B (3:49)

So my reaction, especially watching all the complaining and the underwhelming reactions to it, and I would put this as separate from the kind of terrifying reactions where people miss the sycophancy of GPT of the previous iterations and were able to where it. It always told you that you're right and you're great. I think on this tool calling or which direction should I go? It was really interesting to me because we recorded separately from the Brad Lightcap conversation and Brad and myself kind of ended up having the exact same reasoning as to why this was a different type of intelligence. It's the idea that knowing which tool to call, knowing. Knowing which direction to go actually is different than traditional reasoning models as they existed before. And I still strongly believe that is where things need to go. The idea, and we've debated this for years now, is it going to be one model to rule them all or will there be lots of specialized models and actually from a starting point, not even the model, the actual, the platform or the system or the product knowing where to go to solve a specific problem is where this has to go. Now, GPT5, I think there's definitely some, like, uncertainty over how it was rolled out and what it looked like and how people are used to using ChatGPT. But I actually think. I think this is still where it's going. And I. I'm actually happy because it felt like removing a little bit of the overall kind of hype cycle and actually letting people take a breath, to me, is actually the single best thing I've been hoping for for a long time. So I'm still pro GPT5 release.

A (5:38)

Okay, well, I'm gonna spend a good chunk of the beginning of this episode trying to get you to change your stance, because. Go for it. I have so many issues with this. And by the way, the first thing that you just said, you said that, okay, it is letting us take a breath as opposed to overhyping. I mean, in what universe, Sam Altman was again hyping this up, saying GPT5 is much better than anything he's used before. It's better than most humans at almost every task on Theo Vaughan, basically saying it was everything that we imagine AGI will be without saying the words. And then it came out and it was clearly not AGI. So let's just go. Let's go right there, right now, and then we're going to go back to this switcher thing. Um, is this really a break from the hype or was it just that, like, is the break actually that the model just doesn't live up to it?

B (6:33)

No, I think the. The break from the hype is it did not live up to the hype, and I think that's good. So I. I agree what Sam Altman was promising on Theo Vaughn and wherever else, this definitely is not that. But why I'm saying I think that's a good thing is it's going for a moment like GPT5 has been so built up for so long and now everyone realizes we're not going to get some magical model release that just solves everything all at once. And then AGI comes and robots take over and all white collar jobs are eliminated or whatever else. I'm saying the fact that it did not live up to the overdone hype, that's a good thing. Because now. And we'll get into the Eric Schmidt piece, we'll get into this MIT 95% of projects don't show ROI study. Like, like, I think now everyone can get back to the real issue is how can we actually build to solve problems.

A (7:33)

And so I'm going to actually get to that issue in a moment. But I still want to talk about the disappointing nature of the model. And then I'm going to get to the things that I don't like about this agent idea, even though I was bought into it maybe a couple weeks ago when we talked about it. So look, I mean we've talked on the show for a long time, right? It took them like what two years took OpenAI two years took to go from GPT4 to GPT5. And we said for a long time, just name the next model GPT5 and don't put all this pressure on yourself to release a God model. And so finally, okay, so I think there is something to be said for like, all right, they got the number out there, right? So we'll get GPT6 probably next year. We don't have to keep like, you know, playing up what the next model is going to look like and anticipating that it's going to be a big one. I think Thomas Wolfe from Hugging Face in this FT article about whether AI is hitting a wall, put it really quick, put it really well. He says for GPT five people expected to discover something totally new. And here we really, we didn't really have that right. So basically here's the thing. They spent two years working on this new model. They built it up to be, you know, God level. It wasn't a new thing. I don't even know if it's better. I don't think it is better than for. Than 03. I'll just say one last thing and then turn it right back to you. Ethan Malik, the Wharton professor who quote here often had a very, very good post about this right after the release that I think got left mostly overlooked. He said the issue with GPT5 in a nutshell is that unless you pay for a model switching and know to use GPT5 thinking or pro, when you ask GPT5, you sometimes get the best available AI and sometimes get one of the worst AIs available and it might even switch within a single conversation. So he's basically saying like at the high end you might get this GPT5 high, which I imagine thinks a lot, or you might get this GPT5 minimal and you don't really know which one you're getting. So like you're not, if you're a regular user of this technology, you might actually feel a downgrade happening when you, when your queries get pushed off to this minimal thing versus like you knew you, you Might have to wait a minute or two, but you're probably getting really good stuff O3 and I think the fact that this doesn't feel like an upgrade from O3 is a very, very big problem.

B (9:57)

Okay, so, so on this idea of like where is the intelligence, where's the value? I want to dig into that a little bit because to me and like for listeners, two weeks ago I was talking about how I work for a startup called Writer now and we had launched something called Action Agent, which is very similar in scope that it's. There's a number of predefined tools just from a prompt knowing where to go and what to do with it. To me, the more I've used tools like this, the more I see that's where things need to go and the way to think about it is. And agreed, this is a new type of intelligence, even within the ChatGPT ecosystem. So it's gonna be rough at first, but it's where the world is going. Like if you think about, you know, what, what did Apple promise us like with a prompt even just looking up your flight info and then going and doing like a flight status check. The reason they've not been able to roll that out yet, it seems simple, but the intelligence there is knowing which system to call, what to do, what tool to call next, what action to take next. That type of intelligence is where all the promise of AI is and we have no, no one has done yet at scale. And this is the first, this is like the beginning of this new way of intelligence versus within the model itself. It's going to do a bunch of thinking, it's going to go into its own like training data, it might do a little bit of web search and then give you a result. So, so this is just the beginning of this whole next phase. And so of course it makes sense that I guess I'm going to give Sam Altman a little credit here. The fact that they actually still rolled this out, it knowing it's not going to be mind blowing, but it's setting the stage for where AI intelligence needs to go and then maybe one day Siri will actually work and be able to check your flight status just by asking it.

A (11:53)

Give me a moment here to sort of try to make the case for why we don't need AI agents. So I'm going to try to blow a hole through this entire concept.

B (12:04)

This is the hottest take of all time in this.

A (12:06)

Yeah, you defend and I will be on offense. I have been using GPT5, you know, as Many people have. But my specific use case over the past two weeks has been very intensive. I've been using it all throughout this trip that I've been planning, including trying to figure out like the amount of altitude I should be increasing day by day as I went up the mountain in Nepal. And um, of course like, you know, didn't spend my entire time going up this mountain hooked to technology. But you want to make sure that you're not going to die on the next stop. So when you're at one of these tea houses, you connect to WI fi and say, all right, how far should I go the next day? And I felt like the AI was transforming from, and excuse me for using this term, but from a thought partner to an overeager helper. And not a particularly useful one either. And we had talked about this idea that like GPT5 just does stuff, which was the title of the Ethan Malik piece after he had been using it. And GPT5 is just trying to do things for you. And I think that like we can see AI go in two directions from here, or maybe it does it both simultaneously. But to discard the thought partner type of use case for this overeager helper, you know, agentic use case to me really seems misguided. Like I think, you know, some of the best ways to use AI is to help you round out concepts in your head or like to say, I'm thinking these things, what do you think about them? And then have it come back with like a pretty interesting and considered answer that, you know, might hallucinate sometimes, but ultimately helps you sort of expand your mental capacity because you're just able to think about these concepts with higher horsepower. And I think what GPT5 has done is it sort of deprioritized that use case for like trying to go do stuff for you. And like there are times where I really craved having a deeper conversation about the climb or whatever it might be. And instead what I got back was just like this bot coming back over and over and over again of like, let me make a card for you that shows your turn back times and stuff like that. I'm like, I don't need that. What I need is something to be able to really explore these concepts and thoughts. And I felt that that has been deprecated in this new model. Now maybe that's a selfish criticism, but I'm going to just throw it out there. And I don't understand why this agentic thing, which may work really well in like the when is my flight and we want it When's my flight coming? For that to be the direction and the route and the rule for AI moving forward. I just question whether that's the right move.

B (15:05)

Well, no, so I don't disagree with you here. I think these are two different actually I think that's like a good breakdown. When do I want a thought partner? When do I need to do things? And they are very different use cases and I think like OpenAI's mistake here is they're trying to jam it all into one product and, and again that intelligence of which way how should I interpret this? And I agree like I had a bunch of those where it would let me make you like a presentation when I just wanted a quick answer. I think that's it should be two separate products. I want to do stuff, I want to think and have a thought partner. They're very different systems that are required for those or they need to nail the intelligence of which way to route it at the beginning. And I think to me the biggest issue here is GPT5 is essentially like some whole new GPT1 in this way of intelligence, like in this mode of intelligence we're talking about around taking actions, calling tools, doing stuff versus I'm going to call my training data and maybe do a little bit of web search and I'll give you a nice text answer or an image or something like that like in reality. But, but, but I'll also push back that like web when web search was introduced to all of these tools that is agentic it decides to go search the web because it doesn't have the training anything in its training data to answer the question. And we all have become so used to that even as you're trying not to die and asking and having chat GP depending on ChatGPT to keep you alive to understand your oxygen, your VMO2 max or whatever it is. I'm not a hiker by any standard. I think like we've already, the agentic has already been happening quietly. Web search was the big one. Web scraping and operator like computer takeover hasn't really worked well yet but web search was always agentic and it are. It was this exact flow that we're talking about. So, so we're all, we're all using it. It's just that this isn't amazing yet. This is the V1 of this new way of working and they rushed it out, didn't explain it clearly and conflated these two very different ways of working and thinking into the same tool.

A (17:33)

Right. And so my Concern is that they are going because these companies have a need and this has always been the concern with raising the amount of money that they have raised. They have a need to replace and augment labor.

B (17:47)

Right.

A (17:48)

And so that's where you get into these agentic use cases. My concern is maybe that's not the most useful way that we can use AI and we might end up in service of returning the money to investors, lose some of this, like really useful capability that we've seen so far. And let me push back on the web search part of it. Is it agentic? Yes. But what is it agentic in service of? I think, well, maybe in some ways it is like, you know, find me the opening hours of this, like, you know, of like restaurants in New York City that I'm. These are, these are my criteria. So like in some ways there's that hygienic stuff, but it's also like these are ultimately they are filters of knowledge. And the cool thing about it is they have not, maybe not all, but almost all of the knowledge that's ever been written within them. And web search might be agentic, but it's also just like a way to update their mental models and update this filter with the most current information or stuff that's not in their training data. So yes, that's agentic, but it's also in service of, I think this use case that I've found very valuable. Now as I'm saying this, I'm like, maybe I am shortchanging the do stuff part of it a little bit, but I am mourning a little bit the direction of. I'm mourning a little bit the loss of the old direction. And I find it especially interesting that OpenAI added back the 4.0 model which you talked about, which was like very friendly, it's maybe a little sycophantic, but not.03 because it thought too much or maybe not O3, which was the reasoning. The thinking model has now been replaced by GPT5 thinking. And I wonder if they didn't add it back because it's just too expensive to run and they can just run this version cheaper. Which again gets to this big question we've been talking about of like, was this old direction just financially impossible for the industry to make work?

B (19:58)

Yeah, hold on. I think two separate issues there and I'll get into the financials of OpenAI and how much of this is done to actually meet some kind of valuation criteria, which it certainly has to be. But I think that first part, I, I think if we're going to break it down. It's thinking versus doing. And using it for thinking is. It has been very good. And they're kind of screwing that up actually from like a user interface standpoint. I mean, they really are screwing this up because, like, everything keeps saying thinking. Like the fact that it's. Maybe they should just change that word to doing and it'll start to at least be a little more realistic about what's happening. So, so, so to me, I think even more as we're talking it, these are two very separate things people need to do. The doing side of is going to be. That's how we get to the next. Like, that's the way we actually realize some genuine value from all of this. And so, I mean, whether it's like the labor conversation or is just asking Siri to find your flight info and do something with it, like contact Delta and change my flight time or look up alternative flights, whatever it is, you know, like, there's really, really simple things that should be, should be better. And I think that's the way things like it has to go there, otherwise the entire industry implodes. But also, I think it can work. This is just the first iteration of that. But they hyped it up too much, I think. On the OpenAI financial side, I'll give Ed Zitron had written a good piece on who I generally disagree with on how this is financially driven a bit because it's going to go try to take a cheaper route, but it should, in the end, O3 should not be running when you're trying to rewrite this email for me, you know, like, like different cases should require different models, different tools, and getting that intelligence in place. If anything is ever gonna be financially viable, it has to work that way anyways. So, like, yes, is it financially driven? Sure. But it, but it needs to be like, I, I won't hold that against them. Do I think, like, that this downplay of hype or this backlash could hurt them significantly? Sure. But I'll still give them the directionally strategically correct decision.

A (22:33)

Okay. And I will, I will say strategically incorrect decision. And I think this might be a new debate that we have on the show. But there's also, there has been this argument we should, we should talk about this, that this has been so bad that all of a sudden everyone's running out and saying, yeah, let's stop talking about AGI anymore, or, you know, where we are, we might be in a bubble. And that is what Sam Altman did say in his dinner that he had with some reporters. He says this from cnbc, quoting the Verge. He said Altman said when bubbles happen, smart people get overexcited about a kernel of truth. Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes. And this is from the CNBC story. His comments add to growing concern among experts and analysts that investment in AI is moving too fast. What do you make of this? I mean, it's kind of interesting to get reporters together in a dinner after the release of the somewhat bungled release of your most important model ever and, and say that we're in a bubble now. Maybe the headlines made a little bit too much of it. But, but this idea that like there's a kernel of truth, even though he does say that AI is going to be the most important thing that we're seeing right now, but kernel of truth has led to, you know, some sort of bubble. I don't know. The implication is somewhat concerning. What do you think about it?

B (24:13)

He, Sam Altman knows how to get a headline better than anybody else. Like, I don't know. I mean, will I agree with. Are we in a phase where investors as a whole are overexcited? Is AI the most important thing to happen in a long time also? Yes. Like, I'll agree with that. I feel this. I don't know. I'm going to. This one feels like him just kind of saying stuff, knowing how to kind of get a rise out of reporters, get the headline versus there's any genuine thought put into this statement.

A (24:45)

Well, I think that if you take, you know, maybe we'll just take it as a data point. Right. I don't think Sam got reporters together to try to like, let everybody know that AI is a bubble. I mean, he's currently, I think, just, just raised a huge amount of money and he's probably right. Trying to raise again. And by the way, we have another headline that Anthropic is also raising. They're raising Intox to raise now $10 billion. And the last report was a $5 billion raise at $170 billion valuation. The valuations keep going up, but now we're starting to see talk of maybe this is a bubble from Sam Altman. I think Mustafa Suleiman had a story this week saying, hey, let's not call AI conscious. And I really want to sit on this New York Times op ed from Eric Schmidt and Selina Hsu about why the US needs to stop talking about AGI and why this is going to set the country back. Because it is somewhat remarkable that Schmidt has gone from someone who, like, gave a TED talk who said, we're underappreciating how important AI is to this, now pouring cold water on AGI. He says this is his njoo's op ed. He says reaching artificial intelligence, or AGI, is now a singular aim of America's tech giants, which are investing in tens of billions of dollars in a fevered race. It's uncertain how soon artificial and general intelligence can be achieved. We worry that Silicon Valley has grown so enamored with accomplishing this goal that's alienating the general public and bypassing crucial opportunities to use technology, to use the technology that already exists. It's being solely fixated on this objective. In being solely fixated on this objective, our nation risks falling behind. China, which is far less concerned with creating AI powerful enough to surpass humans and much more focused on using the technology we have now just to we're going to get into this argument of like, should we just use the tech we have now or focus on building AGI? But just a very interesting data point that Schmidt and Zhu add in their story. They say in a recent survey of the association for the Advancements of Artificial Intelligence. By the way, who's that? They call it an academic society that includes some of the most respected researchers in the field. But this is interesting. More than three quarters of the 475 respondents said our current approaches were unlikely to lead to a breakthrough. While AI has continued to improve as the models get larger and ingest more data, there's concern that the exponential growth curve might falter. I mean, it's just interesting to see Schmidt saying this in the pages of the Times and Thomas Wolfe from hugging face saying this to the FT and Sam Altman saying there's a bubble and Mustafa Suleiman saying, you know what, let's stop calling these things conscious. Everybody together is coming together and talking about this AI hitting a wall or reaching a curve, reaching the diminishing returns point of the curve as opposed to the exponential. And so I'm curious, Ranjan, what you think about the context of where AI is heading overall, given that this is what we're hearing? Material security is transforming how companies protect their most critical cloud assets like Google Workspace and Microsoft 365 with modern purpose built security that actually works the way people do. The biggest cloud threats walk through three doors, email, identity and data. Material was built from the ground up front. Day one for Google Workspace and Microsoft 365, not as a retrofit. So it closes all three. It provides continuous protection before, during and after an attack. You detect problems early, contain them fast and recover without chaos. Lean security teams scale through intelligent automation rather than adding headcount material blocks, evasive phishing and impersonation. It protects sensitive content using built in rules for personal health and financial information, or rules you define. It also spots risky applications, unsafe settings and sketchy verification or password reset attempts, all without slowing people down. So get the overview at material Security. That's material security.

B (29:03)

I mean I could not be happier. Listeners can't see how much I'm smiling right now that Eric Schmidt's on team product, team build. It's not the model. Everyone is coming around right now saying let's take the technology we have now and actually try to figure out how to build with it. I think like again, his whole point that like in China people are focused on actually integrating AI into hospitals and farming and other areas of life and medical diagnostics and like that's what we need to be doing. That's what we have been saying for, for a long time. And the kind of over focus on this one model to solve all of our problems. If it kind of feels like that's going away and I don't know that I'm excited by this, we can actually focus for a moment.

A (29:54)

So I read this story and I just had Ranjan in my head because this is almost making the Ranjan case, you know.

B (30:03)

Are you saying I didn't before and now only Eric Schmidt can make my case?

A (30:08)

I'm saying, I'm saying I think, I think Schmidt did a better job making the product case.

B (30:15)

I did not get my times up, Ed, but maybe soon, if you're listening.

A (30:19)

But, but I do think that he's taking this ideological torch from you, Ron John, and running with it. So I mean he listens obviously. So he does. He does. Well, we might get him on the show. Might. Anyway, so he listen, but listen to this. Let me just read it. Because he's talking about it in juxtaposition with China and I think I haven't heard it made. I haven't heard the case made made this way yet. So he says let's look at what's happening in China. The countries and scientists and policymakers aren't as AGI pilled as their American counterparts. They're talking about a deep integration of AI with the real economy. While some Silicon Valley technologists issue doomsday warnings about the grave threat of AI. Chinese companies are busy integrating it into everything from the super app WeChat to hospitals, electric cars and even home appliances. In rural villages, competitions among Chinese farmers have been held to improve AI tools for harvest. Alibaba's Quark app recently became China's most downloaded AI assistant, in part because of its medical diagnostic capabilities. Last year, China started the AI plus initiative, which aims to embed AI across sectors to raise productivity. This is interesting. It's basically very. You know, there's another stat here saying how like the vast majority of folks in China have said that AI has had a positive impact on their lives, whereas in the United States, the vast majority have not. Recent poll said 32% of Americans say they trust AI, compared to 72% in China. Over three quarters of adults in China said that AI has profoundly changed their daily lives over the past three to five years. That share is the highest globally and double that of Americans. I don't know, maybe it's Team Eric and Ron John that have come together to sort of maybe make me believe in the product side of things over here. But I mean it is interesting because it's like maybe we, the AI industry has reached this point where, you know, if AGI was realistic then that would be a good strategy, but if it's not, you have to sort of like get going full speed as opposed to focusing on developing it. And I think that's the core message here.

B (32:37)

Yep. And I would actually add one part to that. He brings up a really interesting point because we've talked for a long time around AI has a branding problem and like instead of focusing on what it's actually doing, meanwhile like everyone has or a large percentage of the population in the US uses AI in their day to day life probably is deriving significant value. Probably writing all their emails, using ChatGPT. Like people are even. I mean when we talk about AI, like having it alter your photo to look better, all that stuff is AI. So, so people are using it. But instead, I mean the way like just the term itself, you have kind of two groups of marketing. You have like when Doge was going through, everyone was like they're using AI to cut jobs and make decisions or just like that very negative connotation or the conversation around job displacement. But then on the other hand you just have really bad marketing. Like remember the Google Gemini commercials? Matthew McConaughey like it, just like it. The industry as a whole has not communicated to people this is how your life has already changed because of AI and like, and people aren't even processing it or realizing it. So I think that was a really telling stat. If 75% of adults in China are saying it has a positive effect, Whereas in the US it's only 32%. I think that's more around how the, the figureheads, the industry, the marketing has just been bad.

A (34:13)

Right. And I think you're only getting this op ed. If Schmidt, who's close to this, who's. I think he was the first investor in Anthropic, he's been watching AI since he ran Google. If he came to the conclusion that we're nothing close, we're not close to AGI, then it is time to make this pivot. So I think that's important context and I think that's a good point. If you take that along with what Altman Mustafa Suleiman Thomas Wolf is saying, you're like, oh, shoot, this might be that wall now. Again, I'm not. We talked earlier, like, were there going to be overreactions to GPT5? Yes. Is the fact that AI is done reaction to GPT5 an overreaction? Yes, of course it is. This is still very powerful technology, even if it stops here, only improves incrementally. But it is interesting that they're all saying it now. And I think that you're totally right, Ranjan, that like there is a branding problem and I think there's a building problem. Right. This gets again, back to your argument. I love it. Which I'm coming towards it because I think it's been underemphasized, I'll put it that way. And there's a lot of truth to the position you've taken. Here's again from this op, Ed. It's paramount that most people outside Silicon Valley feel a beneficial impact of AI on their lives. AGI isn't a finish line. It's a process that involves humble, gradual, uneven diffusion of generations of less powerful AI across society. Instead of only asking, are we there yet? It's time we recognize AI is already a powerful agent of change. Applying and adapting the machine intelligence that's currently available will start a flywheel of more public enthusiasm for AI. And as the frontier advances, so should our uses of the technology. But being too fixated on artificial general intelligence risks distracting us from AI's AI everyday impact. We need to pursue both.

B (36:15)

God damn it. He just, he wrote it a lot better than I've said it.

A (36:20)

I know, but I think that the conversation here has been building. And you know what? I think it's a good point. And I've been on T model for a long time. I've been on go build those better models and everything will take care of itself. Reading it in this way does lead me to believe that. Yeah, I think this focus on GPT5 and focus on the next model has sort of put the US at a disadvantage where it hasn't been building it into everything because it's just expecting a God model to come in and fix everything. And that doesn't seem to be happening.

B (36:57)

Yeah, I think the more we're talking today, I think this is a huge inflection point. Like GPT5 was such a big idea like that hovered over the entire industry for so long and now it's out there and it's underwhelmed everybody. And now everyone's kind of recognized this because I definitely. There's this study that came out from MIT this week where it said 95% of AI projects within companies have not seen any kind of value. And I actually, I think this, this is near and dear to my heart now working in the enterprise AI space, I'm in all these kind of conversations, but really what I believe has been happening is over the last few years that expectations, misalignment has been a huge problem that people are hearing. Everything can do these gigantic massive things and don't even start small and just start learning how to use it and integrate into their companies and just, just getting AI fluent. But I think also this ties back to like the Schmidt op ed, the GPT5 backlash. Like could this, do you think this is a moment where everyone's gonna take a breath realizing the last two to three years, call it two years of actual implementing AI, we've had the wrong approach and actually become a little more sensible and start, start it's, it's. Well, maybe we'll get another. It's time to build OP ed at the right time.

A (38:30)

I don't think so. I'm. This is why I don't think so because of the funding. Now think about the amount of money that's gone into this, right? Like the industry needs AGI. They need this AI that can build these like super powerful tasks to be able to justify the valuations that they've gotten. And you're thinking about like the biggest VC rounds ever. These models need to improve more than they are because they still get things wrong, they're still unpredictable. Is part of the reason why the 95% of projects within enterprise have failed? Is it part of it? Because of Misaligned expectations? Yes. Is part of it because the technology isn't good enough? Yes, I think so. It's still, it's, it can't. We had a reader write in, so it says it's still, it's too unpredictable to be useful in many corporate settings. And I guess like you could say, yeah, well, you've been using it for sales and marketing and it's just not good for that. But even in back office cases, like if these things start, like, I don't know, doing jobs 95% right, but getting 5% wrong, you're going to end up with like a very bad situation for, for the companies that, that have implemented it because there can be errors.

B (39:44)

Well, I, I would push back that in our current human led infrastructure, 95% accuracy in a lot of spaces is not, is not actually great. But, but, but I think one thing, like actually, so I was reading and it was the study, actually it was by the amazingly named Nanda Group at mit. One of the other things it found though is it said the biggest problem the report found was not that the AI models weren't capable enough, although execs tended to think that was a problem. Instead, the researchers discovered a learning gap. People in organizations simply did not understand how to use the AI tools properly or how to design workflows that could capture the benefits of AI. Like so, so to me, this learning moment, people having misaligned expectations and approaching things the wrong way, to me that has to change. And now it feels like maybe in the last two weeks we're hitting an inflection point where it will.

A (40:47)

But here's the thing. Who's which AI company is going to go to these enterprises and say, so about that AGI thing we've been telling you about. You know what, forget about that for now. Let's go with these much less ambitious projects. Even though we've been telling you that like, God, AI is coming around the corner. I think the advantage that China has, aside from like some central planning that maybe forces companies to do this, is that the models there are open source. Yes, they cost money to run, but there's typically cheap to run. And so you're looking at just like basically paying the inference costs and that allows you to like put AI in your refrigerator or your briefcase or whatever it is, as opposed to here where like the companies are selling ambitious projects.

B (41:37)

I agree. The way the funding has been structured at the giant players is definitely kind of rubber meets the road moment's going to happen where they, they're going to have to decide. And even what you, what you said right there, like in your refrigerator, you know, like the promise of using computer vision to see what kind of food's in there and when it's going to spoil. Like this stuff should be happening. The technology is there, the models are there and it hasn't yet. And it's like, felt like we've just been all waiting, so. So again, I agree there's going to be some genuine issues from like how the, the funding structures of a lot of these companies. But. But again, it has to happen at some point. Like you can't just keep promising it unless you believe the technology. You're still team God model and GPT6 will actually move us in that direction. So we can all just kind of wait for our refrigerators to know what food's going to start spoil and not actually GPT5.

A (42:43)

Yeah, I mean, I'm not going to say that. I predict that OpenAI was going to call GPT5AGI and clearly like that was wrong. So I'm not going to say GPT6 is going to be it. Although OpenAI is talking about GPT6 having a lot more memory, which I think is like a really good way to take this technology to really get to know people and to remember things about their lives. That's good. But this MIT study is, you know, fairly brutal. This is the report. This is from Tom's Hardware. The report said only 5% of AI pilot programs achieve rapid revenue acceleration. The vast majority stall and deliver little to no measurable impact on profit and loss. The findings are based on 150 interviews, a survey of 350 employees, and analysis of 300 public deployments of AI. And that 90, these 95% do not hit their target performance. But yeah, because generic AI tools like ChatGPT do not adapt to the workflows that have already been established in the corporate environment. Yeah, this is, this is going to be an issue.

B (43:47)

No, I think for me, like this one, this one hit home hard again, like this is my day to day life now. Learning and talking about and working on enterprise AI. The. I mean, it's real. It's definitely real. The way a lot of people approach these AI pilot programs again, was assuming everything will just work perfectly. I don't have to like fix my existing broken workflows and I'm just gonna layer chat GP on chatgpt on top of it and everything's just gonna work perfectly and it's gonna take no effort. Like, I think that this has been another big area is people approach it as typically software deployment where it's a very, very different thing. That's something I'm probably going to write about soon. But like, it's just so real that I think the way the entire industry, most organizations have been thinking and approaching it over the last two years, we are really hitting an inflection point now. And GPT5 might be that canary in the coal mine, but I think it has to change if we're going to let our refrigerators know when our food's going to spoil and you're not going to die climbing on the Nepalese mountains.

A (45:02)

Well, I'm still alive with the help of mostly folks at tea houses and I didn't follow ChatGPT's advice. I went a little bit higher than I should have and things were fine. A little ibuprofen goes a long way from your altitude sickness headache. But, but yeah, I just to go back again to this why the product matters, why the model matters. Yes, it would be good if like, you know, this sort of rhetoric around AGI started going away and, and you know, companies started putting this stuff into place. However, like one way you can solve a lot of these problems again will be better models. But the question is, is this technology fundamentally unable to do it? I don't know. Okay, but before we end, we definitely should talk about one place that AI is making its way into the workflow or apparently might be. According to this new report from Rest of World, and that is OnlyFans. This is from Rest of World. A hidden network handles chats. For OnlyFans stars, AI could soon take over. That's from the story. Artificial chatbots are starting to take over from low wage workers known as Chatters who impersonate only fan stars in direct messaging with fans. The adult website's creators rely on these remote operators to flirt with fans, earn tips and sell images and videos. Chatters in the Philippines, a hub for this work, told Rest of World that rising sales quotas have made their work more stressful. One said his company plans to replace the worst performers with AI. The Chatters believe that once AI fully masters sales, their jobs could be automated. But for now, the bots cannot impersonate human quirks fully. They said these people are using keyboard smashes and intentional misspells and Gen Z slings. Gen Z sling and I can't do it do that yet. Well, Roger, maybe this is it. This is if you got to start with product, maybe it is the only fans chat. By the way, wasn't it always obvious that the people on OnlyFans were definitely not speaking to the models themselves. Isn't this always like on the Internet? No one knows you're a dog situation. But now it may be going to AI. So maybe there's hope for the product side.

B (47:21)

I don't know.

A (47:21)

What do you think this is?

B (47:23)

It's the product. It's an attainable use case. Leveraging the technology and as it stands today to deliver value in a concrete fashion. It's there, it's real.

A (47:36)

Apparently what happens is the, the when chats come in, these groups of people and often in the Philippines are directed by bosses on Discord to like go out and chat and they monitor their activity. And you know, one, one of them, the story has one of them freezing because there was this like emotional message and the supervisor got mad. And they know that AI will, you know, won't have that. That issue. I mean, I don't know if the technology is good enough yet to replace these human chatters, but I hate it.

B (48:13)

I hate saying this one. If you are an only fans like Chatter, the fake chat person, that it's a job that's going to get displaced by AI out of all of them. I disagreed with Dari Amadeus, 50% of white collar workers or whatever it was. But I'm gonna have to go with this one probably now.

A (48:39)

Now what if it's more than the chatters? What if it's the only fan themselves?

B (48:43)

It's gonna be.

A (48:44)

I mean, because here, this is, this is another part of the story. Automation and only fans is now moving beyond chat. Using AI management. Companies that represent models can generate photos of them in poses requested by subscribers without any human involvement whatsoever. AI images generated by one of these companies using Stable Diffusion reviewed by rest of world were so realistic they could not be distinguished from photographs.

B (49:09)

It's funny because the whole kind of like narrative or meme around like that was written by AI. That's an image generated by AI. And it's obvious to all of us, I'm convinced like a lot of content out there is already quietly generated by AI. Like you had those kind of big splashy PR driven efforts and though. But in reality behind the scenes, like, because those are the companies that don't want to advertise it because like the sheins of the world or whoever else, like if you look at their product pages does not look too natural. So yeah, I think it's already there.

A (49:49)

Let me put it this way. We talked at the beginning of this episode that there's like two main uses for Artificial intelligence. One is this thought partner thing. One is this agent thing. Kind of left out the other one, which is this sort of therapist slash companion.

B (50:06)

I thought you were gonna say thinking versus doing and I didn't want to go there, so.

A (50:10)

Oh, God, yes.

B (50:12)

Because that's where we start.

A (50:14)

Let's move on from this idea. Let's go with the talk about. Yeah, there's thinking, doing and then there's friend or companion. And it's sort of interesting that like with the Asian stuff, you move away from that. And we also like, there's this HBR article that we've talked about that's circulating out there that says like the number one use case for AI is companionship and therapy. And, you know, I don't know, do you. Don't really need AGI for that.

B (50:42)

No.

A (50:43)

Just need better memory. And by the way, that's what they're building towards with GPT6 memory. This is where this is. It's all just going to head this way. Your agent dream is going to die, man.

B (50:52)

It's just. I know, I know.

A (50:53)

My dream of acting your agent dream, my thought partner dream, both going away. It's just going to be AI. Yep.

B (51:02)

Let's. Let's just get ready for it. It's where it's going.

A (51:05)

Are you excited for this Future?

B (51:07)

And somehow OpenAI will actually realize its valuation with just companionship. I mean, maybe that's not an unreasonable thing, like getting people, like the elasticity of demand and just jacking up the price. I think if it's companionship is probably a lot more valuable than. What's my oxygen maximum on a while hiking the Nepalese mountains.

A (51:34)

I think you're right. I think that's where it's going. And then we'll merge. We'll merge with the AI and have a companion slash AI lover at all times in our head. And let's just all war will end and peace will come in our time and people will be kind to each other and that. That will be true. Super intelligence is the understanding that we're all one as a species and as anything carbon based.

B (52:04)

The real AGI was love the whole time.

A (52:08)

Exactly. Well, I'm glad we solved it, Roger. I'm glad we found a kumbaya moment here at the end. All right, man, thank you so much for coming on. Great, great speaking with you as always.

B (52:21)

All right, see you next week.

A (52:23)

All right, everybody. Speaking of merging with AI, we have two brain computer interfaces coming up. The next brain computer interface episodes coming up. The next two Wednesdays, next Wednesday, we'll speak with journalist Sally Adee about the state of the brain computer interface. And the following week, we have the leadership of Precision Neuroscience coming on to talk about their bci. Ronjan and I, of course, will be breaking down the news each Friday. So stay tuned as we cover next week's headlines next Friday. And thank you all for listening. We'll see you next time on Big Technology Podcast.