# Jev Creator on Decision Models, Benchmark Fatigue & Why the Labs Missed It

Markdown copy of https://realalpha.app/v/ham8dROHCrc/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Diogo Almeida argues that AI needs reliable, machine-native decisions before it can automate more routine work.

- Show: Unsupervised Learning
- Hosts: [Jacob Effron](https://realalpha.app/people/jacob-effron/)
- Guest: [Diogo Almeida](https://realalpha.app/builders/diogo-almeida/) is creator of Jev and cofounder of TypeSafe AI.
- Published: Nov 5, 2026
- Length: 1 h 1 m

## The best minutes

- [17:57](https://realalpha.app/v/ham8dROHCrc/?t=1077) [Reliability matters more than inventing new model shapes](https://realalpha.app/s/reliability-matters-more-than-inventing-new-model-shapes/) (4 min)
- [26:55](https://realalpha.app/v/ham8dROHCrc/?t=1615) [The KV cache constrains how coding agents are designed](https://realalpha.app/s/the-kv-cache-constrains-how-coding-agents-are-designed/) (8 min)
- [52:49](https://realalpha.app/v/ham8dROHCrc/?t=3169) [Coding agents may speed engineering by 20 to 40 percent](https://realalpha.app/s/coding-agents-may-speed-engineering-by-20-to-40-percent/) (3 min)

## Also worth watching

- [1:16](https://realalpha.app/v/ham8dROHCrc/?t=76) [Most model calls should serve software, not people](https://realalpha.app/s/most-model-calls-should-serve-software-not-people/) (5 min)
- [21:55](https://realalpha.app/v/ham8dROHCrc/?t=1315) [Semantic checks could make agent software safer](https://realalpha.app/s/semantic-checks-could-make-agent-software-safer/) (5 min)
- [34:39](https://realalpha.app/v/ham8dROHCrc/?t=2079) [Multimodal models are less reliable than text models](https://realalpha.app/s/multimodal-models-are-less-reliable-than-text-models/) (3 min)
- [48:52](https://realalpha.app/v/ham8dROHCrc/?t=2932) [The labs may be strongest at financial engineering](https://realalpha.app/s/the-labs-may-be-strongest-at-financial-engineering/) (4 min)
- [55:49](https://realalpha.app/v/ham8dROHCrc/?t=3349) [True AI agency may require two major breakthroughs](https://realalpha.app/s/true-ai-agency-may-require-two-major-breakthroughs/) (4 min)

Skip the discussion of model categories, the launch and benchmark recap, and the fundraising exchange.

## Full analysis: Diogo Almeida says coding agents speed engineering 20 to 40%

- Almeida estimates coding agents speed engineering by 20 to 40%, depending on task and user.
- He says reliability takes more work than new model shapes, while benchmark hype can erode developer trust.
- Teams should test the speedup in their own workflows, counting review, rework and side effects.

Its 7 sections, what it means for you and what to watch are on [the page](https://realalpha.app/v/ham8dROHCrc/#analysis), free with an account:

- [4:12](https://realalpha.app/v/ham8dROHCrc/?t=252) Software calls should serve machines, not readers
- [18:27](https://realalpha.app/v/ham8dROHCrc/?t=1107) Reliability matters more than new model shapes
- [5:23](https://realalpha.app/v/ham8dROHCrc/?t=323) Developers know the reliability their workflows require
- [27:44](https://realalpha.app/v/ham8dROHCrc/?t=1664) The KV cache limits some agent designs
- [37:47](https://realalpha.app/v/ham8dROHCrc/?t=2267) Almeida says image capability trails text sharply
- [55:14](https://realalpha.app/v/ham8dROHCrc/?t=3314) Almeida estimates a 20 to 40% speedup
- [57:21](https://realalpha.app/v/ham8dROHCrc/?t=3441) Almeida guesses true agency remains breakthroughs away

## Summary

Almeida coauthored OpenAI’s InstructGPT paper and spent two years building Jev in stealth. He is Jev’s creator and a cofounder of TypeSafe AI.

Almeida argues that AI’s gap is not a lack of raw intelligence but a lack of reliable systems that can turn model capabilities into useful software. He says model makers should focus less on new model shapes and benchmarks and more on reliability, while developers should push providers to address specific weaknesses. He sees coding agents as useful but estimates a 20 to 40% speedup on average, with results depending on the task and the user. He says safe deployment requires attention to side effects and believes true AI agency is still about two major breakthroughs away.

### Minute by minute

- [1:06](https://realalpha.app/v/ham8dROHCrc/?t=66) **AI needs machine-native software.** Diogo Almeida says he was motivated by the contrast between AI’s apparent intelligence and its limited effect on productivity. He argued that software should work backward from machine-native model calls: computers usually need bits or structured decisions, not prose. He says Jev is not yet at that goal.
- [6:34](https://realalpha.app/v/ham8dROHCrc/?t=394) **Reliability drives Jev’s roadmap.** Almeida says Jev’s developers need to judge what reliability their applications require, so the company exposes model weaknesses and gives developers more control. He says real-world feedback is still limited and developers should push providers harder to address model “jaggedness.”
- [10:54](https://realalpha.app/v/ham8dROHCrc/?t=654) **Models should be provider agnostic.** Almeida calls it a bad outcome if developers must build systems around fixed assumptions about particular models. He wants model behavior to become reliable enough to work across providers, like SQL. He says Jev should be a small, composable software primitive, while LLMs remain useful for answers that need not be correct and for throwaway software.
- [15:53](https://realalpha.app/v/ham8dROHCrc/?t=953) **Reliability matters more than shape.** Almeida hopes LLMs—or something better—will remain useful, but says people overuse them and should keep a range of models in their toolbox. He expects more model shapes, but says creating a shape is easier than making it useful and reliable. He criticizes benchmark-driven launches for encouraging overpromising and making it harder to identify genuinely useful models.
- [21:13](https://realalpha.app/v/ham8dROHCrc/?t=1273) **Software and models must co-evolve.** Almeida says progress requires both better models and education that helps developers build more capable software. He points to coding agents less constrained by the KV cache, semantic dependent types that verify properties expressed in natural language, and fast semantic checks on tool calls as promising areas.
- [26:55](https://realalpha.app/v/ham8dROHCrc/?t=1615) **The KV cache constrains agents.** Almeida says coding-agent design often follows the KV cache’s append-only state, which limits routing and makes subagents’ context costly and difficult to manage. He describes alternatives such as agents sharing context, live visualizations of another agent’s work, and task-aware memory and compaction. He says the cache is useful; the problem is treating it as the only design.
- [34:39](https://realalpha.app/v/ham8dROHCrc/?t=2079) **Multimodal models need more work.** Almeida says multimodality tends to reduce reliability and expects text tasks to be automated before image tasks. He describes Jev’s image model as insufficient for production and estimates it is about 10% as far along as the text model. He is especially interested in video games and reliable computer use.
- [39:24](https://realalpha.app/v/ham8dROHCrc/?t=2364) **Older models could automate easy work.** Almeida says building Jev five years earlier would have been harder but possible. He argues that GPT-3.5-era models, and possibly GPT-3, were already smart enough to automate many easy tasks. He cites GPQA performance around 2023 as an example of strong measured ability that had not translated into enough real-world automation.
- [41:55](https://realalpha.app/v/ham8dROHCrc/?t=2515) **Almeida says benchmark releases burn trust.** Almeida says similar releases from other labs risk damaging trust when they are optimized for benchmarks rather than real-world use. He argues that lower prices alone do not expand demand: a product must also create enough value for people to use it more.
- [44:04](https://realalpha.app/v/ham8dROHCrc/?t=2644) **Almeida prioritized helpful investors.** Almeida says Jev can run profitably and that the company did not need its recent funding to operate. He says the goal was to bring in valuable people who could help build the business, rather than subsidize the service to create lock-in.
- [45:23](https://realalpha.app/v/ham8dROHCrc/?t=2723) **Reliability beat benchmark hype.** Almeida says the team considered releasing a benchmark-focused version but chose to prioritize reliability. He is glad they avoided another overpromising launch, though he says he cannot know whether it would have attracted more attention. He says developers’ fatigue with benchmark-heavy releases came sooner than he expected.
- [48:11](https://realalpha.app/v/ham8dROHCrc/?t=2891) **Almeida sees financial engineering as labs’ strength.** Almeida says coding agents can be valuable but questions whether their current use justifies the money being spent. He says he is not informed enough to judge Anthropic’s valuation, and sees financial engineering—not models, product, or distribution—as the labs’ greatest strength. He also worries that some expectations assume AI will deliver far more than it has demonstrated.
- [51:52](https://realalpha.app/v/ham8dROHCrc/?t=3112) **Deployment depends on managing risk.** Almeida says coding and math have safeguards, such as human review and version control, that can limit unwanted effects; many other kinds of knowledge work do not. He estimates coding agents deliver a 20 to 40% speedup, depending on the task, the user’s skill, and how well side effects are contained. He says some uses enable work that otherwise would not happen, while others are as slow as doing it by hand.
- [55:49](https://realalpha.app/v/ham8dROHCrc/?t=3349) **Agency remains far off.** Almeida separates risks from ASI, bio, cyber, and trust and safety, and says reliable systems with true agency are still about two major breakthroughs away. He distinguishes AI agency from people using powerful tools to cause harm. At the close, he urges builders to regain a sense of agency and rethink software and interfaces.

## Transcript

Unsupervised Learning, hosted by Jacob Effron, with Diogo Almeida.

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

[0:00](https://realalpha.app/v/ham8dROHCrc/?t=0) As you think about Anthropic, is that something you're optimistic on? It's like, hey, that's undervalued or something you're like, oh, I don't know what's going to happen to that. I get the sense that a lot of the expectations baked in are ones that are not just product level, but based on like misunderstandings of the future of AI and what it can do. So my guess is that because some people might be valuing like building God is basically infinite, there will be some amount of miscalibration there. And I think that, you know, like the let them cure cancer is going to be like Marie Antoinette's let them eat cake.

[0:37](https://realalpha.app/v/ham8dROHCrc/?t=37) Jeb took the AI world by storm about a month ago. And I've been wanting to talk to Jogo, their CEO of TypeSafe, for quite some time. So it was really fun to have him on the podcast. I think maybe even better with a month to reflect on everything that's happened because we talked about a ton of different things. It was really awesome to get Jogo's perspective on the shapes of problems that he still thinks need to be solved by AI models, how he thinks about pushing JEV models to be even better. It was also fun to get his thoughts on Anthropics' upcoming IPO, AI safety, a bunch of other topics. I think folks will really enjoy hearing from him.

[1:09](https://realalpha.app/v/ham8dROHCrc/?t=69) Without further ado, here's Jogo. Well, I've really been looking forward to this. I think ever since your viral launch, I feel like the second that happened, I was like, I have to have you on the podcast. And so I appreciate you making this happen. I think actually, in many ways, even more interesting that were some weeks removed from it. Given that the early results were not what you expected and maybe not as strong as you were hoping, what gave you that kind of early confidence in this research direction of RLCD? I mean, I feel like you kind of alluded to it in the GPT-3 example. It seems like there has to be a leap of faith that any great research effort undergoes. And what was it about this approach that just in the early days when you didn't have the signals early on, you were still kind of confident in? Oh, man. Oh, man. Oh, man.

[1:55](https://realalpha.app/v/ham8dROHCrc/?t=115) this is like i'm getting a little emotional right now um i have a fun little uh side topic that gets into this if that's okay my co-founder was debating on whether or not i was mission driven at open ai i would have described myself in hindsight as one of the lowest percentile mission driven because i think that the open ai mission is not really um it's kind of intentionally left as vague so many people's missions can fit into it but like the grand mission was never something i super believed in even though i truly want ai to be extremely extremely transformative so i'm still trying to figure out like am i was i mission driven then or not because i'm really mission driven now but the answer was you know just a lot of work trying to square like ai being super incredibly awesome which i think it is and ai being really not that useful in every like productivity stat around and this is like the kind of thing that like most people like just focus on one or the other and they never had like a bridge answer or like they never have a sensible bridge answer i think distribution is like a nonsensical answer here um because so many companies have gone all in on that and then gotten burned um it's it's almost certainly not a

[3:05](https://realalpha.app/v/ham8dROHCrc/?t=185) distribution and to me it was like believing that like hey ai is clearly smart and also that the future is clearly going to be machine native the question that always motivated me and like i used to ask this question even to early people joining or even to investors in our early, you know, like our early rounds or even like the precede round, which I don't even know if that counts, like taking a safe. I used to ask this question of like, OK, like you have imagine LLMs are an API, you know, like a like a sub module that is called by software, which even in the most string model ends, they still need to be called by software by something like a harness.

[3:43](https://realalpha.app/v/ham8dROHCrc/?t=223) **Diogo Almeida:** that harness can be a while loop. But what percentage of those calls going into the LLMs are actually meant for human consumption? So for consumption, so things like style matter,

[3:54](https://realalpha.app/v/ham8dROHCrc/?t=234) and, you know, pros and all of that, versus for computer consumption, which is going to be very, very different, you know. Like, I don't think a computer really ever wants a string, except when that string happens to be an extreme cardinality enum, which is a very valid use case. But it's very rare that that is the thing that computers want because they operate on bits. And when I just ask myself or anyone this question, even the most string-filled of people will give many nines of, like, percent for the machine-native component. It's just a question of when that will happen. And to me, it was always like, okay, what does that thing look like?

[4:27](https://realalpha.app/v/ham8dROHCrc/?t=267) let's work backwards from that cognitive core and make it work. And I will say JEV is not yet there. There's still so much more work to do to make it so much better. And that's what really drives us all. Is JEV not yet there? Because I mean, you know, today, the shape of like, you know, calls and tokens actually is more for humans than it is for computers themselves. How much of it is just you're also waiting for the use cases themselves to evolve? And how much of it is, you know, areas you want to push on the model i think it's a bit of both it's it's like this interactive dance and this is also why we released like these dreamer ready features like the things that we are you know like wouldn't recommend for the most stable of use cases because we really want to be honest about our model's properties so we've released a lot of these things because the reality is that the people who best know how much reliability you know robustness like intelligence is needed for their downstream application is not us, it's the developers themselves.

[5:26](https://realalpha.app/v/ham8dROHCrc/?t=326) Because developers are the expert at this and they will know way better than anyone on our team. We have smart people, but they will know better what reliability profile they need as long as we give them the right knobs, which string models don't tend to do. So there's an entire spectrum of this. I think that they need to co-evolve together. The reason why we launched was not like, okay, this is like a good moment in AI. Everyone told us not to launch then, just an FYI. It was that we felt like, hey, this is getting to be something we're really proud of like it's kind of like impressively smart at a lot of things if you prompt it right and we are ready to like keep pushing that limits that those limits with the rest of the field to help us co-evolve use cases figure out weaknesses deal with jaggedness like there's so much we do to focus on jaggedness and insufficiency and intelligence that having people play with it would be so helpful why'd folks tell you not to launch um we launched around the time of the pacing the frontier discussion and people were like hey like it's kind of like a bad vibe to like launch a capabilities product during this like everyone's like really doom and gloom about

[6:27](https://realalpha.app/v/ham8dROHCrc/?t=387) ai and it's a kind of a bad vibe and like i don't know if you guys should do this and we're just like that discussion never really made sense in the first place um or it is at best incomplete in my opinion you talk about kind of you know wanting to push in in so many directions i mean i think one thing you know that you guys do is you're always kind of transparent about the things the models can't do, you know, areas where they're not as strong. How much of that is, like, the roadmap for what you're pushing on? Like, you know, how much of it is other areas? Like, you know, how do you think about, like, where you want the company to push in the next month or two?

[6:58](https://realalpha.app/v/ham8dROHCrc/?t=418) Month or two? Oh, that's a... I mean, in AI world, I don't know if I can ask longer than that, right? I wouldn't say jaggedness is purely, like, our roadmap. There are some things that models truly are weak at. And, like, this is where, like, that taste stuff comes in, right? Like, when I was talking about like the reality of what you can extract from these models the reason we call them system one models and among other things and system one actually people did not want us to name that it either because system one thinking tends to be a bad thing like people use that as kind of like an insult right it's like the the part that just naturally happened from evolution you know from back in the day before uh exactly instinctually are much larger brains yeah exactly but and like people often like say that that is the mistake making side and i believe that the shape that actually ai gets compressed to happens to be like very system one and therefore the honest thing to do is to communicate its weaknesses because i think that on its own can be incredibly incredibly valuable for software and most of what people need It actually is system one stuff. Why can we solve Millennium Prize problems in math, but not automate anything touching accounting? That should be really weird to a lot of people. And I think that that is still our North Star, even if people get distracted by cost, speed, like, oh, it's good for really fast decisions or whatever else. I see these as distractions compared to what is really needed for AI to meet its potential, and that is my North Star.

[8:28](https://realalpha.app/v/ham8dROHCrc/?t=508) And so obviously you put the models out there and you have a lot of people using them in different ways. How much does like the real world usage of the models come and feed back into the roadmap versus just like the stuff you all are planning to do? I would say it's almost none real world feedback. I think people should be complaining to us so much more. I think that like the field is still in a Stockholm syndrome of like, oh yeah, the model providers just do stuff in their one track path. And we happen to be like, the victim is the wrong word, but like um you know like in uh what's that movie wally you know like they're like they're just like the people in the chairs that are just getting fed the ai and i think this is like the state of things right now where like people plan around ai changing but they don't have a sense of agency of where to take the field and i think that this is not the way it should go that's actually really disempowering and kind of lame in my opinion and also i think that there's so much more we could do collaboratively as a field. And I think that the field could push us a lot more to say like, hey, we want to be doing these things. We think it is within the realm of capabilities. Let's

[9:34](https://realalpha.app/v/ham8dROHCrc/?t=574) figure out how to do it together, but not just putting stuff into the model, but figure out how to prompt them, figure out how to unearth capabilities. Let's work together for these sci-fi use cases, because I think that is the way the future is going to go. And I would want so much more of that. My joke internally is that the number one way to get me to respond to something is to talk about model jaggedness because you have like very high up people say like, hey, can we meet with Yogo? And I'm just like, no, can you meet with my CTO or GTM head instead? But if they complain about model intelligence, I am like on that and like debugging that myself for better or for worse, because that I think is the thing that will really move the needle for us long-term.

[10:15](https://realalpha.app/v/ham8dROHCrc/?t=615) So right now we mostly have to go with our old roadmap because we are not, I think people are still in that mode of like, Like, yes, let's just do whatever the model providers let us do instead of, like, let's collaborate on this, like, insane future together. And actually, there's another aspect of that. I think people have a deep distrust of model providers, too, because they train on their data. And, you know, like, this whole, like, discussion of, like, can GPT wrappers even exist, I think is very disempowering, too, because the whole application ends up becoming a chatbot. So the moat ends up becoming the equivalent of a system message that you send every time versus, like, the moat and the intelligence and the behavior and the specialization actually existing in software where we just make little micro decisions for you and you do all of the sick logic.

[11:00](https://realalpha.app/v/ham8dROHCrc/?t=660) Yeah. So there's so much to do. There's so much to do. No, I think your point's really interesting because I feel like when Jeff came out, there became maybe like a common set of patterns or rules that folks were like, oh yeah, like this is a really good use case for Jeff and you should, you know, try it out in this thing or that thing and everyone tried it out and, you know, kind of got that sense of patterns. You know, I almost feel like people are trying to get their own mental models of like, okay, I use like, you know, Opus 5.5 for like this kind of thing. And I use Jeb for like this kind of thing. And I feel like your points, well, you know, your point is like, this is a dynamic evolving thing that you could, there's not strict boundaries, I guess, between those two types of model categories, but like, what's your ish, ish. I actually think that what you're describing is a little bit sad to me. And I, you know, I understand that like, we're playing a role in this. I do not think that that is the most developer centric worldview like for people to start building like theories of mind of the sub models they call and have their systems locked into that is i opinion, a really, really bad thing. I think the North Star for intelligence is going to be a lot

[12:04](https://realalpha.app/v/ham8dROHCrc/?t=724) like SQL. Like, we want to be so reliable that it's actually provider agnostic. We obviously want to the best models, but like, we don't want it to be locked into like, things that only work with Jev. So what is your, you know, mental model for how this all evolves in terms of, you know, this like, thing people are doing today of like, you know, hey, you know, Jev works really well for these things. Maybe like, there's other sets of models that have been not like, to your point, And RLVR works really well for code and math. We'll see what else it ends up working really well for. I have to interrogate that gesture.

[12:35](https://realalpha.app/v/ham8dROHCrc/?t=755) What was that about? Do you want to interrogate it right now? Yeah. Well, really well is quite relative. As far as I can tell, even the best companies still require humans in the loop for all of this stuff, despite superhuman results in some axes. and that still, like, people should have an answer to why that is the case. Like, if they are so unbelievably good at software, like, why do normal software engineers have anything to do? Not that I want them automated, but, like, they clearly aren't and there's clearly something missing there and I think thinking of, like, intelligence as this single axis that is measured by whatever the benchmark of the day is is, like, totally wrong and there's, like, obviously nuance missing there because the people who go all in on not understanding that nuance get burned super hard so i think that rlvr is incredible but he's incredible at human in the loop stuff still and i i think that wisdom comes from understanding reality where it's at knowing its weaknesses and like this is we are obviously use agents a lot but we really value understanding its limitations you know i just wouldn't say that it's like oh super great and it just solves everything personally maybe i'm using the wrong models

[13:48](https://realalpha.app/v/ham8dROHCrc/?t=828) I think a lot of our listeners are trying to figure out, you know, like a simple mental model maybe of Jeb, which I know you disagree with is like, you know, hey, there's some set of tasks or some set of things I'm trying to do that will make sense for this. There'll be some set of things that make sense for other things. And I'm sure 95% of our listeners right now are trying to just like adjust their own mental model of like, OK, what am I like Jeb use cases? When do I use Jeb? that's a very pragmatic take for the short term and i'm in favor of it this is like where it's about giving engineers power because like we shouldn't be overly opinionated about what they use it on and if they need to do things to make it work abso-fucking-lutely hell yeah like if you need to do like the worst dirtiest hacks to provide value in the world like you know like we are not going to stop you uh like that is awesome and also if you have to do too many hacks that's still our fault you know like like in a deal where you wouldn't need to do that and the more it becomes close to like semantic understanding the more beautiful the code is the more reusable it is the more you can plug it in into places that you wouldn't expect it to it's kind of like

[14:47](https://realalpha.app/v/ham8dROHCrc/?t=887) the unix philosophy of like doing like incredibly small useful things but because people have a mental model of it they can compose it together into big ones i feel like this is great and also like we should not be like recreating regexes we should be recreating like beautiful primitives not that i hate regexes i just don't like reading them you know and like you know like it's like It's like more, what's the word? Is it write-only code type stuff? And we want to be beautiful, composable. We want to build on layers like LLMs have failed to do so. And we want stuff to be more like software. I think I have an okay mental model of software engineers at worst.

[15:25](https://realalpha.app/v/ham8dROHCrc/?t=925) It's kind of weird that people even want to update to the latest and greatest models. I hate updating my dependencies. Updating dependencies is the worst thing ever for production software. It's fun for personal use. I used to use Arch Linux. I get it. um and also like why would i want to bear the burden of that and the only reason people keep doing this treadmill of like maybe the next model will fix my problems is because i don't think they solve those problems so well that they can run in the background fuck off for literally ever and maybe cut you know revisit it every like couple of months to see like is there a cheaper or better alternative available but it works fine right now i don't i don't of anything like that.

[16:01](https://realalpha.app/v/ham8dROHCrc/?t=961) And that is a sad state of the world. Like, it's very counter to how software engineering tends to work. I guess in your vision of the world, like, do LLMs have a place in the world in, like, 10 years? Super duper. Absolutely. I hope if that they do not have a place, then something strictly better exists. I'm not opposed to LLMs. I'm not actually opposed to either chatbots or coding agents. Completely different things. And I really don't want to go to a world where they don't exist. The thing is, I think people overuse them for a lot of tasks. And I actually hope the market for that increases too as people get better at it.

[16:37](https://realalpha.app/v/ham8dROHCrc/?t=997) And I'm basically certain that there's tons of use cases for it at basically every single intelligence cutoff that has existed, maybe since GPT-4. All of these models had beautiful, awesome use cases that maybe would get Pareto dominated by another model, but they should be able to run in production. And it takes a lot of nuance and understanding. and it's like really not built for engineers. So I love LLMs. I hope that I can continue to like chatbot anything that I have like, I need like pseudo expert level answers to that don't need to be correct. That is the use case in my mind for chatbots.

[17:14](https://realalpha.app/v/ham8dROHCrc/?t=1034) And I love being able to like spin up, like throw away software instantly when I want custom dashboards and stuff because I think it's the best use case for coding agents. And that should stay for the rest of time. I don't think people should be doing that. and we should keep on building our toolbox, so to speak. Like, I think it's just awesome. Is your intuition that the toolbox will include, like, many other shapes of models in the future with, you know, other RL acronyms? Or if we're talking 10, 15 years from now, like, are there, you know, are there a bunch of different, like, tools in this toolbox?

[17:44](https://realalpha.app/v/ham8dROHCrc/?t=1064) Or is it kind of like your approach and the LLM approach cover most of the use cases? I'm going to end up saying something spicy here. but I would say when I first when we first launched this I was really excited for novel shapes to emerge um really really excited and I still think that that will be the case I think new shapes are freaking awesome we plan to keep shipping them I'm a little bit worried that again what we talked about maybe this was like before we got on the record that ML people especially have like kind of like lost the actual message i was trying to say um and i'm worried that people are now over focused on like the shapes of tasks rather than the reliability the shape of task is incredibly you know like it's super easy to do that when i thought this would take a week i thought it would take a week to make the shape of the task which it did and this is what's happening now with other people making similar models with similar shapes but

[18:40](https://realalpha.app/v/ham8dROHCrc/?t=1120) **Diogo Almeida:** the shape really doesn't matter compared to actually it being useful and that is the thing

[18:45](https://realalpha.app/v/ham8dROHCrc/?t=1125) i think people really really need to focus on so there will 100 be way more shapes than these two even if no one else does it we will do them i think people are making too many new shapes right now and not caring enough about you like the utility of it and i'm worried that this will actually burn the goodwill that people people don't like being uh over promised under delivered to and like these things are like pure over promised under deliver and i think that that's like not good for the world you know like if people have like creative playful energy we don't want them to waste it on stuff that clearly doesn't work. As far as I can tell, people dislike that. And even if they dislike it, I hate it. Like I, oh man. And I think there's going to be a balance. There will definitely be a lot more future shapes to be had. And I just really, really hope that people start giving like unbelievable levels of fucks for utility, usefulness, reliability. What would a developer want in this situation rather than ml people who i think focus on like i want to flag plant you know like i want to be the one who names the thing that gets cited i want to like get um revenue in the short term because people will play around with it and i want to do the same thing that every other ai company does

[19:57](https://realalpha.app/v/ham8dROHCrc/?t=1197) I really hope we don't do this, but overpromising. under deliver in order to get attention because it feels like a lot of the ai economy ish is really driven by that attention and that is so so bad for developers it's like actually the absolute worst for the developers who do not have like the time to test these things themselves because they just get they're the ones who get swept up in all this and this is also why i'm anti-benchmarks it's a it's a it's so bad for those people you know you have to play around with some of this stuff yourself to have any sense of of how good it actually is yes and like that is tragic you know like it makes sense um like i actually think that as far as i can tell people have stopped caring about like benchmarks for our class of model maybe i'm like in the wrong um uh niche for this but my sense and there could be survivorship bias is that people already don't care about new releases here because they've all been lied to so many times.

[20:55](https://realalpha.app/v/ham8dROHCrc/?t=1255) And this is good for us as a company, but bad for the field because actually if someone makes a genuinely good thing, there should be a way to acknowledge it and shout it out and say like, hey, this is awesome. But like in the world where everyone optimizes for benchmarks and demos and things that don't require liability, it's very hard to communicate that other than this very manual process of evaluating models themselves, which is actually super duper hard because benchmarks also don't capture that. One thing I'm struck by, you know, is obviously this idea of like these shapes of problems, right?

[21:27](https://realalpha.app/v/ham8dROHCrc/?t=1287) And the chatbots are very good for some shape of a problem you're trying to solve. Coding agents, like you said, very good for some shape of a problem. Obviously, folks have found JAV really helpful in different shapes of problems. Are there other problems out there that you're still like, God, there's just nothing out there today that's good for... So many. What are like some that come to mind that like of these future areas? Hi, I'm Anne Maris and I'm the producer of Unsupervised Learning. Thank you so much for listening to the show. I love working on this show and I have the challenge of hitting 35,000 subscribers by the end of the year.

[21:58](https://realalpha.app/v/ham8dROHCrc/?t=1318) If I don't hit that, I'll be fired. So please subscribe. Thank you. Bye. What are like some that come to mind that like of these future areas? What I realized is that a lot of people are still stuck in like fairly basic patterns and we have like a gigantic job to both push the frontiers of model intelligence, reliability, all of these things that we really care about. as well as education to push the limits of software. And there's a thing in the intersection of education and pushing the frontier forward, which is this collaboration between research and product, which is, I think, where all gigantic ML wins come from.

[22:37](https://realalpha.app/v/ham8dROHCrc/?t=1357) And this, I think, is kind of an unexplored place. If we had more help, we could do so much more here to encourage people in the world or even fund the kinds of things that we think should exist and really allow people to push the frontiers of what software can possibly do. So I'll give a couple of examples of this. And, you know, this is going to be a little bit of a spoiler for a launch, a release we will have sometime in the future. It is that we want to, like, fund these things. And there's so many cool shapes of stuff missing. Like, one of the things people know I'm passionate about is coding agents that are free from the tyranny of the KVCache.

[23:14](https://realalpha.app/v/ham8dROHCrc/?t=1394) I think that so much coding agent is built around the KV cache. And the KV cache is freaking awesome, too. I don't want to toss it away, but I want to be aware of it and design around it and make different sets of trade-offs to make things that were pure sci-fi possible. Another thing that I think would be so unbelievably cool is... Do you know what dependent typing is? Dependent typing is a lot about adding constraints to your programs that are like higher level than purely the type system. You know, like you can make sure that your function returns an int, but dependent typing would allow you to do things like it is an int greater than zero.

[23:53](https://realalpha.app/v/ham8dROHCrc/?t=1433) And then now downstream applications can use those properties to have their own sets of guarantees, et cetera. To get higher and higher levels of composability with your software, because they're more legible, and they don't require as much unit testing edge cases, and also as much human understanding to understand the whole system. And one of the things I think would be so freaking cool would be a dependent, a semantic dependent typing, where instead of having to represent your outputs as pure types, like int, float, you could actually say specific properties of that in natural language, because there's so many properties you could care about in software, and have a type system that uses JEV-like models verify that everything works together.

[24:33](https://realalpha.app/v/ham8dROHCrc/?t=1473) I think that that would be so unbelievably cool. Maybe it doesn't matter in a world where everyone uses coding agents, but I think that we should do stuff like that just because it's awesome. And that's what early internet energy is about to me. So security is a huge thing as well. As these models get cheaper, faster, all over the place, where can we add extra levels of security kind of to avoid if you guys saw the news about like the cohere quen 3.8 next flash debatably makes a sketchy call to like phone home it's debatable i'm not going to talk about whether or not that's legit or not because i

[25:08](https://realalpha.app/v/ham8dROHCrc/?t=1508) **Diogo Almeida:** haven't read enough about it but why can't we just have checks on top of every single tool call that actually are semantic to your organization rather than whatever anthropic thinks should go

[25:20](https://realalpha.app/v/ham8dROHCrc/?t=1520) into their auto router you know like maybe in this part of the code base they shouldn't be making any calls to this other part of the code base at all and how do you add these checks uh you know you need like extremely fast and cheap security all over the place which could be super awesome you know i'm not saying everyone should build it right now but i think that we should collaborate with you know the making pushing capabilities and software in order for like thinking it's so interesting because a lot of those problems like the problem you're talking about now like you could imagine you know approaching from the perspective of a model company or like there's all these you know enterprise agent security companies that are popping up right that aren't model builders themselves but they're like yeah let's put in some like you know uh some actual like deterministic software to to monitor these things and it's interesting to like think about you approaching a similar set of problems but more from like a model perspective uh versus obviously there's plenty of other folks that are just trying to build software that you can put around in lll so my philosophy is that we should enable those software players to win and just

[26:18](https://realalpha.app/v/ham8dROHCrc/?t=1578) software is so fucking cool exactly but ideally they also tell us that we don't want their data but they should tell us like hey you have jaggedness on this particular axis that makes our thing less reliable let's work together and actually they could just tell us like we are not going to tell you what we work on at all here's the rough shape of this task fix it and i would just be like hell yeah thank you so much for telling us this hopefully it solves your problem and a whole bunch of other people's problems i hope you produce tons of value in the world like i hope you make like a decent amount of money and i hope that this enables other people to build all sorts of other things because now this jaggedness thing is healed and that is the relationship i hope to have and this is why i'm like so so jazzed about software no i love it you said something there that i wanted to come back to because it was just a great phase i think you said that the tyranny of the kv cache in uh encoding agents maybe explain for our listeners a little bit about like some of the limitations uh encoding agents you know due to the kvcache and like how you think about that problem space i would say almost all design

[27:16](https://realalpha.app/v/ham8dROHCrc/?t=1636) decisions around the ux of coding agents like the stuff that actually functionally changes stuff is built around the kvcache and this is natural because like the biggest wins in ml are always like very hardware aware um this is how things have gotten so efficient recently um or pseudo recently um since gpd4's days gpd4 is very expensive like you just need really need to design around this which is great because it's efficient but it's also quite limited it this is why like routing doesn't tend to work routing absolutely works if you're making the equivalent of like um what we call instruction following calls like you know task in task out you know fuck the cash throw it away uh routing is amazing for that actually you can do way more advanced things than just routing in that case.

[28:01](https://realalpha.app/v/ham8dROHCrc/?t=1681) There's so many advanced patterns you can do when that is your assumption. Like, you could actually try it with a cheap model, use an extremely cheap judge, and if it's sufficient, keep it. If not, try a larger model. Or, you know, try multiple times and pick the best. Like, there's so many things you can do. But when you have, like, this stateful KV cache, you are, like, really constrained to it being kind of append only, and you have this complicated trade-off, kind of like an iPython notebook, where you can never delete the memory because it's, like, a really weird, you know agglomerated state and it turns out that it's very hard to engineer around this because the models are trained with it so if you actually start doing fancy stateful operations with the models having free reign over that state they'll end up looping and stuff like that because they're so tuned to having that kvcache and so efficient to have that kvcache so this prevents you from doing like routing effectively it prevents sub-agents from being really effective because now you have this weird trade-off of like what state what context do i pass to my sub-agent because this requires intelligence in order to know what to pass in.

[29:01](https://realalpha.app/v/ham8dROHCrc/?t=1741) If you pass it all in, this is a huge input token cost instead of a cash read cost. And when your subagent makes an output, how do you know what to pass in? And like what subset is relevant down the line? And the safest thing to do tends to be to append a lot of stuff. And now this is the equivalent of like from a state management perspective. Imagine your state is like global variables.

[29:20](https://realalpha.app/v/ham8dROHCrc/?t=1760) **Diogo Almeida:** It's passing one giant dictionary around forever until the models break and they start getting dumber over time because context fraud is a real thing.

[29:28](https://realalpha.app/v/ham8dROHCrc/?t=1768) And this is totally cool, just to be clear. I don't believe that there's anything super tragic about this on its own because it's very useful. I think what's tragic is that's the only way to do it. And I think there's so many more advanced patterns that you could do if you could assume that I could just put intelligence all over these operations. And that is the kind of thing to me that I think from an engineering, architecture, design, beauty, power perspective is so limiting. And that is what that tyranny is, because the people who have tried to do those cool algorithms have mostly failed because they've tried to use LLMs for it.

[30:05](https://realalpha.app/v/ham8dROHCrc/?t=1805) And LLMs are not optimized for, you know, routing logic within the software or making discrete little decisions. They're optimized for making human pleasing outputs that are not super reliable running as a sub process. And this then makes people believe this doesn't work. And actually, a lot of what our original shape looks like, it's not too dissimilar from LangChain. There's nothing wrong with LangChain's interface, right? Some would even say that it was maybe perfect for classifiers on top of LMs. I think, I can't remember back in the day, but I'm pretty darn sure it's going to be pretty darn good if someone made a LangChain-compatible wrapper on top of our model.

[30:46](https://realalpha.app/v/ham8dROHCrc/?t=1846) The problem is people didn't like it because it made models somewhat dumber because the models are not optimized for that type of pattern. And this is also a form of tyranny where everyone just thinks Anthropic and OpenAI or whoever else is like number three right now is like doing their best and giving us the best intelligence possible when the reality is there's so many other intelligences possible. And like there's trade-offs there, you know, like there's truly a lot of trade-offs that are made in this design that people can feel but cannot really measure because it's really hard to do so. What's like an example of the shape of what a coding agent could look like, I guess, with like a different intelligence design?

[31:27](https://realalpha.app/v/ham8dROHCrc/?t=1887) One feature that I think would be really sick is what if you could have like multiple agents sharing the same context? And you're like truly like read writing from the same context. That sounds really complicated. Like there's no way that works right now because of the KVCache. But like nothing stops that from happening. And actually a very common pattern people have with agents these days is like they're just cooking and crazy stuff they don't know what's going on and after it does the thing you asked it to do people are like make a visualizer or a dashboard or a learning tool and it's kind of crazy that that has to be like an extension of the content That has all of the trial and error in it. Or starting from scratch, looking at a git commit. Why can't and rereading everything from scratch? Why can't it share the state? And why can't you have like an agent here? That's actually basically read only for the code base, but allows you to write the visualizer, such that as the agent is cooking, you actually get live visualizations of maybe, you know, using like a slightly dumber model, like Sonnet instead of Opus to write out what Opus is doing. So you kind of get a sense of what's going on.

[32:29](https://realalpha.app/v/ham8dROHCrc/?t=1949) instead of having to read the gibbetsing updates or whatever else. I think that would be cool. Another example here is compaction. Compaction makes a lot of sense for a write-once kvcache that is append-only because, like, oh, it's full. We need to restart it. But compaction is a form of compression, and compression is just not totally solvable. Like, it's very hard to compress without priors for compression. But what happens if you compress based on what problem you were solving before. And now you want to actually ask a problem that is different from that one. And it would use information that was compressed away.

[33:06](https://realalpha.app/v/ham8dROHCrc/?t=1986) Like that agent is like Delta, very much losing hand. But what if how the state gets compressed was actually semantic and it could detect, hey, I'm doing a new task. I should actually think about what the context should be for this new task using an ROI calculation, which should be in code. The code should compute expected value of like cash reads and cash writes and everything and recalculate perhaps a new KVCache for this other task and maybe even keep the other one hot. I don't really know. This sounds like cool sci-fi. One more, one more, then I will stop on this topic. But I'm just so excited about this.

[33:40](https://realalpha.app/v/ham8dROHCrc/?t=2020) One thing that I think would be really cool is, hey, I keep cooking on the same code base and I keep reading the same files and having to re-understand my system's abstractions over and over and over again. You know, there's a partial solution to this, which is like maybe the agent should write markdown files, but then how do you keep it in sync blah blah blah you know it's a kind of a pain you know but what if you could look at the history of old models but like not as like a like i feel like right now it's like a little bit of a hack like oh it's a tool that you can use sometimes if the model deems it as necessary but what if like by its nature one of the first steps you do is even a extremely cheap map reduce over all history to find the relevant stuff here and what if that history could also be shared with your co-workers who are using the same code base so you can actually have insights from what they're working on like these are very doable software problems right the problem is that it's like very unfriendly with a kv cache when all the intelligence needs to be append only and there's there's so many other cool things i love it um yeah the creative juices uh flowing i i guess you know another thing i'd be curious about is

[34:47](https://realalpha.app/v/ham8dROHCrc/?t=2087) I feel like you all obviously have released some, you know, stuff around images, right? And, you know, obviously multimodal, I imagine, is a focus. I think most of our listeners probably think of Jeb in the context of, like, you know, these decision models or type 1 things and kind of in classic, like, LLM replacement. Maybe talk a little bit about, like, the multimodal world, like, how you see the shape of problems there, how, like, Jeb models fit into those. So there's a couple of, like, branches to go off of here. I will say branch one is that being aware of the reality of AI is so unbelievably important.

[35:21](https://realalpha.app/v/ham8dROHCrc/?t=2121) And by its nature, just like long context tends to degrade intelligence no matter the model, multi-modality tends to degrade intelligence too. And this is like a well-known fact by like anyone in ML circles. And actually every modality has different rates at which it is degraded. And this is very measurable stuff, except when you put them into chatbots, because chatbots are optimized for human preference. It becomes a very, very fuzzy thing on whether or not it's better or not.

[35:46](https://realalpha.app/v/ham8dROHCrc/?t=2146) **Diogo Almeida:** It's very measurable that multimodal models will be less reliable than the text models.

[35:51](https://realalpha.app/v/ham8dROHCrc/?t=2151) And my philosophy has always been automate the easy work before the hard work. Like, will JEV models be used to automate software engineering? Maybe one day. But software engineering, maybe I'm arrogant, seems a lot harder than like really easy tasks that LMs fail at. And in order to like be calibrated to the world, we should study why they fail at those easy tasks and like really honestly, like go in depth into studying them, not to cheat them, which I think is like quite common, but like actually like dig in and figure out what is missing and why. And sometimes it's not the model. Sometimes it's the prompt.

[36:22](https://realalpha.app/v/ham8dROHCrc/?t=2182) Sometimes it's every other stuff, you know, like it's a collaboration between software and model. But it's through that that I think we like learn the truth. I have, like, stories about how we got to even be anti-benchmaxing more. But this is a thing that LLMs try to avoid because they will look bad on benchmarks, as far as I can tell. And so that is one aspect of multimodality. We have to take into account text tasks will be automated before the image tasks. Thing, too, is we invested really heavily on text tasks. And very much, like, we've done, like, a tiny amount of work in comparison to the image tasks.

[36:59](https://realalpha.app/v/ham8dROHCrc/?t=2219) Like I am like kind of jazzed that our model does as well as it does. I think that that's great. Like as far as I can tell, our lead in images is not as large as our lead in text, but we do have an overall lead in images is my vibe. Could be totally wrong, but that's kind of what I've sensed. And I would actually describe the image model as completely insufficient. Like I would be truly embarrassed to say that this is prod ready because there's so much jaggedness left in it that I'm jazzed if people find stuff to actually deploy. But I will say that this does not have like my seal of approval yet.

[37:36](https://realalpha.app/v/ham8dROHCrc/?t=2256) We need to do so many rounds because like it's not quite as bad as like other people just shipping a shape of something. But it's like in the grand scheme of it, like our image model is like maybe 10% of the way that our like text model is. As you move from like 10% by your characterization to like 90, 95%, what are the first kinds of things that you're excited about gem models doing on the image side on image land i have a soft spot for video games and computer use video games because i have like a video game background i and oh man i i confession i've also like played around with botting them not for competitive reasons but just because it's cool to automate stuff and i think like being able to play around with that and like i feel like software is a little bit like warping technology to your will and being able to like expand what you can warp to your will is just very cool in general so i think that is really cool in video games computer use i think is going to be super duper valuable unbelievably valuable i i think it's it's like especially combined with like lots of other stuff i think that ai should be so much better for interfaces it's actually kind of confusing to me how it is so bad you know like why there's a lot of people who don't like

[38:53](https://realalpha.app/v/ham8dROHCrc/?t=2333) using computers and they still have to even if they're like old and not experts despite that despite ai being probably better than an assistant they could hire to do a lot of computer stuff for them and i would love for that to be reliable enough that people could like trust it as an interface and you know i guess an additional thing that is like i have a soft spot for and this is why i'm wearing my like unsloppable shirt is i would love ways to just stop slop um for me personally it might be bad for the internet like the internet is funded by ads so like maybe it's a good thing i don't know but i would love if i could have some or someone would make me like a slop stopper where i just don't need to see it anymore that would be lovely but one thing i've been curious about from the outside is like the extent to which you know uh the advances that you've had at jev were enabled by like all the advances we've seen in in you know ai in the last two three years or could you have just done this five years ago like obviously you have a different approach that that is like in some ways orthogonal to the rest of the field like could you have done this five years ago or how much of what you built like does leverage some of the stuff that's been

[39:59](https://realalpha.app/v/ham8dROHCrc/?t=2399) done you know on the LM side? Five years ago, it would have been harder, but doable. Back when I was more naive, oh man, I sound like an old dude. When I was more naive and I thought that this would take a week of work, what convinced me, I love writing documents. I think it's a great way to think. I wrote a document at OpenAI that was titled, We Could Have Had AGI a Year Ago, at least by OpenAI's definition of AGI, which I don't think that people really listen to the letter of the definition, but I think it's actually a really great letter. I think that models from the starting from GPT 3.5 era, possibly even GPT 3 era, could have met the definition of AGI at the time, which is still the same, like doing most of the world's economically valuable work by volume. It doesn't say by volume, but I assume the spirit of that is by volume because the models are so unbelievably smart like gpt 3.5 level models surpassed median level intelligence at all sorts of things and like shouldn't that be used for tons of automation at least for the easy work and so yes i think that's something like this would have been harder but it could have been done five years ago absolutely and you know like on the people who are like obsessed with like for

[41:18](https://realalpha.app/v/ham8dROHCrc/?t=2478) example like the the benchmarks of the day and they see like what tasks ai is doing these tests are really hard. They're really, really, really hard. GPQA was saturated 2023-ish, maybe not totally saturated, but it was surpassing PhD student level at 2023, I believe. Google proof question answering for PhD level intelligence at these extremely hard multiple choice tasks, it was passing that level 2023. How is it that that level of intelligence was not being put to work that is super tragic so i think we've had a lot of like raw material for intelligence for a very long time and we're gonna see a whole lot of craziness yeah i mean speaking of that craziness i think part of i mentioned there's a lot of noise you know i thought there's been a lot of noise in the ecosystem and i bet many of our listeners will be familiar with you know after you guys came with these these models you had open ai and cloudflare microsoft and everyone kind of released these these models um i know you'll you'll say that the benchmarks aren't uh aren't a good way to test them.

[42:20](https://realalpha.app/v/ham8dROHCrc/?t=2540) They obviously have tried to show that they're near equivalent. I guess, how do you think about, like you obviously have now invited a bunch of competition into the space. How do you think about like what, you know, keeps Jev, you know, I guess everyone likes to ask about moats. So I guess I'll ask it on behalf of our listeners. Like, how do you think about that going forward for the business? Oh boy. It feels like other people are digging our moat for us, which is kind of not how I thought it would go. and I think that that's kind of bad for the field. I think that releasing models that are kind of more benchmark-optimized than real-world-optimized actually burns that goodwill for people to try things out, which is the only thing that matters.

[43:02](https://realalpha.app/v/ham8dROHCrc/?t=2582) I truly think that this is not great for the field. And my hope is that just giving unbelievable fucks about reliability is going to be the North Star. I think that will be immensely, immensely valuable for the world. there's just so much more to do there that I think that it's a little bit early to worry about competition. It's kind of like OpenAI worrying about a moat at GPT 3.5. Like, who cares? Like, GPT 3.5's market is nothing, or it was nothing, right, in comparison. It's still, I think, going to be nothing in the grand scheme of things. Like, we truly believe in Jevon's Paradox, obviously, hence the name Jev.

[43:44](https://realalpha.app/v/ham8dROHCrc/?t=2624) Like, Jevon's Paradox, like are we seeing rebound you know more demand because it's so cheap absolutely I believe every good sees rebound are we seeing backfire like the total market size increases this is this is when Jevons paradox holds Absolutely not. A key part of it has to be like, wow, this is so good and so cheap. I'm going to be throwing it all over the place because it makes value in all these places, and my budget for AI will actually increase. Speaking of investors, I guess you obviously recently raised a pretty big round. How did you even think about the amount of capital that made sense to raise for this business? I'm sure people were trying to throw money at you. How did you even think through that?

[44:26](https://realalpha.app/v/ham8dROHCrc/?t=2666) I will say that people probably don't believe it, but we can run our service profitably. And I hope that we never have to do the thing of temporarily subsidizing to get lock-in. I don't want to axe it out entirely because I'm a pragmatic person, but I really hope that doesn't have to be the case. I hope that... I find that to be a very shadowy, sneaky way to get people to use your service. So I and I hope to be the kind of person who's like, you know, always collaborating in like these prisoners dilemmas until, you know, until forced otherwise. So we actually did not need the money. And our goal instead was, hey, we want actually like the most valuable people on our side to help us build this business because we have no fucking clue how to do this.

[45:20](https://realalpha.app/v/ham8dROHCrc/?t=2720) So we really opted for like who are the best people in the eye who will help us. you can see who we got on our board well I always like to end with a quick fire round where I just stuff some overly broad questions into the end that I haven't had a chance to ask elsewhere I'm curious you know over the last month you know or last six weeks of having this product out there like what's one thing you've changed your mind on an AI in the last like six weeks having just seen how that's such a cool question whoa that's a quick fire question holy smokes I will say that we were really worried if our philosophy about caring about like the actual reliability of intelligence was a mistake like we had people in the team say like hey should we just release a bench max version and just hype it up and like do one of those early releases that like obviously is not great but just to like test out if the market exists i'm so glad we didn't do that and i i say seven weeks because like i i don't remember when launch was but like seven weeks i think was before launch we really did not know this we it was like a big leap of faith for us and like we were just really man like i don't think we could be like proud of ourselves if we did the same over promise

[46:33](https://realalpha.app/v/ham8dROHCrc/?t=2793) under deliver and like i think that that paid off super duper well i actually i don't know if it paid off maybe the product is sold itself and maybe it would have done even better if we like really optimized for for attention um but that was like awesome and i actually feel like now that my brand has been solidified as our brand I hopefully never have to do that like even close to that again and that debate no longer exists for us I hope I hope I don't know I truly truly hope because I have like a hard time like even giving like un-nuanced slightly incorrect answers because I think that that's just a disservice to anyone listening to me I would rather just like not answer the question than do the standard like CEO compression of the truth that errs on the side of generous so I hope that I will never really have to do that um but like something i've learned as well is that this reliability saga was sped run so fast for our class of models um i feel like i could be wrong but my temperature check of the environment now is people have already gotten like extreme fatigue of like the benchmarks and not actually useful stuff way faster than i would have thought like i i would have like thought that this would be a concern uh based on what i see because people keep falling

[47:54](https://realalpha.app/v/ham8dROHCrc/?t=2874) for benchmarks every time a new model is released right people are like well cool let's talk about it and then it's It disappoints, except when it's like an open-air anthropic model, apparently. And that just happened. I could be wrong, and maybe we'll see. I think it just happened a lot faster than I would have thought, because our model is so benchmarkable, such that people could blast out so many benchmark-crushing things so fast. I feel like this next one is gonna be very anticlimactic in comparison but I am as a podcast host I'm contractually obligated to ask you you know like we're we're an AI podcast the other big AI story right now is obviously anthropic um you know going public and folks kind of thinking through that business and I'm struck even in the course of this conversation like I actually don't know how you're going to answer this question because on the one hand I feel like you're very disappointed in like so many of the uh the lack of capabilities that LLMs have and maybe that they're like almost a hammer trying to like you know go on every nail of like different sorts of use cases. On the other hand, like you have been complimentary of like, there are things that coding

[48:50](https://realalpha.app/v/ham8dROHCrc/?t=2930) agents are really good for. And so I guess as you think about Anthropic is like, you know, one and a half, two trillion dollar company, like, is that something you're optimistic on is like, hey, that's that's undervalued or something you're like, oh, I don't know what's going to happen to that. I think coding agents are massively valuable. It's unclear to me how much of the actual use is truly valuable, like it being like net positive is different than it being worth all of the money put in that is probably a product question for them to figure out if they actually are truth seeking they probably have a better answer to this like did some people like just token max too hard or are they actually getting the relative value that spending that would make a difference on so i don't think i'm informed enough for that i have my guesses on what percentage gain coding agents are for various levels of projects and skills of engineer um and i wouldn't say that software is anywhere close to solved but that doesn't mean that they're not valuable i just think that there needs to be like a pragmatic take here i think that it's possible that the coding i could see the coding agent market going either way like i could imagine going very up and

[49:57](https://realalpha.app/v/ham8dROHCrc/?t=2997) very down very up in the sense that as people learn to actually get more value can people use it more i believe so i could imagine it going down because a lot of people have like overdone it because like there's so much fomo in the field and actually my sense of the state of ai is not that it's been like an upward exponential but a lot of like hills that keep going up down up down up down and then um eventually having like a little bit more uh at the tro trough trough of disillusionment i actually don't know how that word's pronounced the trough yeah yeah and then eventually it becomes like more stable usage and i think that you know like the the flatlining revenue might be indicative of that people are like realizing that maybe they spent too much i think so from a product perspective i think that there will be better experts than

[50:41](https://realalpha.app/v/ham8dROHCrc/?t=3041) **Diogo Almeida:** me at that. I actually get the sense that the greatest strength of the labs is not models or product or distribution, but financial engineering. This is the thing that everyone who's like very

[50:55](https://realalpha.app/v/ham8dROHCrc/?t=3055) veteran at the labs always talks about, where my understanding is they structure these deals in incredibly beneficial ways to not fuck themselves over, but fuck other people over who are taking the risk and trying to latch on to the rocket ship and are willing to accept whatever terms that would... On the compute side? Yes, on all sorts of sides, but especially compute. I think that financial engineering, as far as I've heard, is their greatest strength. And I have no idea how much of a rat's nest that is and how vulnerable they would be to any which direction of the market going. You know, like a bubble can pop just because expectations are not met, not because it's not truly a transformative technology.

[51:43](https://realalpha.app/v/ham8dROHCrc/?t=3103) So I'm a little I try not to I try to be calibrated in my predictions. I try not to be overly confident when I have no privy information. And I would I'm I'm a more conservative person. I get this. A lot of the expectations baked in are ones that are not just product level, but based on, like, misunderstandings of the future of AI and what it can do. So my guess is that because some people might be valuing, like, building God is basically infinite, there will be some amount of miscalibration there. And I think that, you know, like, the let them cure cancer is going to be like Marie Antoinette's, like, let them eat cake.

[52:28](https://realalpha.app/v/ham8dROHCrc/?t=3148) oh that is a that is a good lot yeah i mean obviously maybe i maybe it didn't spicy then that was that was a great one um i feel like the yeah i mean obviously so many people are underwriting you know the move into from coding agents to general knowledge work to bio to these other spaces and obviously like not the perfect use cases for rl vr by any means i still get the sense that the like the the nuanced way to understand agents perhaps and maybe all of ai right now is what decisions can you trust them with in terms of like stakes for the company and like can you like can you get like value without too much of the the downside and coding is like quite unique in this or like math is actually like like super duper unique where like there's no downside other than the compute you spend right like math models do not cost companies anything except if they're like paying people to validate their proofs which it's unclear if they even pay them for that um yeah to me maybe someone knows um but code also has like somewhat more kept downside because you have like a git right and you have humans in the loop and like there's ways to protect you from unwanted side effects which is the main downside of powerful

[53:39](https://realalpha.app/v/ham8dROHCrc/?t=3219) tools and powerful tools are necessary to do knowledge work because knowledge work often involves like side effects in the world i'm talking a lot like a computer scientist right now My sense is that for a lot of other fields, they don't have anywhere close to the same levels of like safeguards and ability to revert and like tracking and all of that. And that is why, to my knowledge, no one has automated like a drive-thru yet, even though people have tried really, really hard. And the stakes of a drive-thru are a lot lower than the stakes of many other forms of knowledge work. So I think that the truth will be like, there'll be like, nuance and skill into it and the people who will be the big winners are not the ones who just yolo all in but like understand reality and like deploy them um you know in the in in line with the realities of technology so i am i and i'm not given my state of ai i feel like most people are very fomo driven and they do not have a great understanding of the state of ai and they will probably err on the over-optimistic which will get them burned with many side effects or like the over-pessimistic which is business as usual which is how most businesses have been for this

[54:52](https://realalpha.app/v/ham8dROHCrc/?t=3292) ai boom when one thing you said there like you know you i'm curious you said you have an estimate for how much these coding agents actually improve uh engineering productivity what is that estimate okay i will say i have like error bars on this because i think that there's like long term effects, like, you know, skill degradation, and tech debt and stuff like

[55:14](https://realalpha.app/v/ham8dROHCrc/?t=3314) **Diogo Almeida:** this. So my estimates are in the 20 to 40% speed up range. But again, it really depends on the use case. And it also depends on like the skill of the user. And if you can like, you know, because

[55:25](https://realalpha.app/v/ham8dROHCrc/?t=3325) the side effects are the bad thing, if you can isolate the side effects in a way that those are very legible. So depending on the use case for some, it's kind of like an infinite percent speed up you know you would do things you would never you can do things you can never do otherwise for others they are like just so bad that they're as bad as like coding the thing by hand i would guess on average it's like a decent but small boost and i believe that is the consensus that i've seen in the field right now yeah another question i'm not totally sure how you how you answer like um how good you know there's obviously so much discourse around it AI safety these days, like how concerned are you about a lot of these, you know, this AI safety conversation? I think AI safety is not one monolith and there's like a ton of different slices of it. I think that the ASI discussion is actually extremely different from the bio risk discussion, which is very different from like the cyber risk discussion, which is different than like trust and safety. Maybe like ASI versus like bad actors, you know, using bio and cyber and I get the sense that most people concerned with ASI are a little bit divorced from the

[56:35](https://realalpha.app/v/ham8dROHCrc/?t=3395) reality of the technology. And this has been the case since day one. You know, like, I feel like you need to, like, actually build a reliable system with AI to have a voice in the room that something is going to be, like, so reliably, maybe not evil, but misaligned that it can destroy the world. But I think we're pretty far from stuff like that. additionally there's actually a bunch of nuance towards like you know asi rsi and like whether or not the models have a sense of agency on their own like paperclip maximizer is actually very different than bad actor programs super employee to destroy the world type things and like actually i think that those end up becoming quite different in threat model i think that we're breakthroughs away from true agency as much as people like to like read these chains of thought and like try to

[57:27](https://realalpha.app/v/ham8dROHCrc/?t=3447) **Diogo Almeida:** ascribe real agency to it um i would guess that we are something like two breakthroughs of like gigantic magnitude away from this i think even one breakthrough would not be sufficient is my guess

[57:38](https://realalpha.app/v/ham8dROHCrc/?t=3458) but maybe one gigantic one could be um and i think the threat model of like humans with super tools who want to do damage is very different than the ASI-RSI thing. I think a lot of people who also believe in RSI both have not studied the history of AI very much and also have no appreciation of Amdahl's law, where if you assume that intelligence is the spectrum and it's the only one that matters and you have to map things into that spectrum and you can also map humans and chimpanzees, it's a horrible argument in my opinion. uh then it does look bad but the reality is that it doesn't work that way and like the nuance of like the many different types of intelligence and how to do them reliably really matters a lot in this whole analogy could the field of ai development become a little bit faster uh yes but people have been trying this for a long time and it usually doesn't work out and it i don't think that people would count like a 10 speed up on ai research as rsi if it doesn't actually increase over time so i i i'm very open to being convinced by evidence but i i think that it's more of like a philosophical problem rather than a technological one um the kinds of people who like read uh like a bostrom super intelligence and treat it like reality over like a very

[59:05](https://realalpha.app/v/ham8dROHCrc/?t=3545) interesting sci-fi book makes makes me kind of worried yeah um well this has been a fascinating conversation. I thought we covered a ton of things. I want to make sure to leave the last word to you, like anything you want to point our listeners to. I mean, I'm sure they're all familiar with. And certainly by the end of this podcast, certainly hope they are. But like anything we haven't discussed or I don't know, the floor is yours. There's so much stuff to have happened in software plus AI. I think it'll be so fucking incredible. I think that we need to start untraining ourselves of this lack of agency that LLMs have kind of taught us about AI and start really building again.

[59:43](https://realalpha.app/v/ham8dROHCrc/?t=3583) Like, there's, you know, like there's so many possibilities that are going to be opened up and we should get back to playing rather than being like these victims of stuff. And I will do whatever we can to make, you know, playable AI. But it's... I really want. Oh, man, I have like so much to say here. I feel like I will be doing whatever I can for the field to like have this opportunity to just rethink software, rethink interfaces, rethink everything. But it's going to require like real builders to figure out how to do that. And I think that together we can make AI this really awesome force for good in creation and just all around awesomeness.

[1:00:26](https://realalpha.app/v/ham8dROHCrc/?t=3626) you know, like, and sometimes even like really weird, wacky stuff. I feel like that's where technology really shines. I'm Jacob Effron, and this has been Unsupervised Learning, a podcast where I get to talk to the smartest people in AI and ask them tons of questions about what's happening with models and what it means for businesses in the world. As I hope is clear, I have a ton of fun doing this. It's a nights and weekends project, in addition to my day job as an investor at Redpoint. But our ability to get these incredible guests on really comes from folks like you subscribing to the podcast, sharing it with friends.

[1:00:57](https://realalpha.app/v/ham8dROHCrc/?t=3657) It's really what ultimately makes this whole thing work. And so please consider doing that. And thank you so much for your support and listening. We'll see you next episode.