Given some of the results I'm seeing recently, it's pretty clear Codex is a good harness. But it will seem primitive in 2-3 months and we're about to go through another major evolution in how we use AI at the frontier. The next generation of models need more than your laptop.↗

I've been testing GPT-5.4 for the last week. In short, it is the best model in the world, by far. It's so good that it's the first model that makes the “which model should I use?” conversation feel almost over. The biggest surprise: I barely use Pro anymore! If you know me, y…↗

The harnesses are still not optimally utilizing the models Everyone rightfully loves claude code right now, but it is clear to me it is underutilizing the capability of opus. The lowest hanging fruit is /init in existing code bases. It does a good job of building it's …↗

🚨 700+ MODELS on @FactoryAI DROID?! Totally possible. Just finished this VS Code extension for DROID. It’s wild: ⚙️ 700+ models ready to import 🌐 Choose your own provider 🧠 Droid Flow - build model chains (ex: write → GROK, read → GLM 4.5) 💻 Local server for instant and uni…↗

Has opencode ever shown reasoning tokens? I can't for the life of me get it to show any reasoning tokens for any model I try. Same prompts generate reasoning tokens in other clients.↗

We surveyed hundreds of engineers building in AI about everything from which models they’re using to whether they’re using a dedicated vector database. And, of course, if they think everyone will have AI girlfriends by 2030. Some highlights 🧵:↗

i couldn't leave nyc without a parting gift so we're hosting a hackathon with a prize of $25,000 @OpenAI credits. 📅 june 5th, 10am-7pm 📍 soho, nyc if you are a pre-seed builder, this is the chance to build with the latest models & tools and demo for a live audience of VCs.…↗