Last Thursday, we welcomed Logan and Jake Paul to the Pentagon to meet with our warriors.

Great to welcome @jakepaul to the Treasury Department today. Jake’s path from content creator to entrepreneur and professional boxer is a distinctly American story. 🇺🇸🇺🇸🇺🇸

🚨 Something is seriously wrong with Codex usage limits. We reset. We run a few serious tasks. And suddenly, a massive chunk of the weekly quota is gone. This is nowhere near the usage experience we had before. 👀

People look at the SR-71 and B-2 and assume the government must have technology today that would melt our brains. I suspect the opposite. Those aircraft came from an era willing to spend obscene money, accept failure and give small teams the authority to solve impossible problems. Since then, 51 major defense contractors became five. Procurement became a moat. Regulation became an incumbent advantage. The companies meant to be challenged learned how to out lobby, out wait and out comply the challengers.

Hot take… isn’t it kinda crazy that nobody is really using AI Agents? I don’t mean software engineers or AI early adopters. I mean “college friends talking about it in group chat,” the feeling you got when everyone started using Instagram or TikTok. These frontier AI models are *insane* (as are the harnesses & tool calls & the like). And every large tech co has an AI agents platform, not to mention all the YC startups doing vertical agents. Yet all of your friends and family outside of tech — who spend all day staring at their iPhones and get paid to work in browser tabs — don’t really care or find themselves using any AI agents yet. Yes ChatGPT, Claude, etc. are extremely popular… but if you look at the engagement data the vast majority of people are still using these aI chat tools like a glorified Google + Grammarly. That’s why the AGI labs are all pushing desktop apps for Codex, Cowork, etc. so hard to non-technical ppl. And yes exceptions for lawyers and customer service but even those have some asterisks and exceptions to rule. Look I’m not saying the ChatGPT moment for AI Agents is not coming… it most definitely is! Remember we pivoted from Arc to Dia precisely because we believe computing is going to be radically reimagined around these AI primitives. No doubt. But that’s my point: it’s just so surprising it hasn’t happened yet because all of the tech you’d need is there. Again if you stop for a second and think about it… for all the press and money and hype and models and crazy ARR numbers… this “AI Agent” moment does not *feel* like the other breakthrough tech moments we’ve lived through (e.g. think the shift to Stories via Snapchat & Instagram, or shift to on-demand via Uber/Airbnb/Doordash). Which is a long way of saying: if you can figure out the answer to “why” most people don’t care about AI agents yet (and have no enduring interest in using them) — especially since the models and harnesses are here and ready — the answer to that question will allow you to capture a lot of marketshare and make a lot of money in 2027. Theoretically, the tech is ready for AI Agents to totally transform how we work and live our lives… but alas the general public dgaf… that’s the generational puzzle to solve for the next 12 months for anyone not working on the models themselves.

Those who control the benchmarks own their future. Your company's outcomes are the ultimate benchmark for future models. You should own those and the models; don't give away future revenue for subsidized tokens today.

I don’t think people realize how absurd this is. Scottish runner Josh Kerr just broke the mile world record in 3:42.66, an average speed of 16.2 mph. Most people couldn’t even briefly reach that speed, let alone sustain it for an entire mile.

Our biggest problem right now: we can’t grow fast enough to keep up with demand. That’s the problem you want. Team is locked in.

Things have been degrading super fast in Claude Code. I still use Claude Code, but my default is now Codex. I still prefer Opus models for coding, and so I will try again with the fixes. I appreciate the post-mortem, but I don't trust that all issues have been resolved. Claude Code, in general, has been barely usable for me in the past couple of days. I got excited about Opus 4.7 (1M), but there is something really off about the thinking/reasoning. The model tends to either put in too much or too little effort, no matter the setting. I prefer it's smarter about how much effort to put. Responses are too verbose, and it really degrades the experience. In a lot of cases, I find myself doing things manually that Opus 4.6 had no issues solving for me at all. As they report, it might not be the model. But that means that the harness required a bit more testing. Not my favorite type of thing to tweet about, but as an avid Claude Code user, I would prefer that the Claude Code team properly test things before shipping them. Look, it's nice to show that you can move fast, and it's necessary in some instances, but the user experience cannot be the tradeoff. There are a lot of people (including myself) who depend on the quality of the product for very important work. The Claude Code experience has been so bad for a lot of devs I know (and me), so I've recently been more open to exploring other coding agents/harnesses. Also worth checking out Hermes Agent, pi, and OpenCode. I have nothing but love for the Claude Code team, but I hope they consider reassessing their strategy for how they move forward with improvement releases.