Kurt Schrader

What’s this @Korey_AI thing?

@karpathy I've met very, very few non-technical people that have used OpenClaw.

Someone recently suggested to me that the reason OpenClaw moment was so big is because it's the first time a large group of non-technical people (who otherwise only knew AI as synonymous with ChatGPT as a website) experienced the latest agentic models.

Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.

@can Impossible apparently.

someone should enlist coding agents to build an oauth that doesnt suck and then migrate the web to it

This is ... not fast? There is like a 300ms delay on every click there. I cry for our profession.

@mylordcod Real lack of creative ideas honestly.

@kurt but not many people are trying crazy things mostly just settling on chat and sidecars

Everyone just keeps saying that we don’t know what the optimal UI for any of this is yet.

Been in Korea for 24 hours and I haven’t seen a single ad for an AI company yet. NGMI.

Like a breath of fresh air in a world of slop.

Great article about Steve Jobs learning about quality processes and how "brute force and great people" aren't enough. I rarely talk to people who've studied this in our business, but I talk to a lot of people who think that they're the first one to discover small parts of it. ft.com/content/080062…

Watching my kid play Wordle and wondering why there hasn’t been a single viral vibe-coded app yet.

@forssto Maximum friction reduction.

@kurt Yeah with some solid central enablement these teams can absolutely FLY.

The optimal dev team in 2026 is two people pairing with an agent in a sandbox, with full ownership of the feature.

@forssto Same. We’ve pivoted our agent Korey to “how can we help these small accelerate as much as possible” and the (very early) returns on some of the experiments we’re running seem super promising.

@kurt As a person who has militantly defended the classic squad model, I'm 100% ready to flip to this camp. Ambidextrous builders with rocket fuel in their tanks are killing it right now.

@Testimonychrid 🤷‍♂️

@Kurt is our problem not the coaches

The fastest way to speed up your app right now is to point an AI coding agent at the slow parts (and then do it over and over again). They're better at performance optimization than the engineer who wrote the code. Almost nobody I talk to is doing this.

@ivanhzhao Bravo! A few humans collaborating with an AI is how you end up with the best work.

The loudest story about AI is a lonely one. One person with an army of chatbots. Other humans are friction. That gets the future wrong. The best things aren’t built alone. In a moment of change, we want to remind the world (and ourselves) what Notion stands for: — Think Together

Who you gonna call?

I’ve been playing around with some of these TurboQuant optimized models today on my MacBook and, whoa boy, what you can do locally now is fast approaching frontier level. Planning to write code offline using local models on my entire flight to Korea this week.

And 1068 on Twitter/X

I’d forgotten about this from back in the day. GitHub user #2323 here (the double Jordan)