Krentsel: OpenClaw, Pi, and Claude Code hardcode static agent policies
“These are all policy decisions that are static, that are defined for OpenClaw, or for Pi, or for Clawed code, if you look at their Source code. And so that is the kind of, that is the policy of what an agent is, the tools it can use, the skills it has, how it …”
Krentsel: Improving AI models should architect their own agent harnesses over humans
“We believe All of that needs to be improvable by the agent, especially as the agents keep getting better, because as they keep getting better, it's this better lesson. You don't want to over-specialize because you don't want the human kind of deciding all thes…”
Krentsel: Exo autonomously modified its code to inspect Pokémon game RAM
“We've had XO running, playing, playing Pokemon. And while it's running, the system itself decided to try inspecting the like RAM of the game and then went and mapped the RAM to, and people have reversed in the past, people have reverse engineered this manually…”
Krentsel: Harness architecture must enforce agent rules instead of model alignment
“My bet is this though, that people keep trying to get models to do things that align with their goals and alignment is an unsolved problem. We keep trying to like RLHF, like try to align these models to do the right thing. In harness space, we actually have an…”
Krentsel: Decoupling executor from harness enables safe agent self-evolution
“What this gives you this nice split is it gives you an isolated execution environment down here. It gives you protected state, and then it has a very explicit stateless layer that's safe for self evolution. So the executor can propose changes to itself in a wa…”
Krentsel: Exo Automatically Rolls Back Runtime Executor Rebuilds That Fail
“And we have in the exo harness, a special garden guardian kind of process that will allow the executive to be rebuilt at runtime mid step. And also provide some sort of kind of soaking process where After being rebuilt, we try to bring up the executor, let it …”
Krentsel: Agent teleportation is only needed when scaling parallel tasks
“In a world where you're running a single agent, working on a single task, there is like very little need for teleporting your stuff away. Just, you either run it locally while you're working, or by the way, if you want it to just be always on, you have to run …”
Krentsel: OpenClaw agent threads cannot be interrupted during active execution
“Once an agent kind of goes and starts working on a thing in a thread, that thread is not interruptible. As the thing is off working, if you try to ping it just won't respond. You don't even know what it's doing, which is really frustrating.”
Krentsel: The Exo Harness Is Running in Production at Braintrust
“The EXO harness and agents built over top of it is running in production at Braintrust.”
Krentsel: A16z GP Martin Casado writes code daily for the Exo project
“Martine really codes. Martine's really in this. Scoting day to day. So it's been super fun, the three of us working together. If you look at the commits, we're all there.”
Krentsel: Self-improving AI requires explicit objective evaluators to avoid misalignment
“In all these optimization problems, the final difficulty is
You always have to provide some evaluator and that evaluator has to kind of reflect your goals, because otherwise there's no signal for the thing to hill climb or to optimize, or that signal, it might…”
Krentsel: Exo autonomously re-architected its Discord adapter, cutting costs by 96%
“We asked it, Hey, I noticed, I asked, Hey, but how much did the last message cost in the discord adapter?
And it was like, it was.
It's like, are you serious?
16 cents.
That's actually crazy.
Like.
Go work on driving that down.
And so it went and re-architecte…”
Sewitz: Crickets consume far less feed and water for equal protein
“Crickets consume far less feed, far less water, and produce equal amounts of protein, but the hiccup, obviously, is that they're crickets, and we have no history or tradition of eating insects in America”
Lewis: Novel foods mainstream when hidden in familiar formats like California rolls
“There's a long history of foods that at one point were viewed as bizarre in the U S you could look at sushi, for example, which 50 years ago, eating raw fish was a little bit strange to most Americans, and the way it became popularized was that a chef in LA cr…”
Lewis: Kickstarter excels at generating press and concept validation over capital
“Yeah, in hindsight, Kickstarter probably wasn't the easiest way to raise money, but it was an amazing way to get attention and get press and prove the concept.”
Sewitz: Novel food products succeed on taste and utility, not intellectual interest
“You also had to really convince them that it was going to taste good and also meet the nutritional needs they were looking for from a protein bar, because ultimately the product was going to live or die on whether people kept buying it for the use case that we…”
Lewis: Tim Ferriss, Robb Wolf, and Mark Sisson invested in Exo
“We were able to leverage what at the time was almost all of the names in the paleo world, whether it's John Duran, Mark Sisson, Rob Wolf, Tim Ferriss, all of those people who came on board as advisors or investors in the business.”
Lewis: Exo did roughly $1M in revenue in its first year
“The first million dollars or so, which was roughly what we did in the first year after launch, was faster and easier than anybody would have predicted for a company selling cricket bars.”
Lewis: Selling Exo felt like a failure because his identity was attached
“It definitely felt like a failure for a while. And I think not only in the sense that we hadn't achieved the mission we set out to achieve, but we also self-identified as the cricket guys. We'd go to Expo West and trade shows, and we'd have our t-shirts that s…”
Lewis: Magic Spoon outperformed Exo primarily due to product-market fit
“The truth is, we're doing some things differently, but we're doing a lot of the same things, and it's working so much infinitely better now, partly because we've learned from some mistakes from XO, and we spent five years banging our heads against the wall, bu…”
Ismail: Exponential organizations perform 10x better, faster, and cheaper than peers
“So the definition we have for an EXO or exponential organization, as we call it, is that it's delivering 10 X better, faster, cheaper than its non-EXO peers.”