 56:59Nathan Labenz?

  And, you know, at 1% the cost, like, there's a pretty good rationale for it. It seems like we're headed.

• 57:06Nathan Labenz?

  For just, you know, kind of a consistent advance on that front. I'd be surprised, you know, I'd be surprised if they can hold the line, you know, or by day, I mean, like any of these kind of interest groups that might try to say, you know, there should be no direct consumer access to an AI lawyer or an AI doctor, you know, it must be done this certain way.

• 57:29Nathan Labenz?

  I don't think that's going to hold for super long, but maybe a few years, you know, five years, like, wouldn't be, wouldn't be crazy to me if we, you know, are sitting here in 2028 and we're like,

• 57:40Nathan Labenz?

  because we've done, you know, we've done dumb stuff, right? We've put a lot of people in jail for nonviolent crimes. Like, there's certainly moral outrages where it's like, how is it that the, you know, the quote unquote, little people in our society are getting, you know, treated so badly based on some high level conceptual argument about, you know, safety? Now, that, that could happen again, but, you know, we've

• 58:06Nathan Labenz?

  largely backed off of the war on drugs and people largely seem to agree now that, like, way too many people are in jail. And I think this kind of, it's going to be just even way more obvious, I think, in the case of the doctors, like, who's really being harmed here? Who's really being protected? Yeah, hard for me to imagine that it goes five years before people, you know, have pretty good direct access.

• 58:26Erik Torenberg?

  That's optimistic note to wrap on and maybe wrap here, but I'll, I'll try one more kind of line of question, which is this idea of, you know, earlier we were talking about, hey, we're going to need to figure out what, what people are going to do. And it is interesting because the last, you know, decade or couple of decades, we've really worried about blue collar automation and what the truckers, et cetera, are going to do and how they're going to find meaning once we no longer need them. It just, it turned out we, we need them longer than we, we thought we would. But we thought of all sorts of things from, from UBI, but then we said, hey, you know, is UBI really going to fill their sense of, sense of meaning? And I think it's interesting because to the degree that it's coming for actually

• 59:09Erik Torenberg?

  white collar work, you know, people who work in white collar work are probably more sensitive to to feeling useful, to to to to status concerns about their importance, given they how hard they've worked their their whole life and to rise up these respective ladders. Of course, blue collar people work, you know, just as hard in their own way, but white collar people are used to climbing up the ladders of success, being told their whole lives, you know, how how successful they are and how special they are and how important they are. And it's going to be just a huge shift in morale. And e- even like myself, I've started to feel like a little deflated when I see, you know, my

• 59:50Erik Torenberg?

  abilities to synthesize information or summarize or or, you know, create conversations or have certain analysis on certain things, just be dwarfed by by by chat GPT, et cetera. And so thi- thi- it's not a question of UBI as as much as, you know, the resources as much as how are people going to feel important? How are people going to feel needed in in in ways that, you know, fulfill the the ways in which they wanted to be needed or prepared to be needed their entire lives, going through all these hoops, et cetera, of school, grad school, jobs, et cetera?

• 1:00:25Nathan Labenz?

  Yeah, I mean, in some sense, I think it's a another way in which this AI technology in general is kind of a great leveling force. It seems like there's a definite strain in politics that is kind of resentment of one class of people who feel and, you know, say that they got where they are based on merit and their unique abilities, you know, and the wisdom of the market, you know, that values it highly. And, you know, I think that we're all going to be in much more of a similar boat, I think, there, you know, before we know it. So.

• 1:01:05Nathan Labenz?

  You know, maybe we, you know, one optimistic take on that would be like, maybe we can all be a little bit nicer to each other, knowing that we're kind of all under some similar pressures in that respect. You know, that could be in some sense, good. I do think people are gonna, you know, react all sorts of different ways to it. And we've been seeing things like, I just saw something from a 3D artist who was like, you know, in fact, I think you've sent it to me. That was like, you know, I just lost everything I loved about my job. Now I sit here and prompt mid journey all day and it kind of sucks. And I'm like, I don't feel creative in the way that I loved being creative before. There's probably going to be a lot of that. I don't really know that any way around it, honestly, at this point.

• 1:01:47Nathan Labenz?

  You know, the flip side of that is some people are tremendously empowered by it. You can, you know, for every one person like that, you can also go find somebody else. It's like, I've always had great stories in my mind, but I was never able to visualize them. And now I can. So I think you do see, you know, both sides of that trade off. Another idea that I've, excuse me, thought is really interesting is this flip from humans seeing themselves as like distinct from the animals and sort of, you know, it being our justified privilege to rule the world based on our intellect to now there's kind of the reverse of that where it's like, well, what makes us special relative to the AIs is that we

• 1:02:29Nathan Labenz?

  feel, you know, that we have these kind of animal like, not animal like, animal, you know, traits that we at least are pretty confident don't exist in the AIs as they exist today. So our, you know, our feeling, our emotion, the sort of realness of that is, is kind of, you know, what some people are now trying to kind of set up as like the specialness of humans. But it is interesting that that, that stuff is largely shared, you know, with the animals. So, you know, what separates us from the animals and what, uh, you know, puts us kind of in the same bucket as the AIs and what separates us from the AIs puts us in the same bucket as the animals. We're, we're like in the center

• 1:03:10Nathan Labenz?

  of that Venn diagram for the moment. But yeah, I think it's gonna be tough. I do think there's, there's certainly going to be a loss of kind of sense of specialness to the intellect, um, you know, the, the things that we can do. Like I saw another, I saw Tik Tok of this with a doctor sitting there at. At ChatGPT, talking GPT-4, getting like a amazing diagnosis and feeling like, guy's like, "Bro, I went to med school for four years. This thing just spits it out." Like, you know, and he's feeling defeated, like exactly like you said. But what was really interesting to me about that video, I'll see if we can find it, maybe we can put it in the show notes.

• 1:03:52Nathan Labenz?

  Top comments, this was, you know, this video went viral, right? So tons of views on it. The comments had thousands of hearts. The top comments that I saw were all women saying, "Well, maybe at least ChatGPT will listen to me when I go in and talk about my problems." So, you know, I don't know. That's, that's really, that's tough. Like you got the doctor that is,

• 1:04:21Nathan Labenz?

  you know, feeling like you described, defeated, like, "I worked so hard, you know, I was on this path to success. Now like my status is in jeopardy." But again, you see, you're just in that same, you know, one little TikTok interaction. A real pent up sense that like, "Yeah, but you're not, you know, not you, this one doctor." But like, doctors aren't necessarily serving all of us super well. And we do see a lot of upside in an AI that might be a little more patient, you know, that might actually listen to us when we feel like we weren't listened to by, you know, doctors before. So we're gonna have to find something else, you know, I think, I mean, we, relationships obviously is a,

• 1:05:01Nathan Labenz?

  you know, in just about every like wellness, you know, life satisfaction type of study, quality, number and quality of relationships are some of the biggest factors determining how people feel about their lives. It seems like an emphasis on relationship and relationship building is probably a big part of where we might go.

• 1:05:24Nathan Labenz?

  That's again kind of connected, you know, my earlier idea about like highly bespoke, highly idiosyncratic, you know, highly local custom services, you know, that's kind of light commercialization of just like relationship, you know, and kind of community. Some of it, you know, is maybe paid and some of it isn't, but it seems like there's a lot of potential there for people to enrich their lives. As long as Replica doesn't get so good, you know, that that also gets crowded out by language models. And certainly, you know, I think that will be a, an element of the future too, but I don't feel like I have answers, you know, for what happens with society, you know, it just seems like everybody's incentive, we're headed kind of to a new equilibrium

• 1:06:08Nathan Labenz?

  where like everybody has the same incentive and you kind of can't not do it, you know, if, if all of the other doctors' offices in an economy are like now 24/7 scheduling, you know, just to take the beginning of the transformation, like how are you not going to do that? You know, how are you not going to sign up? How are you going to be the one that's like, well, actually, we only take calls nine to four, but it's also more expensive. As a, you know, the good thing is it's less, it's less convenient and more expensive. So, you know, that's why you should continue to choose us. Like, it's just tough, you know, the incentives are all toward rapid adoption and like, I

• 1:06:50Nathan Labenz?

  think we're going to kind of have to sort out our feelings separately from how, you know, the actual structure of how stuff gets done evolves.

• 1:07:00Erik Torenberg?

  It is interesting, like one of the prominent narratives that even, you know, people as, as left wing, as Ezra Klein are promoting is this idea that even liberals are, although well-intentioned, are making the sort of government regulations so complex that we have kind of a, use someone else's word, I can't remember who, a vetoocracy, where it's just so easy to, to, to veto things, to, you know, shoot them down. And I think we'll, we'll see in the next few years, or even much sooner, how, how much power these, these organizations, these relator agencies actually have, where in the face of a Obviously, evidently, you know, better solution for the customer or constituent,

• 1:07:42Erik Torenberg?

  you know, will they prioritize that or will they prioritize the provider or the or the other group that's that's providing it the service even when it's so clearly at the expense? Maybe we'll we'll wrap on on on that. Does that sound like a good, a good, complete stop to the conversation?

• 1:08:00Nathan Labenz?

  Yeah, well, we'll certainly we'll certainly have time to continue it. And in kind of keeping with the the general philosophy of this, you know, the the note to listeners would be like, I think this is an honest take that we really aren't ready for, you know, all of the change that is coming at us. As Tyler Cowen said, like, a lot of copes coming from a lot of directions. This is an attempt to be the sort of cope free zone and, you know, really try to understand the the technology on its own terms, what it can and can't do on its own terms, you know, how it is likely to be applied based on the the strengths that it clearly

• 1:08:45Nathan Labenz?

  has and the weaknesses that it that it still has as well. That unfortunately, at this point in time, does not lend itself toward, you know, tidy answers or like comfortable, you know, high confidence outlooks. And so, you know, we just have to invite you to think about that a lot on your own as well. And, you know, continue to participate in this conversation. I think that's really all we have right now is kind of figuring out as fast as we can where things are going in an environment where there is a lot of uncertainty.

• 1:09:21Erik Torenberg?

  Great to to close on. Nathan, as always, thank you for having a wonderful conversation. And until until next time.

• 1:09:30Nathan Labenz?

  My pleasure. Thank you, Eric.

• 1:09:32Erik Torenberg?

  Omnikay uses generative AI to enable you to launch hundreds of thousands of ad iterations that actually work, customized across all platforms with a click of a button. I believe in Omnikay so much that I invested in it, and I recommend you use it, too. Use Cogrev to get a 10% discount.