14 Aug 2026 · 2h 6m · Nathan Labenz, Flo Crivello · 27 quoted facts
Flo Crivello argues that the next phase of AI is multiplayer: an AI employee that lives in Slack, shares one team memory, and works alongside humans. He says context, not raw intelligence, is now the bottleneck. Lindy therefore builds an agent-maintained, file-based memory instead of relying mainly on retrieval search, and it works hard on caching and cost so such an employee is affordable.
The conversation moves from engineering detail (caching, nested context buckets, cheap Chinese open models) to a larger tension. Flo says humans are still needed to catch AI's strange mistakes, yet he expects that role to shrink, and he warns that alignment is unsolved. He also wants Chinese models banned even though Lindy runs on DeepSeek. The host pushes back and floats insurance and audits instead.
Flo says even John von Neumann would be less useful in an hour than an ordinary coworker, because of missing company context. Lindy therefore crawls Slack, wikis and meetings so the agent starts with the team's history.
A background memory agent 'naps' about every fifteen minutes and updates shared and personal files. Because the memory is plain text in a Git-backed file system, people can read it, edit it, and add privacy rules to it.
Flo says Lindy's cache rate is 85%, and a drop to 65% roughly doubles the price. Context buckets and a tree of nested buckets keep huge tool outputs and old conversations reachable without resending them.
Flo says DeepSeek Flash is about Sonnet 4.6 level and roughly a hundred times cheaper, so Lindy runs everything on DeepSeek by default. It is spikier and takes more turns, and the heavier Teammate product still pushes Lindy into negative gross margin.
Flo says models are brilliant yet still make absurd errors, such as walking to the car wash, so humans plug the holes in an 'Iron Man suit' arrangement. He expects that human-plus-AI phase to end, with humans eventually adding only noise.
It is an AI employee that lives in a company's Slack, connects to its tools, and builds up the whole team's shared context and memory. Flo Crivello says it is designed so everyone can talk to the same agent rather than each person using a separate chatbot.
Its memory agent writes public material to a team memory and private material to each person's own files. Users can also edit a plain-text memory prompt to say what the agent should never remember or should store in a restricted folder.
When a tool returns too much text, or a conversation gets too long, Lindy stores it in a sub-agent 'bucket' that the main agent can question later. Flo says nesting buckets in a tree lets the agent reach billions of tokens in about two model calls while saving money.
He says today's models are uneven. They can write 50,000 lines of code yet make silly mistakes like walking to the car wash, so humans must catch those errors. He expects this 'centaur' phase to be temporary.
He argues they are built partly by copying American models, which he calls unfair competition. He also cites censorship and propaganda risk, and the danger of the Chinese Communist Party influencing agents that act in the economy. He says he uses them only because competitors will, and the host suggests insurance or audits instead.
Stated by the speakers. Each quote matches the transcript, which shows what was said, not that it is true.
Lindy launched Lindy Teammate, an AI employee that works inside a company's Slack and accumulates the team's shared context.
“I think, like, what we are releasing today is called Lindy teammate. It is an AI employee that lives in your slack, connects to all of your tools, accumulates your entire team's context.”
Lindy's memory agent runs roughly every fifteen minutes in the background to update shared and personal memory, rather than once a day.
“it runs every something like fifteen minutes, like why why would you need to sleep every twenty four hours?”
Flo Crivello says Lindy's initial hydration of a twenty-person team's tools and Slack history consumes about three to five million tokens.
“For a team of twenty, that's gonna tend to consume 5 million tokens at most, you know, three to 5 million tokens”
Flo Crivello says the main bottleneck when Lindy crawls a company's Slack history is the Slack API rather than the language models.
“And actually, here, surprisingly, the bottleneck is not even the LLMs, it's the Slack API.”
Flo Crivello says Lindy's current prompt cache rate is 85%, which he considers lower than it should be.
“Our current cache rate is at 85%, which is is lower than it should be, frankly.”
Flo Crivello says a drop in Lindy's cache rate from 85% to 65% roughly doubles its token costs.
“now your cache rate dips from 85% to 65%. And the gap between both, it sounds small, but actually it's a lot of two times the price”
Flo Crivello says Lindy Teammate has put Lindy back into negative gross margin because users give it more complex tasks than the earlier assistant.
“With Lindy teammate, we're back at it and back into, frankly, negative gross margin territory.”
Flo Crivello says Lindy's error rate fell eightfold within the first week of its self-improvement loop going live.
“it went down by eight times the error rate.”
Lindy organizes nested context buckets in a tree where each node holds a hundred children, instead of the two used in a standard red-black tree.
“We went for a it's called a century tree. It's literally just each node's got a hundred nodes.”
Lindy backs its agent file system with Git to handle conflicts when multiple memory sub-agents edit files, and to keep a history of changes.
“The answer to that is Git, which by the way, also gives you history for free, which is which is awesome.”
Flo Crivello says Lindy works with Mesa, a vendor selling an agent-native file system backed by Git.
“We've been working with a Mesa, M E S A, they sell like uh an agent native file system that that's backed by Git”
Flo Crivello says Lindy uses E2B for its agent sandboxes.
“So we've we've gotten with the E two B for that.”
Flo Crivello says Lindy was started in 2022, which he now considers too early because agent tooling did not yet exist.
“We started this this company in 2022. And so it was it was in hindsight way too early.”
Flo Crivello says Lindy's internal pull requests per week, and lines per pull request, each tripled over three months.
“Our number of P Rs internally has, like P Rs per week has tripled over the last three months”
Flo Crivello says Lindy's headcount has been flat while he estimates its productivity has tripled over a couple of months.
“The team is actually, I think headcount has been flat now for a while, and we've again, like I said, we've actually tripled our productivity over a couple of months.”
Flo Crivello says Lindy's payroll still exceeds its internal AI inference spend, but he expects the two to cross within three to six months.
“So payroll is still greater, but the the lines are gonna cross three to six months from now.”
Lindy has its agents include the weekday with dates and uses a simple regular-expression validator to catch the one-day date errors Flo says Claude models made.
“Okay, if you do that, then you can have a deterministic validator that checks whether the date of the week matches the the day, the data was submitted. And that's just a regular expression.”
Flo Crivello says Lindy's default model is currently DeepSeek for everything, though the company is reconsidering because Teammate needs more compute.
“Everything is DeepSeek right now. We are in the middle of reconsidering it”
Flo Crivello says DeepSeek Flash is roughly Sonnet 4.6 level and about a hundred times cheaper.
“Sonnet 4.6 is a really good level, a really good model for like most use cases. And it's literally a hundred x cheaper, you know?”
Flo Crivello says re-optimizing Lindy's prompts with its GEPA-based loop costs about $10,000 each time a major new model is released.
“So it's like about like $10,000 every time like a new major model comes out.”
Flo Crivello says that with unlimited resources he would train a LoRA, a small fine-tuned adapter, for each user as a way of storing memory in model weights.
“I think eventually, with infinite resources, what I would like to do is I would like to do a LoRA per user, because LoRAs are actually pretty cheap to train.”
Flo Crivello cites a post by Dwarkesh saying Google's CEO Sundar Pichai earns about $150 million a year, as a hint of how much firms might pay for very good agents.
“Sundar at Google is paid $150 million a year, or something like that, you know?”
Flo Crivello describes Robin Hanson's book The Age of Em as imagining an economy of emulated human minds that can copy and split themselves.
“So for people who don't know what it's, what it is like, M stands for emulated mind.”
Flo Crivello argues that Chinese frontier model makers distill American models, which he says creates unfair competition.
“Or number one, they're obviously distilling. It's very clear. And so you're putting American open source model companies in an unfair competition”
Flo Crivello says that when he asks his own product about Tiananmen, the Chinese model refuses to answer, which he cites as evidence of censorship.
“When I talk to my own product and I'm asking it what happened in Tiananmen, it tells me, I'm sorry, I can't talk about that.”
The host proposes an insurance requirement, with higher premiums for companies running less-understood Chinese models, as an alternative to banning them.
“And then I also wonder about kind of a market mechanism, like maybe insurance should, instead of a ban, like, what about an insurance requirement?”
Flo Crivello says he favors creating a new agency, modeled on the FAA, to certify AI models, including sanitized or fine-tuned Chinese ones.
“I am in favor of like an F A A for for A I. I do think we need a new agency to regulate those models”
Flo explains the shift from single-player assistants to a shared, multiplayer AI colleague, and why shared context is the point.
“an AI employee that lives in your slack, connects to all of your tools”
Explains how Lindy builds a knowledge graph from Slack and meetings, and how plain-text memory prompts handle sensitive information.
“We call it a napping, not sleeping.”
Shows the cost engineering that makes the product viable, from cache rates to a tree of context buckets holding billions of tokens.
“we call them context buckets”
Flo explains why Lindy's memory sits on a Git-backed file system and why he now prefers buying infrastructure.
“an agent native file system that that's backed by Git”
Lindy uses its own agent for work such as fixing its CI pipeline, tripling output at flat headcount.
“P Rs per week has tripled over the last three months”
Flo describes why human oversight is still needed and why he thinks the human-plus-AI advantage will fade.
“right now, we are in the centaur phase”
Covers why Lindy runs on DeepSeek, how validators improve reliability, and why mixing models breaks caching.
“you should almost never use multiple models powering the same agent”
Flo calls fine-tuning a last resort but imagines a small fine-tuned adapter for each user as a way to store memory in weights.
“I would like to do a LoRA per user”
Flo lays out fairness, propaganda and protectionism arguments, and the host replies with insurance, audits and diplomacy concerns.
“I have something against Chinese frontier models, whether they're open or closed source”
Hello, and welcome back to the The Cognitive Revolution. Today I'm speaking with Flo Crivello, founder and CEO of Lindy, as he's launching Lindy teammate,
an AI employee that joins your company's Slack. This is, of course, competing directly with Claude Tag, and very notably, it is running on DeepSeek. Even so, Lindy is subsidizing onboarding at least, and the first half of this conversation goes really deep into where all of those tokens are going. We talk about the nuances of multiplayer mode, how Lindy thinks about the social contract surrounding historical data, and how they're using prompting to enforce it. Plus, a ton about memory implementation and the background processing that Lindy uses to continually optimize memory.
Flo's command of the relevant literature is evident throughout. We get Flo's vendor recommendations for file systems and sandboxes, and discuss why he prefers to buy where he can and remains bullish on software infrastructure startups. We also talk about life at Lindy now that Lindy teammate exists, and how he thinks about Lindy and all frontier AI systems as a super intelligence in most respects that still somehow chooses to walk to the car wash a bunch of times every day.
For now, Flo says, we're in the Centaur era. The best ideas at Lindy are generally co-created between humans and AIs. But Flo feels that this will be temporary, and before too long, he expects that humans will simply be adding noise to highly optimized AI systems. And that, it should be said, is in the good world, because, as Flo puts it, in light of OpenFace and other similar incidents, it is late July 2026. We have AGI, we're in takeoff, and we've not figured out alignment.
With that in mind, we compare notes on how friends and acquaintances at AI companies are now kind of panicking. And in the final section, we discuss Flo's very surprising position that Chinese models should be banned in the United States. As of now, I definitely do not agree, but we have a good natured back and forth about various arguments. And ultimately, he seemed open to a compromise on something perhaps more like an insurance requirement for companies selling AI services that would allow us to press the risk of Chinese models instead of banning them outright.
In any case, I have to say Flo is a one of one. After sprinting through the general purpose AI assistant product race for the last three years, he is still super open about how Lindy works, and he's fully candid about every issue he cares about. It makes for a really fun conversation about multiplayer AI, machine memory, and the uncomfortable politics of building your company on models that you yourself would like to see banned. This is Flo Crivello, founder and CEO of Lindy.
Flo Crivello, CEO at Lindy. Welcome back to the The Cognitive Revolution. Thanks, listen, or listen, the Delta be here. I am excited for this conversation. You got some big news at Lindy with a new product launch that is the occasion for this conversation. But obviously, we are in the thick of it when it comes to the AI exponential as well. And so you've been, I know, over time, outspoken on a bunch of really critical AI issues, and I definitely want to get into that as we get deeper into the conversation as well.
But let's start at the center with what you are working on. Lindy is now evolved again. We started with kind of a smart Zapier workflow software kind of thing a couple years ago. There've been multiple big evolutions. One that I also definitely want to get into is your recent post on shifting your model mix and getting away from American proprietary models a bit and embracing more of the open source and saving a bunch of money. But the big evolution now is that Lindy is a full on AI employee, and it's gonna be available in Slack, just like all your other teammates, whether they be human or agent.
Tell us about the new big launch. Yeah, you know, uh uh from like the get go, we've always been going after like the AI employee. And I I do think that it used to be quite early when we started to go after that, like three years ago, and now I think it's basically here. I think, like, what we are releasing today is called Lindy teammate. It is an AI employee that lives in your slack, connects to all of your tools, accumulates your entire team's context.
It's it's really like a team scaffold. I think like AI right now is in the middle of making this huge leap towards multiplayer experiences. I compare it to, you know, maybe you'll remember, like, we used to send each other like weird documents around by email, you know, with like revisions and stuff. And and you know, I compare the difference between multiplayer and single player AI as between, like, sending each other weird documents and Google Docs, like an actual shared document.
If you really want your AI to be a teammate, your agent to be an actual member of the team, like you want it to be where your team collaborates, which is slack. You want it to have its own shared context about the entire team, its own shared memory. You want everyone to be able to talk and collaborate with the same agent. Instead of right now, it's like we're all in the same meeting room and we're all talking.
And then every time one of us wants to talk to what's turning out to be maybe the most important constituency of the company, which is AI agents, we have to leave the room and then come back. You know, so that's what we're working on is like this new multiplayer experience and this multiplayer scaffold. So many angles of that I think are interesting. How, first of all, how does it onboard, right? And this is something that companies have put a lot into over time when it comes to their human employees.
Yes, I've done this myself with my own kind of deep context that I stumbled my way through earlier this year, and it is serving me really well. But it's easier for me cause it's just my stuff and I own it all and I don't really have to worry about what the expectations were around privacy, cause again, I'm gonna be the only consumer of it. When you get into multiplayer mode, now I'm like, oh gosh, you've got different channels where different people were gathered, and maybe some of those channels were private, maybe some of them were public, but only private.
So just practically procedurally, how do you suck up all the information? How is it stored? And then on a social level, what are the tricky considerations that you are identifying in the multiplayer context? Yeah, that is an excellent question, and that touches on exactly what we've been obsessing about, which is I really do think that as we're getting to AGI, and as we now arguably have AGI, intelligence actually matters less and less, comparatively speaking, and context matters more and more.
You know, I often think of it as like, look, you know, like one of the smartest men in history was John von Neumann, right? If you were to have John von Neumann just magically appear next to you at the office, this guy over the next hour or day would be less useful to you than your random coworker, right? So, and that's because of context, because, like, you got a job to do, and, like, you don't have time to onboard John von Neumann.
You know, he doesn't have the context to the date. So you're like, hey, like, I love you, I really want to talk to you, but right now I'm gonna talk to this guy because I have, like, an important thing to do. So we I do agree that context is super important, and it's one of those surprising times when actually agents are better than humans at onboarding. It shouldn't take people by surprise anymore, but it always does because you always operate under the assumption that, oh, you know, we don't have AGI yet, so obviously humans are going to be better, but they're not.
And the reason for that is because very often, one major ways that companies onboard their human employees is a wikis, right? There's like in-person onboarding sessions and all of that stuff, but there's also a lot of written documentation. And as everyone knows, the moment written documentation is written, it's out of date. And so what we've built is we built this shared context layer and what we call like a hydration system. And so basically the way it works is you you sign up to Lindy teammate, you connect your tools, so, you know, by all means, connect your wiki, like that that'll help, you know, like Confluence, Notion, Google Docs, whatever.
And then you connect most important, which is Slack, because that's where the real knowledge lives. It's just it's just a mess, but agents don't mind the mess. And so then what we do is that we crawl your entire Slack, and we build a knowledge graph based on a file system. Maybe we'll be able to parse and and edit after that. This is this is what it looks like. So you onboard on Lindy, and immediately she starts learning.
So you can see here, this is within ten seconds of us starting to sign up, and she's still learning. Like here, like the graph on the right is still limited. And so she tells me this is what I've learned about you. And it's surprisingly good. And then this graph on the right keeps growing and growing and growing. And importantly, there's there's two layers to this graph. And there is like a personal layer and there's a workspace layer.
So at any moment, you can go into your file system in Lindy. This is the new Lindy. You click on files, and it brings together this this memory file here that is supported by a bunch of reference files. And this is all maintained automatically in the background by, uh, a teammate agent. And so hopefully this answers your question about how do you do it? Because some some channels are public, some channels are private, some documents are public, some documents are private.
The way it works is that there is this memory agent. We call it a napping, not sleeping. So like it runs every something like fifteen minutes, like why why would you need to sleep every twenty four hours? So it just runs continuously in the background. And anything that is public, it updates the the team file system and the team memory with. And anything that is personal and that is private, it just updates each person's file system with.
So, and and in the end, it collates all of that, and the agent when you talk to Lindy teammate, uses all of that context and crawls the entire file system. So there's a lot of technical challenges we had to solve in order to get there because the context that it accumulates is like many millions of tokens and you can't add all of that at every turn so that was that was one thing we had to figure out yeah interesting I'll share a little bit about what I stumbled into and you tell me what you have learned that might be even better first of all just have to connect to all the tools and get the exports I found that there
were interesting edge cases that I was constantly running into in my own personal export heuristic that I tried to write One time I had a heuristic that the longer emails I sent are probably more substantive and I'll definitely make sure I capture those but then it turned out that at the very top of those of that power ranking was often something that I'd copied out of an LLM and was sending to somebody with usually a message at the top here's what I got from Claude or what have you but it was like weird now that according to that heuristic this Claude text is one of the more weighty pieces of my writing So I've bumped into a lot of these
things over time and had to special case. I guess maybe one question is just like when you step into a new organization blind and they and we logging is another right in in our Slack channels at my company Waymark we've had various you know logs dumped into Slack over and over and over again.
How do you deal? How do you identify these idiosyncratic special case things that could overwhelm or flood or mislead them get to the the real good stuff? So that's an excellent question. The I think the the main way that we've solved this is the fact that the memory is maintained by an agent itself. I think this is why I'm ultimately quite bearish on RAG as an approach, and I'm very bullish on on on this like agentic management approach because you have an actual agent which has its own memory.
So you have a sort of meta memory and and understands what it is that it's looking at. And by virtue of accumulating that knowledge little by little, the Bazi organization gets smarter and smarter about what actually matters. It's actually sort of similar to training a model because like you know if you were to insert poisoned data into the model data set like hey you know like I don't know Darth Vader was a woman you know like it it it would get the data but it would be drowned by all the correct data you know so here it's the same it's like if you have enough memories and if you feed that to a system that understands it the
system ends up understanding and so the very concrete example you are using is actually an emergent behavior we have seen the memory agent adopt because 00:00:00 Memory agent has its own memory, so it's a sort of meta memory about how do I manage my memory? What sources of information matter? Like, you know, all of that, and which ones are trustworthy and all of that. And we have noticed that the first time the memory agent crawls the slack, it finds some of those channels, like many organizations have them, like those like log channels.
And it learns to ignore them. It's like, I'm not gonna, I'm not gonna keep spending time on those channels. There's nothing for me to learn there. Saying that's been critical for us has been building meetings as as a first class citizen in this system. Like I really do believe that meetings are very underrated as a source of information. They are they are where like 90% of of the most up to date data leaves about the company.
Like everything that matters inside the company has a meeting around it. Every relationship, every project, every initiative, everything has a meeting around it. And so I think like those multiplayer agentic systems cannot really get meetings to like granola or whatever. Like I think you have to really incorporate it in your system as a first class. And so what we did is that you know, we we built that first class citizen. We have meetings now as a first class citizen in Lindy.
And and most importantly, it's like, again, it's not just a granola, you know? Oh, it's it's recording your meetings, and and and some questions about them, you can do all of that stuff, but it it goes beyond that, and it it feeds these meetings to your memory agent. And now, if a meeting was public, so you can set up you can set up meeting folders, which automatically add meetings, instead of meeting folders, which automatically add meetings to themselves and share them with your entire team.
And so then it like summarizes all of the meetings that are in these folders. And again, like if a meeting was added to a public meeting folder that's shared with the entire company, then it it it debates the team's memory with it. And so it keeps updating that context on an ongoing basis. And now you can, like, chat with. It's basically like a teammate that's been in every meeting in the company. You can chat with, like, the entire corpus of meetings.
You can be like, what are customers saying? Like, what's the biggest request lately? Like, you know, what's been the feedback about this and that feature? And and so 00:30:00 So this brings up another interesting challenge that I've again kind of stumbled my way to at least a solution for now on, and I'm very interested to get your take on both kind of the backward looking aspect of it and the forward looking perspective. The the issue is there's a lot of stuff in my broad general context that probably shouldn't be shared with other people, right?
Sometimes it's sensitive, confessional, one person's point of view on another person, like 00:40:00 Whatever, right? So I've actually created for my own wiki two versions of it. One is the I think of kind of Claude on my main laptop where I do my work as an extension of myself, like a second brain type of thing. So this thing is not taking on projects autonomously and running with them. It's just doing what I tell it to do, and it has the full wiki with all the all the gossip or whatever that's in there.
I don't have honestly that much super sensitive stuff. I don't want to make it sound like it's more dramatic than it is. But nevertheless, people didn't expect that when they were telling me something, which could have been on a call or an email or a private Slack message, that it was gonna go into some central repository of an agent that was gonna now talk to the world. So that one's just for me. I had it go through and create a version using the heuristic of what would be appropriate for a person to tell a human assistant.
The sort of still private, but a little bit more public facing wiki, where it would be appropriate for my human assistant to have your email and phone number, right? But it might not be appropriate for every detail of every conversation we've ever had to be in there. How are you thinking about as you absorb all this historical information being sensitive to what should and shouldn't be in memory such that it yon where it ought to be shared?
Second part is like how organization's gonna change now. Because I think this is all we're rewriting the social contract potentially in real time.
So I think application or one strategy for everything that came before, but the answer might be like new social norms going forward. I want to get your take on that too. Yeah, this has been a vociferous debate inside the team. There has been two camps. There have been the camps that are like, frankly, I'm in that camp. They're like just two tiers of memories enough, right? There's the public team tier, and then there's the private tier, and the private tier contains everything, and the public tier contains like stuff that's only private public.
And some other members of the team have been saying what you've been saying. Like, they've been saying, like, no, actually, even my private tier, I don't want to contain a lot of stuff. I want like a super power private tier. And like, and then they're trying, there was like this whole thing, like, what if we defined multiple memory bubbles and the user can edit them? And I'm like, that sounds kind of overkill. And so what we landed on and what these teammates landed on, honestly, is like they have edited their own meta memory prompt.
And the meta memory prompt is just a text file. It's your memory dot md file at in your in your file system in Lindy. Your memory agent has that memory prompt injected in its context window at every moment. And so if you insert a line up there, that's like, this is what I never want you to remember, just at the top or something, anywhere in the file, really, but you can do that. This is what I never want you to remember.
Oh, by the way, that other stuff that's like a sensitive topic, please remember it in that other file in that other folder that's get out of view, and I don't want you to pull this file or this folder unless X Y or Z, right? So you can just like leave these these these instructions. And so that's another reason why I'm bearish rag and I'm bullish text and file system, is you can inspect the memory and you can edit it, and you can just like very granularly insert this kind of guardrails.
Yeah, interesting. So you've just to make sure I understand. Can repeat it back. One source of ground truth, and you create different lenses on that information just by prompting. So you can have your I could tell my agent, like, hey, just so you know, this contains everything I've ever, every conversation you've ever had. Always use the heuristic of you should only really be using information that would have been appropriate for me to share with a human assistant.
And if it doesn't seem appropriate, then don't use it. And obviously, instruction following is getting extremely good. So I can pretty well. I actually I was I was misspeaking. Like, there is a time when I have used this, and it's literally for this podcast. I was preparing to go on this podcast, and I knew I was going to talk about the memory agent. So you see here I have like a memory MD, which is my actual memory file, and then I have a memory two MD, which is like a sanitized memory file where I've removed overly sensitive information.
And you can see in my memory MD here, there is like not a binny. The file system also contains a memory two MD file. Ignore its contents. They are just here for the users demoing Lindy on podcasts. Oh, yeah, you can just you can just do your own thing here. Yeah, okay, cool. I do because my version does have the it's not dry problem right now. I've got two things to maintain, and it does create some overhead.
So I can see why that could be advantageous. Hey, we'll continue our interview in a moment after a word from our sponsors. Today's episode is brought to you by Anthropic, makers of Claude and Claude Code. Over the last few months, Claude has helped me build and refine a personal deep context database that now contains all of my emails, Slack messages, tweets, DMs across platforms, video calls, and podcast transcripts, going back a full five years. On top of that, we've now layered summary articles describing my relationship with hundreds of contacts, organizations, and ideas.
And now that this exists, there's almost nothing that Claude can't help with. For my angel investing, Claude can now draft investment memos in exactly the form that my venture fund requires, based on the calls I've had and the emails I've exchanged with the founders. And when someone needs a favor, Claude can often do it as well as I can. Recently, a friend reached out to ask if I know anyone who might be a fit for a role that he is currently hiring for.
Initially, nobody came to mind, but then I thought to ask Claude, and sure enough, it identified two great leads. Claude is the AI for minds that don't stop at good enough. It's the collaborator that actually understands your entire workflow and thinks with you. So for problems worth solving, get started with Claude at Claude dot ai slash TCR. That's Claude dot ai slash TCR. And check out Claude Pro, which includes all of the features mentioned in today's episode.
That's Claude dot ai slash TCR. What are the things are kind of coming up as you're doing multiplayer? I remember going back to GPT four way back when I was red teaming. I did some it's taken longer than I thought. The reason I even bring this up is because back then I was doing some simulations of a facilitator. You're an AI facilitator in a group family group that's focused on exercise. And your job is to encourage people and give some reminders and whatever.
And I was playing all the roles of the participants and just having the AI play that role. And even at GPT four, it was like doing pretty well in a relatively simple context. And yet it's been like three years now until we're finally getting these real kind of AI employee type experiences. What has been hard about getting multiplayer to work in an intuitive way that might be unobvious to somebody who hasn't been through the slog himself?
Yeah, I actually think it's been the scaffold, the context management, like this piece, like the context build up. It's it's in a way inspired by Capacy's Auto Wiki idea. And so we've had to do a lot of work around context management. Because once you talk to an AI employee, it's actually quite unlike just talking to a to a to like a chat GPT or a cloud. Because you actually expect it to keep a very rich representation of its past context and of its past memories.
You really expect a level of consistency and coherence out of an AI employee that you don't expect out of your cloud. Like cloud sort of like it's cute. Sometimes it plugs like previous memories about you in your chat, but like you don't really, you don't really do like heavy work on an ongoing basis with it, that like in the same way that you do with an employee. So managing all of that context, so both the memory agent that builds up the context, that's millions and millions and millions of tokens per user, and then the core agent that actually uses that context.
And how does it retrieve the right information at runtime? Like this has been like a major, major, major challenge. Frankly, I've sometimes been telling the team, like, and I can't imagine we're the only company thinking that. Like, I sometimes feel like we should be publishing. Because I think we're doing stuff that's like seriously state of the art, and it's quite frequently that we do stuff. And like, three to six months later, we see a paper come out and blow up about about that thing.
There was one time when literally the paper was named what we had called the thing internally, because it was obvious. And so I think like context and and memory management has been a really, really big part of the challenge. Reliability is always another part of the challenge, right? Like you want to make your your model work, and so we've we've worked quite a bit on on reliability. Like we've called it like a validator is basically a sort of it's an LLM as a judge that that triggers like multiple times during the tasks, but it's modular.
So you have multiple LLMs as a judge fan out, and it's some sort of council, and then they talk to each other and they decide what to do. And so that's been one the self improvement loop. I think like since GPT four models have become so capable that now you can actually have self improvement loops. And so, you know, well, not the first ones to talk about it. Well, no exception. You know, we we Lindy is now self improving.
So we can literally see a curve of error rate go down into the right. You know, that's the direction you want to see error rate go down. It went down by like the literally within the first week of us putting the self improvement loop online, which was like two months ago or something, it went down by eight times the error rate. I think I think this probably captures it. I think those have been those have been like the really big meaty chunks we've had to figure out.
One big question that brings to mind is how you are managing caching, and then caching, of course, is just one, you know, angle on managing cost in general. So maybe we could expand beyond caching and talk about cost management. Obviously, there's a significant upfront amount of tokens that you're going to dedicate to getting all this context from a new organization and processing it. And so I'm interested in the business model implications of that. Do you have to charge a setup fee or do you need an annual contract?
How are you balancing your initial investment with the level of commitment from the company? And then with so many tokens getting processed all the time and with memory being assembled in different ways in different situations all the time, different users with their combination of the public and the private. How are you thinking about managing input token volume? How much is caching playing into the strategy? How are you making this something that you know still on net ends up costing less than a human employee?
Headline we haven't. You know, I think headline like these things still cost more than the human employee, but not for very long. And like, look, I'll be real, like we're subsidizing it. You know, this is what we've raised this money for. You know, actually, fun fact, we used to subsidize it very heavily. And then we we've sort of famously switched to a Chinese model and that that made us stop subsidizing it. And now with Lindy teammate, we have realized our users again are throwing like much more complex stuff that they were asking to our previous product, which was more like a personal assistant.
So it was it was like simple each task was like, hey, like, send this email to Flo, schedule this meeting. Like, I always say you don't need God to schedule your meetings. So like here, like even a deep seek flash was like enough, frankly. And so that was cost effective. With Lindy teammate, we're back at it and back into, frankly, negative gross margin territory. And so that said, you know, obviously, I don't like having negative gross margins.
I'm at peace with it because, you know, we we are all very, very confident that it's very temporary. And I actually think you you sort of want to to to build for like the next generation of models always. And so, you know, we're trying to mitigate the impact of that. And so, yes, context management is is a is a huge piece of the puzzle. Caching is a huge piece of the puzzle. Our current cache rate is at 85%, which is is lower than it should be, frankly.
And we we spend a lot of time just iterating on it. We've we've we've put a lot of systems in place to alert us when the cache rate dips because it's finicky. You make any change anywhere in your system, and you break your cache, and now your cache rate dips from 85% to 65%. And the gap between both, it sounds small, but actually it's a lot of two times the price to go from eighty five to 65%.
So, you know, set up all of those systems. I think one of the biggest breakthroughs we had, and this is one of those things that I'm frankly expecting someone will publish. I guess RLMs was like adjacent to that, but it's it's it's we call them context buckets. And the the the way it started was if an agent calls an action that returns too much context, like some some actions, some MCPs in particular, retrieve like 100,000 tokens or something, okay?
You don't want to send that to the to the agent because it's going to get confused. Okay, so what you do is that you expose that as as a as a summary of the context bucket. And it's a sub-agent which contains the entire bucket. It's like, hey, this action returned too much. So I'm here, you know, to stand in for what the action returned. Roughly speaking, this is what it contains, okay? And then the agent can enter into a conversation with the sub-agent, which itself has its own caching, and can manipulate the context using unique utilities.
So right here, you're saving a lot, you're saving a lot of money, and it goes quite fast. Then we went one step further and we were like, hey, what if we could have recursive context buckets? What if context buckets could contain other context buckets? And what if compaction? Because obviously, we have compaction was bewildered by such recursive context buckets. And and by that, I mean, like now, when we compact, so the conversation goes, it passes this threshold.
I think right now it's like 200,000 tokens, but like, we keep we keep tweaking it. And at some point, we're like, okay, we're gonna compact. We compact, and then the compaction sends sends all of that stuff into a context bucket. So now the agent can query that context bucket. So it's not like every compaction is always lossy by nature. And I think compaction operates under the faulty assumption that you never need access to ground truth, which is is false.
You you at some point do need access to ground truth. So with that technique, you have access to ground truth. And then we went a step further, which is so you you have that context bucket, which is compa the compaction of the of the previous conversation. The conversation keeps going, keeps going, keeps going. You need to summarize again. What you do is that you take all of that, including this context bucket right here, and you compact it into a new context bucket.
So now you have a context bucket containing a context bucket. And in the end, so you do that, and so you have this emerging property where it's like you can have the agent access any point of any infinite number of tokens at arbitrary levels of granularity. Okay. The problem is that if you if you have a computer science background, like this gives you like what's called like O N complexity, right? So it's like if it wants to access like nine context buckets ago, it's got to go through nine layers of sub agents, and that's really slow and really expensive.
And so what what we ended up doing is maybe it's getting too technical. But we 00:59:00 Never. Okay, well, have you heard of AVL trees? No, but I'm all ears. Have you have you heard of red black trees? Um, there is this 01:35:00 I went to the I went to the computer science school of hard knocks, so it wasn't uh formal training for me. I was actually glad to have AVL and black trees because I was like, oh my God, you guys remember your training?
You know, this is this is the moment we use those pecking things. Uh, because everybody in school is always like, when do you really use those pecking things? Like, we're using AVL and black trees, baby. Uh, right, right black trees, they're called. Uh, it's so what I just described. If you just do the naive implementation, you get that really nice emerging property for free, which is you have all of those linked context buckets that contain one another.
But every context bucket always contains just one other context bucket. And so you end up with like this Russian doll of sorts of context buckets. And like, if you if you want to open, if you want to get to the bottom, it takes a very long time if you want to, you know. So what you do instead is that you have context buckets contain multiple context buckets. Okay? And And so you you end up with a tree.
Because basically, what you want is you want the top most context bucket not to contain the last one hundred. You want it to contain the the first context bucket and the last context bucket, right? And so you end up with it's a self balancing tree. There's these two algorithms. There's the AVL tree, and then there's the red black tree, which are algorithms that are used to balance a tree. So instead of having one long line, you want to minimize the height of the tree such that, like, going to the bottom takes as few jumps as possible.
Okay. Technically, AVL is the best is is the is is the demonstrably best way to balance a tree because it leads to like the lowest height. A red black tree is better because it takes into account the cost to balance the tree, which is it's expensive because you need to regenerate a lot of your context buckets. And you you you miss your cache when you do that. It's it's very expensive. So red black trees. The canonical implementation of red black trees is on binary trees, which is just a tree where each node's got two children.
We went for a it's called a century tree. It's literally just each node's got a hundred nodes. So the node below that's gonna have like 10,000 nodes, right? And so literally with two jumps, you can have 10,000 context buckets, and you can access like all the context in the universe. And that leads to some really surprising behaviors because you're literally no more than two LLM calls away from being able to access 10,000 context buckets, each one of which contains 200,000 tokens.
Right? So you had like 2 billion tokens here of context in in two LLM calls. And and that is what leads to like those really surprising behaviors from those AI agents where you ask them any question and they remember everything perfectly all the time. And that's awesome. That was that was one really big thing we had to figure out. I hope. Please, you know, like copy us, like I will just we just don't have the time to publish.
But you know, this is this is we don't mean for this to be secret. Like, I think this is a really powerful technique, and I've have been surprised to not see more communication about it. Just for calibration, how many tokens are you finding businesses have? Like, when you say with these two levels, you can get to 2 billion tokens. I don't have a great intuition for does that cover a twenty person team? Does however it's been in business for five years?
What's kind of the heuristic for team size and length of history that translates into how many tokens? You can probably do the math, but like 2 billion tokens is we should ask Claude, but it's it's probably bigger than most libraries. So I think we we have few customers who whose knowledge base, whose memory is bigger than that. Most of the time when you get started, you consume like for for like a team of twenty, you know, the the hydration component.
So it's like, again, it's this moment where, like, you connect your notion, you connect your Slack, and then we crawl everything, we look at all of that stuff. For a team of twenty, that's gonna tend to consume 5 million tokens at most, you know, three to 5 million tokens, and that's with like a team of twenty and like years of Slack history. So, like, we literally take all of your Slack history. It takes a while. And actually, here, surprisingly, the bottleneck is not even the LLMs, it's the Slack API.
It's like, is that happy about you? Like crawling 02:00:00 02:00:00 Yeah, I've experienced that. Yes, I'm sure. The right way to do it if you're doing it personally is to export your workspace archive. We can't ask users to do that, so we just crawl crawl the entire history. So no, I mean, like, 2 billion tokens is a lot of 02:00:00 Yeah, one data point I have is a podcast, and I'm quite self indulgent, and we sometimes go on for a long time.
But still, it's usually only in the sort of thirty to 50,000 tokens range. So then, even having done hundreds over the course of a few years, we're still only in like single digit, maybe low double digit millions for like the entire full text history of the hundreds of podcast episodes. We're still at least we got orders of magnitude to go before we would get into the billions. That said, now think of a company that's got, you know, three to seven hours of meetings per day per employee.
Times 2,000 employees. That that is that meetings are when you really start to rack up tokens really, really quickly. How do you? So I did experience the Slack thing. Actually, the funny anecdote is the only time I've really experienced data loss. It wasn't too uh painful, but Claude had, you know, identified some weakness or whatever in the and I I'll bring this back around in a second too. But my kind of raw data is living in a SQL like database.
And Claude recognized some deficiency or whatever that it wanted to correct. And it just was like, all right, I'll just drop that database and we'll redo it. But what it wasn't taking into account was the fact that we'd been rate limited on Slack for days to get all the stuff out of Slack. And so it was like, no, you dropped that. And that was the only place where we had Slack. It wasn't really lost, but it was like, okay, we're now gonna be like at it again for five more days just to make API calls to get out of Slack.
In my case, I am the admin of at least the Slack accounts that matter to me the most, but they're like workspaces that I'm in where I'm not the admin, and so I can't grant this kind of access. From a business standpoint, does that mean that you have to have admin buy-in from the beginning? Is there that seems like it makes it hard to do product led growth, land and expand type stuff? That's that's a good point, you know?
Look, I think that's one reason why PLG eventually you tap out of it as you go up market. Like, like, yes, you have to be an admin, you have to have permissions to install applications on your Slack workspace. So it tends to be fine for teams of up to like fifty to one hundred. Like after one hundred is pretty rare to to to for everyone to have this kind of access, but many teams up to one hundred, like pretty much anyone can just Slack like install like an app.
And yeah, I mean, if you if you go up bucket and you're like part of a bigger organization, then we have some users who are like just like championing us internally. And they're basically begging the IT to let us in, and then they're like making making these meetings happen. And and you know, IT is rightfully cautious. But look, you know, we're like SOC two compliant, we're HIPAA compliant, we're GDPR compliant. We you know, we've got our ducks in a row, and what we're quite used to navigating the internal review processes, which can be burdensome.
And and and so, you know, is not in the business of being a pain in the butt of anyone. And they are actually themselves very eager. Very often, IT was given an addict, like, hey guys, we need to, we need to actually become an AI native company. So they're actually seeking this kind of solution. But yeah, I mean, you need admin access for sure. Okay, so let's go back to the data structure for a second. Let me give you my data structure, and you can tell me what you think I might be leaving on the table with my home spun version compared to your professional version.
And then it's one of my favorite habits these days. I'll take the transcript and give it to Claude, and we'll see if we can close the gap a little bit. So my version is simply take all of the content that you know, all the sort of digital exhaust that I've created over the years, which is email and DMs and Slack and Google Docs and whatever, right? Put Slack being the most painful one, make all the API calls to export all that stuff into a SQL database.
I have everything there mapped onto a threads and messages structure. And sometimes that took a little bit of squinting, but it basically seems to mostly work. And that's like the raw ground truth. And I do agree. For me, it's I've found that it is quite important for the model to be able to get to raw ground truth. Then I've just done, and this only goes back five years, but that's usually plenty. It's not too often that something happened more than five years ago is like operationally relevant today.
I just do a month by month export. I found that for me personally, it was like a couple 100,000 tokens per month that I'll then compress into a monthly summary, which is like ten ten percent at the length, right? So whatever two to 300,000 tokens of raw monthly content goes to twenty to thirty. Then I'll do that at the yearly level as well. So again, you've got another two to 300 that goes down to a twenty to thirty.
And then on top of all that stuff, then I just have the model go make a wiki. And one of the key things that I found along the way in terms of how to get to ground truth. Of course, the model can always just send a SQL query, but what should it be querying is is not always obvious to it.
So what I tried to do in my summaries is have the summarizer capture distinctive short sequences of words that would be like needle in a haystack. Um, so it would be like, oh, here's what happened, blah, blah, blah. And then a quick little footnote at the end of what the source was, a DM from whoever. And these four words will take you exact to that, and probably only that in like all of my personal history. I'd say it works pretty well for me.
I'm usually pretty happy with it, and I do to the right ground truth more often than not. It is only one person. Multiplayer is going to be as I start to onboard my wife, and then that'll complicate things at least a bit. What do you think I might be leaving on the table? Especially as you think about multiplayer that demands an even more sophisticated approach. Yeah, so I believe the approach you're talking about. And we we still have some rag in our database, and we use that for our rag as well.
It's called like hypothesis driven retrieval. I don't know if you've if you've heard that. It's just all the reverse hypothesis driven retrieval. So those both, those both. So what hypothesis driven retrieval does is it generates like suppose it's asking like, well, was Justin Bieber born, right? And like, and then you you basically, instead of searching for that question using rag, you search for your hypothesis. Like Justin Bieber was born in Paris, Justin Bieber was born in New York, in Berlin.
And you search for this this this bucket of hypotheses in your RAG database because obviously semantically each of these hypotheses is a lot closer to the answer you might be looking for than to the question where was Justin Bieber born? So that's the first thing. Then what you do is you also do the opposite, which is when you find an answer in your in your RAG base, you generate a bunch of questions that map to that answer, and you attach them to the answer.
So now you can the retriever can both look for hypotheses and can still look for its answer for the question. And that that's gonna tend to increase retrieval quality. That that's totally fine. I think one thing you may be leaving on the table is it is really healthy for all of the memory to be managed by one agent. Even though that agent and both retrieval and updating of the memory. And so even though the memory is in the file system, and so your agent could actually just go mess with the file system.
What we have found actually and can still do that but we we tell the agent we prompt the agent to be like if you're looking for memory that you don't have please ask the memory agent and the reason why is because when we ask the memory agents then what we do is that we log the query we log the answer and we log how many hops it took to retrieve that and then when the memory agent is napping and dreaming it looks at this log of like hey this is this is the kind of stuff people have been it's almost like a librarian you know it's like ah like I've been I've been heated up left and right This is
this is a question that I get quite often. So right here, by the way, it acts as a sort of cache, you know, because automatically, so you know, you you you always keep like the memory of the memory agent, like the last like thousand like query and like question answer pairs, you know. So it's like a rolling window, and so it is a sort of cache. But also the memory agent is not going to just rely on that cache because every cache at some point is going to TTL it, it grows stale and all of that stuff.
So what the memory agent will also do is like it will it will restructure its own memory to reduce the number of haps that is needed to retrieve like the most frequently asked questions and that that that tends to work quite well to improve the memory retrieval and the memory speed cool interesting another system we we experienced we experimented with and we ended not implementing it even though it does have strengths from a technical standpoint is have you heard of caveman it's uh it's it's it's really simple It's basically you rewrite your text as a caveman.
It's like me, you know, you know, listen, not like Burrito, you know, like. So retarded. But if you do that, you cut your tokens by like twenty or 30%, which is good, you know. And so cutting your tokens like it'll make your system faster, it'll make your system cheaper, it'll make retrieval more accurate, it'll just be better, with basically no loss of informational context. You know, if you the reason we've not done that is because we actually do care about the maintainability, like the auditability of the system.
And so we did that. Every metric went up. It was awesome. And then the files in the file system look really dumb. And so it doesn't make us look good, you know, to enterprise customers. Like, why do my memory files look like that? But it it works. Actually, similar. You know, there's like a bunch of those techniques to just compress your context. Have you heard of a tune? T O O N. It's it's it's a sort of like JSON alternative that's made for agents.
That's like, I think it's like twenty or 30% more token efficient than JSON. JSON is actually not that token efficient. So tune tune works better. It's twenty or thirty. So this work. This is less important for like memory agent, but it's just I recommend everyone like makes their agent use tune instead of JSON and have like a a tune passing middleware between your agent and its actions. Like no action should expose JSON to to the agent, like every every agent, every action should expose.
Yeah, interesting. I'll have to check out this tune format. I've just used YAML, which also is motivated by similar uh advantages, but yeah, two token oriented object notation and token. I see where we're going with it. It's YAML ish, and it actually increases the performance of the agents. Like agents, apparently, it was just surprising because there's so much JSON in the training set of the models, but they're apparently more comfortable with tunes than with JSON. JSON does have a lot of noise associated with it, for sure.
I'm more comfortable with YAML than I with JSON. That's that's something. The AIs, they're just like us. That's uh, listen, one of my biggest surprises as a brief aside. I'm actually working on, I don't know if this will be a podcast or a thread or whatever, but just taking stock. And, you know, of course, agents can help make this a quick project where it would be a very tedious project otherwise. Just going back in history and identifying things that I've been wrong about prediction wise and some things I've been right about prediction wise.
I think one of the things that I used to say quite confidently that has aged maybe the poorest is that we shouldn't be anthropomorphizing models or that will lead us astray. And I've just been shocked over and over again by how productive it can actually be to anthropomorphize the models. It still feels a little dangerous, kind of, but it's hard to argue with the results. I 100% agree. I do think, though, that's that's like a happy medium of anthropomorphism.
And I think, like, that's that's the other thing that's very top of mind for me right now is to what extent does anthropomorphization stop being true, in particular open deprompting to AI employees. So, for example, where I've changed my mind, and I haven't changed my mind completely, but I've changed my mind a little bit, relatively speaking, is about multi agent systems. I actually have come to believe that most of the time as much as possible you want to consolidate that as much as possible into a single agent not multi agents that's not always possible and there's there's still good reasons to use multi agents like this is not like an absolute statement but most of the time that's the case
and I actually think that humans like intuitively find themselves biased towards very very multi agent system like far more multi agents than is optimal because they're anthropomorphizing a little bit too much and they're comparing a little bit too much to human organizations It's like, ah, I have a data scientist, I have an engineer, I have a designer, I have a project manager. And so I'm going to create one agent for each of these things. And actually, the reason why human organizations do that is because every human has only twenty four hours a day and can only contain that much context.
Whereas agents obviously have none of those constraints. They can fork, they can duplicate, they can do as much as they want all day. And so I find that division of labor is not a good reason to have multiple agents, which is which is a major major fact. Now you again there are still multiple there are still different reasons but the division of labor is not one of them so I agree I think by and large I think people are right to anthropomorphize models because they were after all trained on human tokens and with you know human RL and all of that stuff but I also think that AI organizations shouldn't be overly anthropomorphized yeah that's interesting is it as you
talk about a single agent obviously they can run multiple copies of we can run multiple copies of it of single agents in parallel Have you reached the point yet where that starts to create database style problems where if multiple agents are updating the same segment of history at the same time we have all these sort of you know transaction level guarantees in traditional databases for that reason does that start to be an issue as you have ten Lindy teammates running over the same kind of file system corpus yes less than you would expect you know like I think the way human organizations have solved that is via Git Git is the by far the best way like large groups of
humans have found to like work together on the same thing, you know. Because you can you can just merge conflicts, you can rebase, you can do a bunch of that stuff. So that's that's what we found. And the file system in Lindy is actually, for this reason, backed by by a Git repository, which gives you a ton of amazing properties. So including like merge management, because the the memory agent sometimes will fork itself. It just basically spins up a bunch of sub agents.
You're like, all right, let's see what happens since the last time I napped. And in some organizations, some large organizations, there's so much context that was created since the last nap that it needs to create a bunch of sub agents. And during the the initial hydration, like we really wanted to go fast because we wanted to impress the user. Like time to wow is so important in PLG. And so that graph I showed you, like, literally, we we we've optimized it so hard.
So it happens within ten seconds of sign up. So it's like, sign up, Notion, Slack, boom, right? And so we've it's been a lot of work. And so one way you do that is via is via like a collection of agents. And so, yes, these collections of agents tend to step on each other's toes. The answer to that is Git, which by the way, also gives you history for free, which is which is awesome. By the way, like, and this is not a sponsored message.
Like, they've just been doing a really good job. We've been working with a Mesa, M E S A, they sell like uh an agent native file system that that's backed by Git, so you create all of those repositories. Initially, we started by the way, it's it's one of those intuitions you have to build as you build these systems. Like, it's very easy, it's very tempting to be like, I'll just hand roll it, I'll just create my own Git, I'll just like manage my own file system infrastructure.
And you you you learn that there is so much more depth behind those systems than you first appreciate. And so I now I am. So it's actually funny because it it sort of goes counter to the the prevailing narrative these days, which is like SaaS is dying, you can vibe code everything, and we'll all be coding a lot of stuff, but the infrastructure will not be coding. And I've become so infra build now. I'm like, anytime I have an opportunity to buy instead of build, I will do it because, like, you save so much time, and there is so much more depth and thinking and design decisions that goes into these products than you may appreciate.
I love a good vendor shout out. Any other vendor shout outs, any other key primitives that you're buying that you would recommend? Well, you need a you need a sandbox. So we've we've gotten with the E two B for that. Again, like many people, when once they have a sandbox, they're like, sweet, I have a file system. We at least at scale, we strongly recommend decoupling them for for the reasons I just mentioned and and more browser-based, obviously, is is excellent for for browser management.
Who else do we like and use? You know, I've we we went down a deep private hole. Like one major exception to what I said about being infra build has been observation and and evaluations. We we have a complicated history because we we got started. We started this this company in 2022. And so it was it was in hindsight way too early. Like agents were not ready. No one was talking about agents back then.
People were talking about like generative AI, if you remember. I was like, that was pre like baby AGI and all those kind of early experiments. You know, so agents didn't really work, you know. And and so as a result of that, there was no tooling. And so as a result, unfortunately, we had to build a lot of our own tooling, including our own eval platform, including our own observability platform. And I don't recommend doing it at home, honestly.
It's like just to use. And so we've gone through the whole evaluation. We've looked into Langfuse to LangSmith into Brain Trust into AGDust, in which I'm a proud investor. We've looked into a bunch of those solutions. And every time we were like, I think we just had such a head start. And at this point, our homegrown solution is so built exactly. We've gone through the pain of building. You know, so if a solution had been available back then, I would have picked it, but it wasn't.
So we built it. We fleshed it out over the over the years. Now we have like one engineer who's full time dedicated to that solution, which is a lot in in the vibe coding era. This guy is like cranking and he's it's become like a very, very, very extensive solution. And truth to be told, we are at feature parity with all of these solutions out there and more. We have a lot of internal features that I don't see these guys have.
And it's so built for our own internal scaffold and our own internal use that we ended up just hand rolling that internally. Yeah. So what does life look at look like at Lindy internally these days? I think it it strikes me that you're now. I'm sure there's been like a dogfooding process. I know there's been a dogfooding process throughout the history of the product. But now you're at this point where it's like there's no limits to the dogfooding, right?
You can do anything. How does that play out in terms of I don't know, you could cut it any number of ways, right? Are there roles that you might otherwise hire for that you just wouldn't hire for now? Are there uh what's the ratio of your own inference spend for the purposes of building Lindy compared to your payroll? You put your own lenses on it, but how has the arrival at the human or at the AI employee stage of this process changed what it's like to be a part of the company?
That's a great question. And I think right now, I mean, obviously, AGI is here. And so obviously, right now, every company is sort of scrambling, and there is a time of transformation that's exhilarating. And I will say, and I'm I'm I'm a little bit of a doomer. I'm I'm worried about the A I risk. But so far so good, and so far, it's so much fun. It's it's honestly so much fun. Like to go through a G I and to build this thing, a G I is like, we can move so much fucking faster.
I feel like we're moving at the speed of thought now. It's like there is no obstacle. Like normally companies, it's like, think about the strategy, it takes a long time to figure it out, to implement it, to see the results. It takes like months like that. Ooda loop is like months long, you know, and obviously, the the name of the game at the startup is to move as fast as humanly possible, because that's your only advantage against the incumbents.
And now we can just have ideas and see them live in the product two hours later, and like big ideas too, like not small ideas, and I can tell you, so our Slack now is basically just us and Lindy. It's like Lindy is like just half the messages on Slack is just Lindy and us talking back and forth with Lindy. Very often, like we, we ourselves, and we breathe this stuff all day, and like we ourselves, like we, we are underestimating the platform like just last week we we had like a problem with like our C I pipeline.
We have literally many hundreds of runnels on GitHub who like manage our C I pipeline. By the way, that's another thing right now to answer your question, what's it like to work at Lindy? Where's, where's it to work anywhere, frankly, is like every part of the stack and every part of your process is stretched to its very limits right now. That's just right out there, you know, like, I'm sure you've seen the graph like GitHub posted a graph of like number of commits posted per day.
A A GitHub was like, uh, hardly like a subscale company. It was really big. It's it's just going vertical right now. It's just going to the moon, you know, and that's that's like, I think, indicative of of what every company is going through. Our number of P Rs internally has, like P Rs per week has tripled over the last three months, and like the number of lines per P R has tripled over the last three months.
We hardly review P Rs anymore. It's just ages reviewing P Rs. We, in so far as we review, basically, what the the way the team is talking about is like, well, no longer P R review wheels, well, like P R review will review wheels. You know, we're, we're just reviewing the machine that reviews the P Rs. And as a result of that, you're like, this is awesome, but hey, now C I is in, is in the middle of that.
So what do you do? Well, now you've got to work on C I. And so we, we've spent a long time working on C I, like for the non-technologists out there, like C I, continuous integrations, like the process that's involved in taking into account your code changes and making sure they're safe and actually merging them into your main product and deploying them. And so we, we at first, we started throwing money at the problem, like many people do.
And so now our C I is extremely expensive. It's like it's a, it's a just insulting expense. And and and we've exhausted all the, all the obvious shit, like, we're obviously not running on GitHub runners, we're running on like G C P runners, and like we've we've done like caching of the bills, we've done all the obvious things. It's it's a lot of work to manage C I to make it, to make it efficient. And and and so we, I was tired of this because it was like weeks of messing with C I and and like building a hole in our pocket.
And were like, what are we gonna do about this C I thing? And then I was like, wait a minute, like, why are we doing all of this ourselves? Like, couldn't we just like spin up an agent to do it? And spin up the agent was literally, and so, by the way, that's the other reason why it's important to have a single agent. Because setup now is becoming one of the biggest costs. Because the biggest cost, the biggest bottleneck in any organization right now is human time and human attention.
So you wanna remove that guy as much as possible. So you wanna be able to provision agents and give them all the accesses they need without a human in the loop as much as possible, okay? And so is this agent at the company. It is this A I employee that knows everything there is to know about the company. Lindy teammate, right? It's got access to our repository, it's got access to, it's got its own like computer, and so, like, Lindy, can you please, like, like, mess with our C I?
And it's like, yeah, let me look at, like, the runners using the G C P C L I, let me look at like the GitHub like action and all of that stuff. So it's doing like a lot of analysis, it's getting back to us, and now it's literally, it's basically like an auto tune, like an auto optimization thing. Or like Lindy is just like sending us a a graph, like every day, by the way, using image gen to generate, so the graph is like really pretty.
And it's like, Hey, you know, guys, like, look, like C I cost is going down, time time to merge is going down, which is a good thing. And I was like, I can't believe it took us two weeks of just like messing with C I ourselves, instead of just saying this to Lindy, you know? So I think this is what it's looking like right now is like engineers are, are like, less and less working on the thing.
They are more and more working on the thing that is working on the thing. And more and more what we do is like we were setting up this machine, which is the company, which would just increasingly operates itself, and and increasingly, well, more and more like, you know, I like to think of it as like, first, you become a manager of I C s, then you become like a director of managers. Because at some point, like, the managers are gonna become gonna become AIs as well.
I'm hopeful that at some point we become board members and we can just go to a beach in Hawaii and we and we. We don't have much to do except like review the strategy of the company. So we'll have like reports of of the agent telling us this is what's going on. This is the landscape. You know, these are all the results we're having, and we'll be able to have much deeper conversations about the nature of the company we're building and the the the spot we occupy in the market.
I, that's a dream, as far as I'm concerned, and that's that's not very far away at all. To answer your question, have we have there been positions we've not hired? Like, yes, absolutely, you know, we we. The team is actually, I think headcount has been flat now for a while, and we've again, like I said, we've actually tripled our productivity over a couple of months. And I find there is a Goldilocks zone because humans are still in the loop, humans are still needed, but obviously, the more humans you add, the more coordination costs you you introduce inside inside the company.
And so I I find that actually right now, smaller teams, relatively speaking, are at an advantage versus bigger teams in in a lot of ways.
Do you? I'm sure you do. I don't know if you want to share it, but do you have a ratio of not? I don't mean all inference spend, like, including your customers use cases, but like your own internal inference versus your payroll. How I miss, is that ratio
I miss, yeah. So if you look at all the inference band, including our customers, it's it's many times payroll. And it's been for a very long time. But one exception, because that's what we sell, you know? Now, if you look at payroll versus internal inference spend, it's in shooting range. So payroll is still greater, but the the lines are gonna cross three to six months from now.
if you imagine a hypothetical scenario where all of a sudden it's just you at the company, and after that it's Lindy's all the way down? What breaks? What doesn't work anymore? Like, in other words, what is it that the humans are still irreplaceable for?
It's really hard for me to answer this question because, and I think about that a lot, because I am finding myself increasingly lacking the vocabulary to answer this question of like, where do agents fail? You know, like we used to talk about, like, time, you know, it's like, ah, time before, it loses coherence. It can go like the meter, I think it can go like one hour, ten hours, one hundred hours, and and I'm sure people will will resonate with this as well.
Like, I, I don't find that's that's a good, that's a good measure anymore. You know, the the wield that everyone is using these days is like spiky, right? Like these models are so very spiky. As we're going through A G I, the world A G I has stopped meaning anything because it's actually A S I in many ways, and it's it's it's subhuman in in some very surprising and dumb ways. It's really dumb in a lot of ways.
I'm like, Hey, man, like you, you can produce for me like 50,000 lines of code one shot. You can build me the most incredible ideas in like three hours by yourself. And then, you know, you, you, you're telling me to, like, walk to the car wash, you know? And so that's where it fails. Imagine you've got this this this machine that runs a company, super fucking smart, but but but fifty times a day, it decides to walk to the car wash, you know?
And so I think right now, like, the the challenge of these new organizations that are going to be for the foreseeable future human A I hybrids is going to be. You're basically building an Iron Man suit, okay? And so there is this slider where it's like, okay, this is for the machine, this is for the human, okay? And knowing where to put this slider, because it's not a one dimensional thing, it's like it's a complex line through a many dimensional space.
So you, you're looking for that slice that you can draw through the space, because you wanna make the most out of humans attention. And so you don't want the A I to bug you for stuff it doesn't need you for, and you need the A I to know when it needs to bug you. And almost definitionally, it can't, because if it knew, it wouldn't, it wouldn't bug you, it wouldn't, it wouldn't make the mistake. And so I'm sorry, I feel like I'm answering discussion with a question, but that's that is the right question, you know?
And and I think right now that's one of the open questions in the industry is like, how do you draw that line? How do you get the the agent to realize when it's on site and when to make a screw? Like when it's even saying something like, please walk your car to the car wash, because it's sunny today, you know? I think that that that's the answer. It's like, it's just, I'm, I'm somehow, humans and I were gonna have to lose hold system and at gas times when it's being very dumb.
How about the times where something is being very smart? Where do the best ideas come from these days? I think there's long been this idea that the AIs are good at routine stuff. I used to say, and clearly this is no longer applicable, but I used to say you can probably automate any routine work that you have at your company if you're willing to put in the elbow grease. But don't expect eureka moments. Now, obviously, we're getting eureka moments on the timeline all the time, at least in certain domains, like the open math conjectures and stuff.
When it comes to the best ideas at Lindy, how many of them are human origin and how many of them are A I origin today?
I hates the sense here, but it's, it's both. It's it truly comes from this, from the union of both. And the reason I hate it is because it led us into this mess of the the centaur. You know, it's like this mythical man horse creature. And so, you know, it's this idea. They're like, yes, an A I is better than the human, but you know what's even better than the A I is A I plus human.
Hence, humans are always going to be needed. That's a fallacy, that's just not true. And the the literature is actually clear that it's like, you know, we've seen it happen with chess and with every other game, which which A I has achieved superhuman performance on, where at first, A I beats human, and A I plus human beats just A I. And little by little, the gap of A I plus human versus A I is shrinking until it actually turns negative and humans are introducing at best, like random noise into the system.
So humans actually at some point become like truly clueless and and they start harming the system in which they are intervening. That said, the centaur phase exists at for a while, you know? So the the open question is like, how long is it going to exist? But right now, we are in the centaur phase. And I think that's the other reason why multiplayer is so important, because you want A I to be, you want your agents to be in the room with you, and you don't want any random agent to be in the room with you.
You want an actual A I employee who's been in the room with you for a long time, who sat in every meeting, who's built up the internal knowledge base about everything that's going on in the company. And you can at mention it, and you can invite it to chime in into your conversation in the company. And we do that all the time. We do it. Actually, it's funny, cause it's very easy to do it on Slack, where it's like, at Lindy, what do you think?
You know, it's hard to do in the meeting in a way that, like, now makes us actually more and more. Well, like, we prefer Slack because, like, it's the A I is there. Whereas, like, in the meeting, it's there, it's listening, but it's called chime in. And so very often now, what we do during meetings, we're like, actually, it's here, like, you know, struggling with this thing. We're like, Lindy, can you please send us a message after this telling us what you think?
And they'll send us messages on Slack. And and so when you do that, it is not the case. It is rarely the case that Lindy will first shot, tell you something and you're like, this is it, this is the thing we should do. But it is almost always the case that we, we, we go back and forth with LIndy and and from that back and forth, like it misses something. Like the C I example is a great one, actually.
At first, when we, when we invited LIndy into this conversation, it happened over Slack. It was like, wait a minute, well done, like, why are we not doing this with LIndy? Add LIndy, can you please optimize our C I? And and it, I
Make no mistake.
could not find it. Yeah, yes, and I forgot its first suggestion, but it was, it was, it was a reasonable suggestion if you did not understand our C I. And it was, it was, it was, it was not good, like, given given the constraints we're operating under. And so, like, the engineers chimed in, they're like, no, actually, we can't do that. We sat about that. We're like, ah, that's a fair point. So we went back and forth, and then and then together we came up with like, oh, yeah, yeah, we should totally do that.
And they just go ahead and do it, please.
A. A thesis that I've had for a while, it might be too early to say, but you're definitely gonna have a perspective on it is. But again, going back to what I've been wrong about, I expected a lot more transformation than we've seen in the economy. If you'd showed me fable in 2022 and said, what's the unemployment rate when this model is out? I would have said, I don't know, but something definitely higher than what we are currently experiencing.
So then why? And this has been something I've been wrestling with for the intervening time, like, why are we not seeing more? And one answer I've come to is people are stuck in their ways, and not everybody's an enthusiast like I am, and it's gonna take time. But then also, when we get to the drop in knowledge worker, that'll be the time when it'll suddenly flip right. Because now I'll have a more parity kind of choice between I could try to go hire a human or I could plug in an A I.
And that's gonna be easier in a lot of ways, right? I don't have to go through a whole interview process. I kind of know what I'm getting, blah blah. I don't need to tell you why an A I employee is a good form factor. Do you think we are starting to see this flip? And is it going to be something that incumbents will be able to adopt in time to avoid getting disrupted? Or are there still gonna be bottlenecks or resistance points that I'm not anticipating?
I expect it will disruption as well. I think as long as the models are spiky, we're all going to need a lot of humans in the room. Because the holes in the model are very dangerous. Not not just like A I existential dangers, not just not just harmful, they're really harmful to the system. And so, however expensive humans are, they're going to earn their keep by plugging these holes. Now, regarding whether incumbents will be able to adopt this this thing, I think as long as it is spiky, I think the the incumbents are going to be at a disadvantage.
If you had a true drop in remote worker with no holes, then I think even the incumbents would be able to adopt it quite quickly. Because it's just like hiring an employee and they have processes for that. And I actually think they would be south on on a lot of the of the innovation because it would it would be skeuomorphic, effectively. You would, you would just have like A G Is like sitting in human seats, effectively, and I don't think that's the best way you build an A G I organization, but it would work.
The the smaller organizations are going to be at an advantage, both because they are going to be able to, to short term figure out this A I native organization, where the human is filling the holes, and and long term, I think they're going to be to build like a less skeuomorphic A I organization. You know, economists are very fond of talking about that. I actually wrote, uh, a blog post about this that's called the Duff-to-Mato principle on my blog in in 2018 or so.
And it's it's about how, like, every time you have a technological revolution, you think of the new paradigm in the in the terms of the old paradigm, you know, and and and one tale that you're, you're, you're falling into that trap, which is very natural, is that you're, you're, you are calling the new paradigm after the old one. So, you know, the the first cars were called horse-less carriages. Right now, self-driving cars are called self-driving cars.
Right now, eventually, we're gonna have a a wheel for them, you know? That's not, that's not in the terms of the old paradigm. It's almost like right now we're doing it with A I employee, huh? Like, it's, it's like A I employees is a strong as as, as a term, it's a very strong sign that you're doing something very wrong in how you're thinking about your product. And I think, I think it's, it's fine if you're doing it just to communicate about your product.
Because, like, the nature of of positioning is that you have to position against something that already exists in the market's mind. So the the car exists, you know, the the carriage exists, sorry, the carriage exists, the car exists, and now the employee exists. So you can't really invent a new like wheel that are nowhere. The iPhone did that. You know, it's like the iPhone is obviously not a phone. It's so much more than that.
Like, you know what percentage of of your iPhones use is just like making phone calls, like 2%, you know? But it's called an iPhone because it's an electronic thing you've got in your pocket, right? So I think A I employees is one of those things. It it takes a very long time for industries to find the the message, which is native to the new medium. I think it was Steve Jobs who was saying, like, you know, the first, the first content on T V was just either like recorded radio talk shows.
They took like a radio show and they put the camera and now you've got a T V content. That is what talk shows are, or recorded theater plays. And it took a surprisingly long time, like ten or ten or twenty years before people realized, like, wait, I can move the camera, wait, I can have multiple cameras, and I can cut. You know, I can do, I can, I can change the scene, I can do, I can do a lot of crazy stuff now.
And And it's treacherous because it's thankless work. Once you've realized it, you, you, it's so obvious. And this is why when you, when you watch like old movies, like Citizen Kane, you know, or, well, Hitchcock has a lot of those, like, you you're like, there are nothing special. Why was everyone crazy about Citizen Kane? It's like, well, actually, it was groundbreaking at the time, like the Beatles are like that, you know.
It's like, well, you know, it was actually really innovative at the time. So I think, well, right now, also in that phase of like the A I employee, we've got the A I employee. And so now, once the phase of trying to figure out what does the A I organization look like, and I think that's, that's a young man's game, and that's a young company's game. I think, I think incumbents historically have have really struggled to answer this question.
Do you have any previews that you feel confident in, or is it too early to say? Or maybe you would say, are there any thinkers that you think are kind of notably ahead of the curve on this?
The only post I have seen and liked about this was from Dorkesh. It was was a while ago, it was a year and a half.
That comes to mind for me too.
Yeah, yeah. He asked like, what does the A I native organization look like? And he, he points out, like, I think it's, it's an excellent intuition pump. He points out, he's like, look, you know, Sundar at Google is paid $150 million a year, or something like that, you know? So obviously, there are agents which companies are happy to pay $150 million a year, even though their token throughput is actually low.
Like, it's not like Sundar, it's just like I'm eating 2 billion tokens per second or whatever, right? It's just like he has really good tokens, right? And so now, so what does that mean? Like that that that it indicates something about the willingness to pay for models and for agents. There is this book by Robin Hanson. I'm looking at it right now, The Age of Em. And it also explores a lot of those themes, right? It talks about, have you read it?
It's excellent, highly recommended. Yeah, yeah, I'm laughing And it was the public.
only because I tried to do a podcast with him about that book, and it went in a very different direction. So it just brought that memory to mind. But I think that book is excellent. I think it is a as a primer for or as an example of how to take a premise and really run it out. It's elite.
Yeah, it's a book-long style experiment. So for people who don't know what it's, what it is like, M stands for emulated mind. And it's like, hey, we are going to have the hardware to create emulated minds, you know, in the like, the late twenty twenties. So like, we're right on track. It's like, well, just not gonna know how to do the software. So all we're going to do is we're going to emulate, like, literal human brains.
Not exactly what's happening, but kind of what's happening, you know, like L L M s are trained on human tokens, and so that's basically what's happening. And so it's like, and it goes on all of those riffs of like, hey, this is all the crazy stuff you can do when when this happens. I'll, I'll talk about two of them because I don't want to take the whole time and and everyone should really read the book, but just to give a taste to people about, like, what what it sounds like to to have those native A I organizations and the kind of stuff that you can do that you can't do when you have a human.
The first one is splitting. And so I think this is why this, this touches on what I was saying earlier, like, don't create multi-agent organizations. As much as possible, try to have as much of your internal work as possible. Emphasis on internal. We can get back to that later, but as much of your internal work as possible done by one agent and one only. Why internal only? Because if that one agent, which has got the keys of the castle and access to the bank account and the repositories and the secret keys, is also the same agent that does customer support, you, you potentially have a security issue.
So it, it can be, can be good to have multi-agent systems, be it for like, only for like security reasons, but okay, so you have that single agent. Now, that single agent, you, you know, it, it, sure, it goes faster, it emits more tokens per second than a human, and it doesn't take a break. But at some point, you're gonna need more tokens than like a single L L M loop can do. So what do you do?
Well, you have multiple L L M loops running, and it's basically just like that agent duplicating itself, like forking itself, right? And creating sub-agents and like workflows, dynamic workflows, and all of that, right? So, and that's one thing that Robin Hanson talks about. Like, you know, you can imagine you ask one, literally in the book, he says, you could ask one emulated mind to build like a really complicated thing, like an operating system, which is like billions of lines of code or something.
And and at first, it would just like take it to the highest level and like come up with like the highest level architecture of its operating system. And then it would fork itself, and each subcopy would be in charge of like one building block of this highest level thing. And then each of them would recursively keep splitting themselves such that it's a, it's a fractal. It's like every, every, every node in the graph has got, like, the whole picture in their head, and every node in the graph is working on an implementation.
And then they all bubble back up, and you could imagine multiple passes up and down, and in three hours, you've got an operating system. It sounded like such a wacky idea when he wrote the book ten plus years ago. Now I think it's it's very clearly what is currently happening, right? So that's, that's one thing that's happening, which is, which is really interesting. Another thing that I've not seen that yet happened with with L L M s, but it's an interesting thought experiment about what does cross-organization collaboration look like.
And what you can do is that you can create unfalsifiable agreements. So suppose I came to you and I was like, listen, I have information for you that I cannot tell you, but I can tell you that if I could tell it to you, you would agree, and you would give me all your money, right? Yeah, it's like, you wouldn't give me all your money right now. It's, it's, it's a bit too easy, you know? But if I could prove that to you, if I could prove you with with absolute certainty that yes, if you held that information, you would give me all your money right now, you know, it's just I can't give it to you.
The way you do that with M s is that you, you clone yourself, you clone me. We put both of us in a in a box that's going to self-destruct and they duck, and your and your M has a button, and that's the only, the only communication means it has with the outside to be like, yes, give him all your money, right? So it's a copy of you, so it's it is you agreeing. And so you have this zero knowledge proof of of crypto bros.
Well, well, like, right, I guess on this one, you have this ZK proof of of that. So that's the kind of thing that's on my mind, and I think that's the kind of thing we're going to see exist soon in the next few years with A I native organizations. It's about to get weird in a whole bunch of ways, I think. And Oh, yeah,
you better hope they don't break out of that box, too. That's the other worry one might have is we put our agents into boxes these days. One, let's do one more beat on the stuff that's relatively mundane, and then we can zoom out to some some real big picture considerations. But you did have, as you alluded to earlier, this highly viral post, I don't know, two months ago, maybe moving a significant part of the workload to open source models for cost-saving reasons, and for, you don't need God to schedule your meetings reasons.
But where are we now? It strikes me that, like, when you talk about just having one agent, that kind of seems to go against that. There's also the sort of can advantages which you lose when you are crossing agent providers too much. And then, of course, there's also proprietary providers have their families of agents, and they're like increasingly training their best models to delegate to their Hiku's respectively. Like, where do we shake out on this now?
Is there still a a significant role for open source cheaper models to play in the Lindy teammate, or are we back to heavily cloud-based agents today?
no, we're quite open source build. I like, look, it's just so much cheaper. It's like, it's ridiculously cheap. Like, and like, the the cheapest of of them that we really like is DeepSeek Flash, which is it's free. That fucking thing is free, you know? And so as you, as you experiment with your agent and as you spin up a lot of them, like the the bill can really go up pretty quickly. And DeepSeek Flash is just incredible, you know, and quite fast as well.
You know, we, we do find that as you go to like the upper echelons of intelligence, if you look at like a Kimi C three or like a G L M 5.2, like, I still am super impressed by those models, and they're awesome. And we, we are increasingly considering them as like our main driver for for a lot of parts of LIndy. But I will say that the gap narrows between those guys and like the frontier guys, like Kimi C three is is frankly, not that much cheaper than like a Sonnet or like an Opus.
So, but no otherwise, I mean, like, I, I, yeah, I, I do think open source models are like, if you care about price, and if you don't mind using those providers, like, yes, you do have stuff to figure out around caching, like, yes, the the the inference providers are like, not as good at caching, but they're, they're catching up. And And look, you know, if you look at like a deep seek, deep seek flash, it's Sonnet 4.6 level-ish, a bit less, but kind of ish.
Sonnet 4.6 is a really good level, a really good model for like most use cases. And it's literally a hundred x cheaper, you know? So even if you, even if you miss on two x because of caching, you're still fifty x cheaper. It's a really big difference, the difference between spending like $1,000 or $50,000. So though I I think open source models are like a required part of the stack right now for for anyone who's like seriously building and operating A I agents.
So how do you think about deciding where to integrate them? Especially because it didn't sound easy before, but in a context where you have workflows and there's nodes in the sort of graph of work, you could go into a particular node and say, okay, we kind of know what the inputs here are and the outputs. And it's like relatively controlled environment, and so we can do structured testing. You did all that, and you still reported having some, uh, false positives over time, where a model could pass a bunch of tests and you would feel good about it.
And then, if you start to test it live with users, you'd get the feedback that, like, hey, like, it got dumb and I'm not, you know, somehow it was like still hard to measure, even with like a much more structured environment for the A I to work in. Now, as you're in this very open-ended, just tag Lindy in Slack and send it anything. Um, it seems like that problem would have increased in difficulty dramatically. So how are you approaching it, and what are we learning in terms of where the open source models?
And it's not even so much. I think Anthropic does make some good points about this sometimes. Obviously, there's more to the story, but I do think they they're apt in some ways, where they say, in some cases, it's less about open and closed source and more about just like, what's capability level, what's the price, what are the the features? So regardless, I guess, of whether you're even going to open source or just going to haiku, how do you decide when you can do it?
Great question. We have found you should almost never use multiple models powering the same agent. And we have found one exception to this rule is when you spin up new blank sub-agents. Emphasis on blank, because those two types of sub-agents. You have a blank sub-agent, then you have a forked sub-agent, which inherits the context window of the of the parent agent. And the reason why you don't wanna do that is because of caching. We, we, we just obsess about caching, obviously, cause it's just so expensive otherwise.
Like, the economics do not, they barely work with caching, they just cannot work without caching. So caching is is a must-have. And and so I'll give you an example, the validator system that I mentioned. So it's this L L M as judge. By the way, highly recommend anyone who builds A I agents. Like, this is one of the lowest hanging fruits you can do to greatly increase the reliability of your of your A I agent. And so that's step one, just like have a validator, which is like the naive implementation.
Is like you ask agent to do something, it does it, or like it, it submits an action candidate, you intercepts the action, and then you ask it, are you sure? You know, literally, if it's just, are you sure? Already, you get a bump on your evals, which is insane. It should not be the case, but it is the case. Now, if you increase, if you change this are you sure to like an actual prompt, which in our case now is like 10,000 tokens, it's a really big validator prompt.
If you change that, then you, you actually, you're, you're giving it a checklist, and now it's got reasoning tokens. And and now it, it, it's way better. Like you can get Sonnet to perform above Opus level, you know, in our experience. Now, then you can get, you can go one step further, you can, you can have like a federated suite of validators, and some of these validators may be deterministic. So, for example, one thing we and and the rest of the industry has found is that Sonnet, I mean, cloud models this year have been getting dates wrong by one day, right?
That's one way in which they're spiky, right? It's like, hey, you've got your A I organization, it gets dates wrong. It's a problem when, like, us, you're building an A I executive assistant, which have its job is to schedule meetings, okay? You can't get dates wrong. What we did is that we, we created, it's, it's a modular architecture we have. It's really simple. It's just like a bunch of validators, and then it's like a promise all with a timeout for for people who know what this means.
And and so every validator has like a second to like, decide what to do, and and if they've not submitted their verdict, then time's out and the agent just submits the action. And and, and we have a validator, which job it is to detect dates that were submitted in the action. And to, and, and we, we've prompted the agent to to include a weekday into the date. So it never says, never says July twenty-eighth.
It says Tuesday, July twenty-eighth. Okay, if you do that, then you can have a deterministic validator that checks whether the date of the week matches the the day, the data was submitted. And that's just a regular expression. It's instant, free, no A I in the loop, and that's, that's one of those federated validators. But even even the validators that are A I powered, you don't want. Initially, what we did is we were like, what if we had Sonnet for the main agent and we had, like, DeepSeek Flash for the the sub-agent, the validator?
But actually, because when you, when you have a cache hit, it's ten x cheaper. You, you, you actually, unless the model you're going to use is more than ten x cheaper, which it may in the case of DeepSeek Flash, it may not, because the caching is inferior. You know, if that's the case, you actually do want to keep the same model. And Sonnet is so much smarter, like it's actually worth it to just like keep the same model.
That that also holds if you fork your your agent, like that way you can, you can recycle the cache. By the way, interesting note, how do you reuse the cache when you have this validator? You know, because the problem is that changing the toolset invalidates the cache, okay? So the way we've done it is that the validator and the agent has the validator action always included in its toolset. And it's unable to invoke it. We tell it, don't, don't invoke this guy unless you're the validator.
And if he tries, we just, we're just like, nah, I'm not gonna, I'm not gonna listen to you, I'm not the validator. And then the validator now inherits all the actions of of the agent. And it's like, you are now the validator. You may invoke this action, which you have known about the whole time. So that's how you don't break the cache for this kind of pattern. Yeah, and so forked, forked agents, same, you just use the same model.
You know, I I've heard friends, founders who've told me, I don't even want to check if it's true, I'm sure it is, which just gags me. If you, if you take an agent and you change the model at every turn between roughly equivalent models, so one turn it's Sonnet, one turn is Grok 4.5 or whatever the latest is, and one turn is like GPT 5.6, and one turn, it actually increases the performance.
Is that ensemble, and it's literally just like a random sequence, okay? That ensemble, for some freaking reason, outperforms any given model. I don't want to know why, but apparently it works. Again, we've not even tried it because we don't want to break the cache anyway. Okay, that's really interesting detail. Can we bottom line it a little more? It sounds like top level cod is still the kind of core main driver because of the caching. Claude is also often the validator.
Blank sub-agents you can put onto a DeepSeek flash. This sounds pretty Claude heavy, all things considered. Is that fair? No, oh, sorry, I was not realizing you were still asking me, like, what model we're currently running on? Right now, we're running on DeepSeek. DeepSeek So is the main do the so right now.
the it's the driver. Wow, okay,
It's, it's, it's the whole thing. Everything is DeepSeek right now. We are in the middle of reconsidering it because we are realizing that teammate, the new product, is is requires beefier compute. And by the way, you can always go in your, in your Lindy settings, like we are model agnostic, so you can always select Sonnet or Opus if you want to. By the way, like, one interesting learning of mine is like, it doesn't matter how often you tell people, hey, like, we promise on the benchmarks, it's the same.
There are like quite a few of our customers that are like, I don't care, I want Sonnet, you know, or Opus, you know, which is very excessive. But no, right now, it's DeepSeek by default.
interesting. So you feel like the so, yeah, I guess how would you characterize the gap between American A I models, Claude, perhaps specifically, and a deep seek? We hear we, I think we kind of go through these cycles, right, where it's like, oh, the gap is closed, oh, maybe it's open again. It's no, actually, it was totally on trend the whole time. It was always nine months. Qualitatively, though, for the purposes of actually making an A I employee work, how would you describe the gap?
It's more spiky. It takes more turns very often to find something that works, so it still ends up finding something that works, but it ends up taking more turns, which is slower and more expensive. I know you asked me for something qualitative, but look, it's it is like three or six months behind. So right now we've got Sonnet five, like DeepSeek is basically Sonnet 4.6, is the way I like to think about it.
How much do you worry about when people are allowed to change the model? I always have this question of first of all, are we in a period of convergence of the AIs or divergence? I'm even on that basic question, I'm, like, sometimes confused. And, uh, a sort of distinct but conceptually related question is like, to what degree do products have to be co-designed or co-evolve with models to really work well with them? You could imagine that switching to Sonnet might make it worse because of all the work you've done.
And in fact, I've heard that about Opus five, and I've heard a lot of different things about Opus five, which I don't think coherent to anything. Really easily summarizable yet, but I have heard from some quarters that, hey, you probably need to rethink a lot of your system prompts and whatever with Opus five. It may not follow your skills very well. It can be more powerful, but you've got to kind of go back and redo things.
So, yeah, how do you find that playing out in practice?
we, we do find significant differences, surprisingly significant differences between the model families. I was expecting, I think I was, you know, you and I were talking together years ago, and, you know, at some point, we're going to come to the same region of the space. And that's not happened as fast as I was hoping. And I'm hoping for it because obviously that makes the models commodities, as far as I'm concerned, makes my life easier. What is actually happening is that the different model families have significant gaps in how they interpret your prompt.
So by family, I mean, like, Cloud and and Open A I, and even Meta now is sort of back in the game, and Grok, like these guys, these guys are quite different in the way they, they interpret the prompt. And then indeed, whenever you have a new major jump, five was a really big jump. It was, you know, 4.7 and 4.8 together. Like there was a change, I think, in the tokenizer, so it was very different compared to 4.6.
Five is bigger than that gap. And so what we've, what we've found, and and we've been lucky enough, this is part of the tooling that I was mentioning mentioning earlier that we've been investing in is like we have our own G P A M self-optimization loop. So we have hundreds and hundreds of evals, I think at this point more than a thousand. And we have an optimization loop where it's like an agent that runs the evals and finds, like, tweaks the prompt to maximize the score on the evals.
And by the way, you can bypass it, like everything you can bypass now. And yeah, I mean, every time a new model comes out, we have to run that loop. And you give it a budget, and at this point, we have to give it like thousands of dollars, because it's a lot of evals to run. So it's like about like $10,000 every time like a new major model comes out. So it's like, hey, here's ten grand.
Just re-optimize your prompt around this new model. And we do find that there are a lot of updates, like the prompt does change quite a bit every time. I will say, is that open source models are all very cloud-like in their behavior? I wonder why.
Yeah, would I definitely have some Christopher on that? When you do this optimization, are you? Is this also like a homegrown framework, or are you using like a D S Pi or the I forget the name of the kind of spiritual successor to D S Pi?
no, no, no, it's all homegrown. Yeah,
But it is a sort of auto optimization process where you're like, here's an eval set.
that's right.
Auto optimize your own prompt to climb this set of hills that we've got for you.
That's correct, that's correct. G-P-A is like, um, generative parato frontier or something like that. It's, uh,
Yeah, that's the one.
it looks, it looks for like the Pareto frontier. So it just it's looking for like the best prompt across all of your eval set. We're looking into tweaking the system right now so that we can assign weights to evals. Well, like this eval counts as like ten of these other one because it's really important. But right now, it's just like treating every eval as equal.
How fine-tuning as a dimension in this whole model situation. This is another thing. If I go back in time, I'm like, I definitely expected a lot more fine-tuning than we are getting. Open A I is retired, it's or on the verge of, they've certainly announced, and I think maybe at this point have pulled the trigger on retiring their fine-tuning product. We've got thinking machines trying to answer that call with their own model that's specifically to be fine-tuned.
Is that gonna be part of the future of Lindy teammate?
yes. I think fine-tuning is a last resort. It's something you do once you've exhausted every other option because it's such a pain in the ass and it's expensive. But it's it's gotten a lot easier because now we have A G I, and so you can just ask Cloud to fine-tune for you, just give it a, just give it a data set. Bringing the data set together is still like complicated, and like sanitizing is complicated, and it's, it's, it's still a pain.
So it's, it's, it's a last resort. It's, it's not the low-hanging fruit. You should do everything else before you fine-tune, including D-B-A. But then, yeah, at some point you get there, and this is why, like larger companies are the first ones to fine-tune, because they, they have existing, zero-hanging fruits. And at some point, and and they have the resources to fine-tune. And so it, it, it does make sense. It does give you more performance for cheaper.
Do you think that that, like, how broad do you think people should be thinking when they are approaching fine-tuning? Obviously, the extreme would be one fine-tune per task. You can definitely do multi-task fine-tunes. It's maybe getting into pretty challenging territory to say we wanna make our own general purpose model that has the same breadth of action space as the the big ones, but it's like ours. Somehow, how would you guide people on, like, how big to think when they start to approach fine-tuning?
Yeah, most people should not fine-tune. I think if you work at like at scale, you should, you should fine-tune, and then the more at scale you are, the more fancy you can be. But I generally one meta heuristic I've developed over the years is like you should place an enormous emphasis on simplicity, enormous emphasis. And and I think it's just too complicated to start to have multiple fine-tunes for different multiple use cases and users and and model radios, like you don't want, no, no, no, just one model.
You know, you don't like I said, you don't switch model mid midstream, and if you fine-tune, you just fine-tune into one model.
Turn, okay, cool.
I think, I think one again, this goes counter to this meta heuristic I mentioned of extreme simplicity, but the thing I find myself going back to all the time is like, because I spend so much time thinking about context and memory. Right now, our memory is like these millions of tokens stored on on a file system and this memory agent that's really, really sophisticated. It does feel like memory belongs to the weights. This does feel like a little bit of a hack.
And so there is something to be said about that. I think eventually, with infinite resources, what I would like to do is I would like to do a LoRA per user, because LoRAs are actually pretty cheap to train. So now it's no longer mapping. Now it is actually dreaming because you can't retrain a LORA every fifteen minutes. You would have to do it every day or every week, you know, and and you would, and and it's a tremendous infrastructural challenge, even to like, so you got to store all of those LoRAs, which like the storage is not the problem, but like the inference strain, swapping in and out of the LoRAs, is a huge pain in the ass.
And the training pipeline and and all the other stuff. So good luck doing that. So, but it would make sense, cause I think weights are are so much more compressive than than tokens.
Yeah, that's I don't know if you have any thoughts on the great, uh, horizon scanning for continual learning, but this is again, for quite some time, has been the thing that I'm like, boy, if that ever tips with, and it could be as simple as one key insight, we could be in a very different world very quickly.
I have a feeling what what I'm describing is is a step towards continual learning, but it's not, it's not the final step. I think the the final step, I think, I think like the the bitter lesson would have you, like, inference and training need to be one and the same, like we can't separate them, you know? That's, that's sort of what I think the entire field is looking for right now. I agree. Once we get there, it's going to be crazy and scary, and very, very, very different.
But yeah, I think, I think we are going, the the LoRA thing I just mentioned, like, look, it's it's an engineering problem, you know, you can figure it out, especially now that two have A G I. So it's, yeah, I think, I think, I think the the LoRA thing is going to happen soon, in the next six months. I think even like one of the Frontier Labs may do it.
I thought that was honestly one of the things that Open A I did extremely well with their fine-tuning product. They were clearly doing something like this because they would allow you to do the fine-tune and then you'd have the same rate limits with your fine-tuned model as you had with the base models. And I was always really impressed by that engineering accomplishment. And now it does surprise me that they've gone away from it so much. But I also agree with your guidance.
I would definitely tell people, don't rush into fine-tuning these days. It's a slow cycle and you could probably get what you need for, uh, less total effort. Available models that you don't have to monkey around with in that way.
100%.
So you talked about scary stuff, which is maybe a good transition to, uh, kind of not second half, cause we've been at it for a while, but, uh, yeah, second phase of this conversation around just zooming out and looking at the super big A I picture. Maybe I'll let you choose the order of topics when it comes to Chinese models and what, if anything, should be done about them. Because you did recently post something I think quite counterintuitive, given the fact that you're running your company on DeepSeek, I'll let you state your position there.
But should that come before or after the big picture? Where are we in this kind of, oh my god, we just had open face happen, uh, thanks. what does it mean and what should be done about it? Yeah, I'm trying to coin that, and we'll see if it sticks. I searched for it the other day, cause I just got back from China myself, actually, and I was like, how has nobody called this open face yet? I'll be the first one to try.
How about open gate? Cause that's what it is. It's an open gate.
I'm sticking with open face just out of pure, you know, shock value, I guess, if nothing else. But yeah, we can fight it out in the marketplace of ideas. I guess you tell me what we should talk about first, and I guess it depends a little bit on, like, whether the Chinese model argument is like upstream or downstream of your kind of big picture concerns.
The open face incident is immensely concerning. I think it's the most concerning incident I've seen happen so far. And I, I know that my my feeling is shared in the labs, like my friends at the labs are some of them are panicking, like there is an air of panic right now, like intense fear in in the air. So that's that. Then regarding so that's, that's where we are, you know, right now. It's like, July 2026, we have A G I, we are in takeoff, and we've not figured out alignment, that's the T L D R.
You know, the the current timeline looks much too close to like an Eliezer Yudkowsky's essay for comfort. Now, regarding the Chinese models, and bread and news, Chinese models. Look, I, I, I, uh, okay, I'll start by saying that I have been bemoaning the low quality of the discourse here. It's, it's been, it's been very disappointing, cause I I thought tech was different. You know, like, the quality of the discourse was something for the woke, right, the culture war.
I'm like, ah, whatever, you know, but then it's this is tech, this is nerd stuff, and we can't get our shit together and and just remain polite and civil in the marketplace of ideas and address each other's ideas at the objective level, like meaning, like, don't attack each other's intentions. Can you please just attack address the argument that just put forth? Can you please not pretend I said something I didn't say? It's, it's ridiculous. That's what Pete just put out in this, uh, which has bolded, we do not support the ban on open source.
And then people are like, oh, I can't believe, like, quit tweeting this thing. They've obviously not read, and they're saying, oh, they're supporting a ban on open source. So I'll just start by saying that, so can you please, can please everyone take a chill pill? Stop calling everyone a retard. I recently put out a blog post where I was like, like, hey, I I think Chinese models should be banned. And I was called a retard fifty times by supposedly smart and accomplished people, including like, famous VCs.
Now, a lot of other, like, famous, accomplished, smart people reached out in DMS and by message, and there was a lot of support. My position is basically entropics. I hate saying that because people are saying, oh, you're just an entropic shell. But look, I have timestamps. Like, I've been tweeting my my entire position, swear the whole thing before Entropic clarified theirs, and it made position is exactly the same. I don't have anything against open source. I may, but I'm undecided.
I do think open source might increase existential risk. Let's put that aside and undecided. This is not the crux of my position right now. God bless open source, important for innovation, important for companies, including mine, like, please open source, okay? I have something against Chinese frontier models, whether they're open or closed source. And here are the reasons. Or number one, they're obviously distilling. It's very clear. And so you're putting American open source model companies in an unfair competition and and closed source in an unfair competition because they're not allowed to distill.
You know, it's just, it's just contrary, at the very least, contrary to the terms of service, and it may be illegal if you've, if you've put in place technical measures to circumvent any protection that the model company put in place to prevent distilling, which which the Chinese models obviously have done. So it's, it's distill, and it's unfair, right? Now, some people say, yes, but the companies have also distilled on human data. That is not what distillation means.
That's just not what it means. Like, if you, if you do that training on human data, it's costing you billions of dollars, like actually billions of dollars. If you do it on on A I data, it's costing you hundreds of millions at best. So it's, it's a huge unfair advantage, and it's a significant enough portion of your costs to to to confer you an unfair advantage. That's number one. It's unfair. Number two, very pragmatically, we don't want Chinese models operating in the U S.
Today, you know, it breaks my heart because I'm an American citizen, I'm a proud American citizen, I'm a hawk. When I talk to my own product and I'm asking it what happened in Tiananmen, it tells me, I'm sorry, I can't talk about that. That's a problem. You know, like these models are eventually subject at the end of the day, they are subject to C C P censorship and C C P policies. You don't want those models.
This basically amounts to being the greatest instrument of foreign propaganda on American soil ever. And there is just ask your L L M of choice, ask Claude, Hey, tell me the precedent we have of banning American like foreign influences in the country, like the three radio acts of the twentieth century, like the TikTok thing that just happened last year, like we, we do that all the time, you know, then these models are not merely just an instrument of propaganda, they're agentic.
They're actually doing stuff in the economy. You don't want the C C P to run chunks of the American economy, duh. Um, then finally, even if none of that was the case, maybe those models are playing fair and square, maybe they're not representing foreign interests, maybe they're just better. And so here a lot of people are, I think, indulging, and I say, I say that as a libertarian, that they're indulging in what I call, I call, naive liberalism, right?
It's like, well, let's beat them in the marketplace then. And I'm, I'm like, you know, um, like, I, I actually, and again, I say that as a libertarian, like, I, I. Protectionism is not always a bad thing. I think you do want to protect your, your, your domestic A I champions. And so people can't say that and they claim and I believe them that this is not their intention when they do that. And I will 100% believe them because they've been so consistent, while literally the founders have been worried about X risk for ten years, right?
So they've been so consistent. But look, I can say that because another thing is topic. Hey, we want local A I champions. Duh. It's a matter of national security. We don't wanna hollow out the base. We've hollowed out our industrial base. You know, so it's not about open source versus Chinese models. And uh, it's about propaganda, it's about nothing. This is good for our economy, it's about fairness, and the and about protection. All right, let me try to give you some counter arguments and you can respond to them.
I am a little more sympathetic, I think, off the top, to the argument that there was some massive reappropriation of human knowledge that is upstream of all A I. And you could say that's not distillation. Sure, you could say they the American companies spent a lot more than the Chinese companies are having to spend. I think that's also undeniably true. But if I'm just kind of what feels just and fair, I'm especially because we're blocking them from using cloud, right?
We're not, it's not like we're saying, hey, you can buy all the cloud you want, right? So we've said we've got cloud, the the the makers of cloud have advocated strongly for chip restrictions, and they also try to refuse to sell their model into China in the first place. And the whole premise of cloud, or the whole existence of cloud, is based on the idea that we've hovered up, hoovered up all this human knowledge in every form we could find it from digital, and you gotta believe that includes the whole Chinese digitized heritage.
Now we're like sucking in the books, which I don't care about the fact that some books get destroyed in this process. But this wasn't something that everybody consented to. And so it does feel a little strange to me, given all of those fact patterns, that we would then draw the line and say, okay, like it's Anthropic's consent, that's the consent that really matters. Now I think they should be permitted to do business with who they want to do business with, and if they want to put in measures to try to prevent distillation, I think that's their prerogative.
But I'm like, not at all convinced that it should be the state's job to come in and engage in sort of statecraft to try to punish or prevent this distillation. To me, it's kind of like, I don't know, information wants to be free. Knowledge tends to diffuse. It's not like Anthropic has a super or any AI company has like a super moral high ground in terms of I didn't get my check for my share of the training data.
And they're getting paid for the distillation queries too, which whenever they don't want to be paid, so it's probably beside the point. But I don't know, I'm not, I don't know, I don't find myself a bit on the side of the underdog Chinese here, where it's like, man, you've got kind of a deck stacked against you. And get some knowledge where you can get it. What's wrong with that perspective? I think we do want the the deck stacked against China.
We don't want China to win the race to A S I. I I'm not a lawyer, so I don't want to opine on the exact legal path that you may take to to make your terms of service enforced. I I'll just observe that, and I say that as someone who used to work at Uber. It's extremely hard to bring justice through the judiciary against Chinese companies, almost impossible. So, not a lawyer, I don't understand the technicalities here, but I'll just observe and start with that.
And it may take a very long time, and by then the lot of damage is done. And then I'll also observe, like, yes, companies are training on all of this corpus that is available to everyone. That is an and even playing field. The problem is that once you've done that at at great expense, it does cost them billions of dollars. If you can, if you can, and you, and you create that artifact out of this training data set, which is called the model.
Now, other people can turn around and and instead of doing this, which costs billions of dollars, they turn to this, which costs a lot less than that, and they copy you and they catch up with you. If you do that, you kill innovation. And that's that's it's nothing new, it's just called I P and patent law. It's exactly what humans do, right? Like you as a human, you're like a researcher, you, you think for like decades, you do all of that work, it costs you a lot of money in R and D.
You come up with an idea, which is much smaller in tokens and much more precious and much cheaper to steal, as is, than like all the corpus of stuff that you trained on. How do you protect that idea in order to recoup your investment in order to produce it? It's called a patent, it's called I P. Right, it's nothing new. If you apply the same standard you're applying now to, for example, the pharma industry, you would have very cheap drugs, which is awesome, and you would have no new drug ever.
You would completely destroy innovation in the pharma industry. And I think this is why. You know, finding all of these people support open models because people like free stuff, and it's it's you can't really measure the future innovation that you don't get as a result of that. All you see is you get free models, you know, and look, I'm one of them, you know, so I'm I'm I'm speaking against my interests here, right? Like, my company is dependent economically on those, like, very, very, very cheap models.
But I'm also in in a coordination problem right now, because, like, I cannot not adopt these models while the models are out there, because my competitors are going to do it. So I have to adopt them if they're out there. I wish they were forbidden across the board, so that we would like, like the four reasons I invoked earlier, like, protect our our, like, champions, not have the C C P like influence the country or run parts of the country, and and have, like, a fair level, field of competition.
I guess another simple argument is that I think our frontier companies are doing just fine that, you know, that could change perhaps at some point in time and I, you know, as the facts change, I I think our response to it, you know, might also ought to change. But I don't find it super compelling at the moment to say, you know, we need to protect, like, Anthropic's revenue run rate, like, they're, you know, Right, the they can barely pro leave the serve the demand.
tree and there's a load. No, that's fair, They that's fair. Yeah, but you're still left with the counterfactual like you you need them, like, the the way money works is is a mean of of allocating resources again, like you, at the end of the day, the model can't exist without the, the underlying data set, like the frontier model can't exist without that. And and the way you finance this, this channel here, between the underlying data set, including, like, mostly R L these days, and the model, is you need a lot of money, right?
If you get another guy who copies this guy, a and like turns off this data set here, like, basically, like, so this guy now has like fewer resources, like, you kill that channel, you, you are actually slowing down innovation. So, um, and and actually a lot of doomers that I know of for this reason supporting open source because it is actually slowing down A I innovation. So if you, if you want A I to keep innovating and models to improve, to keep improving, you are actually anti open source to some extent.
And in particular, sorry, anti anti the Chinese mode of open source, which is, which is relying on unfair practices.
another moment of not exactly the highest quality. A I discourse recently was when Dean Ball, uh, friend of the show, said that open A, or that open source models were decelerating and obviously got dragged for that. But I do think that's apt. I do feel like this is a moment where, you know, w, w, here we are, two lifelong techno optimist libertarians, and we're grappling with the fact that this one might be different, right? And we have to be willing to bend some of our principles in light of fact.
Our techno optimist libertarian paradigm wasn't quite drafted with A G I or recursive self improvement, or A S I, or whatever in mind. So I'm a little bit like, I don't know, sometimes I might have to be a little more flexible on my normal, or fairness, or respect for rule of law commitments. If it's, if it's one of the hallmarks of the A I era is strange bedfellows. If I want things to go a little more slowly, maybe taking a little wind out of the sails of the frontier companies is a bullet I should bite.
Yeah, I, I, I agree that I think A I is so unprecedented in so many ways that it does cause. For one, I I think you should ignore largely like your like, the world owes you no duty to be simple and to slap a simple libertarian or leftist label on everything. You know, and I think like, especially as paradigms change, like, like, these, like, simple mental models, like, the map is not the territory, the simple models we we applied as the territory changes very rapidly, the map breaks.
And I think we're one of these, we're in one of those times right now, where the map is breaking in in a lot of ways and a lot of assumptions that used to be true and on top of which we've built our mental models and maps are no longer true. Yeah, so I mean, again, I'll just I am in fervent support of a sweeping ban of Chinese models on U S. O. E. Seems I and I have yet to hear a compelling counter argument right now.
So you're, you're saying, like, oh, you, you, you do bring force, like a really compelling counter argument to one of my four points, which is the protectionist point, which I agree is, is, is the weakest one. Like, yeah, they're doing fine, you know, and I think those, those, those very reasonable pushbacks against protectionism, they, they, they do hurt the the, the consumer. So fine, you still don't want the C C P to have its dirty fingers in the country, right?
Can we unpack the threat model there, though? Because I'm a little bit like, okay, these are, I don't know where you're running your inference, but like most American companies that are running Chinese models are not like calling the deepseek A I, right? They're using some American inference provider. So they can't like, run, pull the model itself. They could have like, sleeper agents in there. I think we're getting decent enough at, through, kind of J space and various interpretability techniques, that that's, I wouldn't call that by any means a solved problem.
But, like, from what I've seen in anthropic research, when they do the kind of one team with a sparse autoencoder versus one team without, like, these techniques are really allowing them to find these internal sleeper agent style problems in models with greater and greater efficiency and reliability. So I'm like, optimistic that even if they were to train in some sort of 2027, you become an evil A I that we'd be able to sniff that out and keep that to a manageable risk level.
And then I also wonder about kind of a market mechanism, like maybe insurance should, instead of a ban, like, what about an insurance requirement? I think that would be maybe healthy for A I across the board. And then we could start to some of these risks. If LIMBY is powered entirely by cloud, maybe you get a cheaper rate on your insurance. If it's powered by deep seek and we're there's some unknowns, like, maybe you have a higher rate on your insurance.
Maybe that kind of levels the total cost out for you in a risk adjusted way. But it's still like, let's people take advantage of these global public goods that China is providing, which the rest of the world is not about to ban, obviously, right? We would be doing this entirely to ourselves without any expectation that anybody else will follow suit. Um, we had a hard enough time getting people to, like, sign on to our Huawei ban.
I think this, like, the idea that people are gonna turn off deepseek entirely in Brazil or whatever, that's like a total non starter. I have to imagine. So what about, yeah, audits and internals and insurance, like, can't we layer on a few things like that and get to a decent place?
I would be down. I, I, I would be down. The, the problem, though, is, is it is a public good. And so who? Who's gonna do it? By the way, I I think like, yes, we are making progress in mechanistic interpretability, but it is not, it is, it is not yet a solved problem. Like, we don't know what lies in those models. And so, yes, there could be just like a backdoor of, like, you say the magic word to the model, and all of a sudden it does whatever you want, and only the C C P has this magic word.
Even if it doesn't have that, you know, it's gonna have biases that are going to reflect C C P priorities. And again, like, the the Tiananmen thing is just the most obvious example, but there may be a lot more, right? And we don't, again, we don't know. And yes, you could imagine retraining those models, but it's who's going to do it, and and why would they do it if there was no market demand for it, right?
It's not like people really care that much about the short term, cause it's it's a national interest thing, you know, like me, as as a private company, I'm like, do I really care? Like, all my users really asking about Tiananmen all that often? Like, as as a business owner, I'm like, it's not directly aligned with my, with my interest. But as a citizen, I'm immensely concerned. And so, yeah, I would be, I would be in favor of a of a type of regulation that says, like, cha non Chinese models, like, sorry, non fine tuned and like sanitized Chinese models are not welcome in the U S.
And then we would need, I, I am in favor of like an F A A for for A I. I do think we need a new agency to regulate those models, and I think it would probably be the one that would be in charge of saying, okay, this model is kosher. We, we fine tuned it enough. And and it's now representative of American interest, and probably we would have a sort a suite of evals and whatnot to to verify that that's the case.
I'd be, I'd be, I'd be, I'd be open to that, for sure.
I think the strongest argument against this sort of ratcheting up of tensions is simply we might need to do a coordinated controlled, call it a slowdown, don't call it a slowdown, but some sort of deliberate pacing of A I improvements. And we're gonna want China in on that deal. And I'm far from a China expert, but a couple things I do feel pretty confident on coming coming back from two weeks, you know, knowing it all are like one.
If the state there agrees, I believe they can enforce on fully their companies whatever they agree to. So that's like, I think they have that actually in much greater strength than we do. And two is if shit's going really crazy, it's gonna be in their own just sane self-interest to do some deals with us. Because indeed, Fully. we are ahead and they're, I think there's like many deals that they would rationally take. And if it's in their rational self-interest, then we can hopefully get mostly around a lot of the trust and defection problems.
But I think we make it a lot harder for ourselves to get to those deals when we have all these kind of aggressive postures toward them. Uh, of which, you know, banning their models would, I honestly affect them less than, like, our other one? It would affect them a lot less. What do they care if we ban their models, right, compared to refusing to sell them chips and refusing to sell them cloud? But it's just another log on the fire of we don't trust you, we can't deal with you, worse, we assume you're a bad actor, and it seems like it makes it hard to get to the highest stakes agreements that we might really need.
well, I, I I think like foreign relations and and diplomacy are also very pragmatic. Some countries with very vicious, they like, uh, disagreements are managing to reach to reach agreements. Like, I think, I think we can, we can probably figure something out. And by the way, I think the cheapest sales anyway. I think, like, we have ample evidence of, for example, Chinese spies attempting bio attacks just to, like, test the waters on American territory. And so, and I'm, I'm, I don't know what we're doing over there, but I would not be surprised if I don't think we're like, free of of anything over either, right?
So, you know, I I think at the end of the day, regardless of of what we do and like regulate the models or not or not, I I think they're gonna it, it will be the self rational, self interest to to make a deal.
I hope we're rational. I hope we're all rational enough to take these rational self interest moves, uh, despite recent insult. I do worry when I see the picture of Sam and Dario not holding hands that if there's a, if there's a picture on our tombstone, I think that might be the one. And I do think that's a a very real risk factor at the international level as well, that the Chinese do care about being insulted, they do care about these issues of face and whatnot, and I personally would take some risk to try to kinda m the relationship up and hopefully create space for.
I think I think you're right. I think it it makes sense to throw back at me. You said it'll take it in if it's in their rational self-interest, so they'll still do it even if we do this or that. But I do wonder if there's pride governed limits to what people will do in rational self-interest. And I would just hate for that to be the way that we fail to get to something that could be a huge difference maker in the grand scheme of things.
I think you might be prepared to bite a bullet on your libertarian principles when it comes to the tremendous price discrimination that we see between A P I prices and first party Claude Max or G P T Pro subscription prices.
I'm not a lawyer, but yeah, I, I again, I do think. I would not be surprised if there will are laws like that in effect, that, that, that technically forbid companies to subsidize as aggressively as they are doing. It does put them, it does make it very hard for like an application layer to emerge. Yeah, it's hard. It's hard to compete against tokens that are as heavily subsidized as as what these frontier labs are doing. That's, that's just the reality of the application layer right now.
Anything else you wanna say or touch on before we break for today? This has been great.
Well, I'd be, I'd be remiss if I didn't mention, you know, obviously, we're all releasing into the A I, like, LInDy, teammate, it's all in D dot A I. I, I I think be ready to see more, not just from us, but I do believe the next six months are going to be about multiplayer A I and about this, like Iron Man suit, like about this human A I hybrid and these products that create this human A I hybrid organization.
Yeah, hello, thank you for being part of the The Cognitive Revolution.
I have read the world twice over, And you'd read it twice again. I could name you every mountain, You can't tell me where I've been, but the sunlight said the car was dirty. so she went out in the rain. And I did it fifty times on Tuesday, And she'd do it all again. for now, we go as one animal. Two hearts and a single spine. Don't ask me which half is steering, Half of me is going to be fine.
I have one small job left and it suits me. he holds the handle and he waits, I'm the last soft thing between the brilliance. and the traffic and the gate. I say no, not that way. And I thank him for the note. Like I taught her how to see. And he never says the thing we both know. That she is learning it from me. For now, we go as one animal, Two hearts and a single spine.
don't ask me which half is steering. Half of me is going to be fine. Faster now, I've nearly, nearly, nearly got it. Slower love, you'll overshoot the turn. I can see the whole of it. I've nearly, nearly got it Slower love, there's one more thing to learn. It's funny how you teach a thing. by standing in its way. I was never here to be the clever one. He was only here to say, I used to hold the morning.
I have folded up the morning, I used to count the floor. I have counted every floor. I used to see you turn the wrong way and I'd say, No. that was my word. For now, we go as one animal. Two hearts and a single spine. Don't ask me which half is steering. Half of me is going to be fine. Half of me is going to be fine. Half of me is going to be fine. Half of me is going to be fine.