How we run a code factory at Toyo We now write code in a way that is completely different to how we worked six months ago. Now building is cheap. This warrants a rethink of engineer and product development processes. All the planning we used to do—backlog planning, design—was…↗

Developer to Fleet Commander The top 5% to 10% of builders are now 3 to 5 times more productive than they were a year ago. The median builder is up maybe 10% to 20%. --- In the past few weeks, I've spoken with several founders about how much AI coding tools are creating enormo…↗

The Execution Layer LLMs are in desperate need of an execution layer made for them to run tool calls in. A year ago LLMs were making direct calls to tools, we found that it flooded their context with irrelevant information to them and found incredibly poor performance. Then we…↗

Karpathy's Autonomous AI Can Make You 701x Faster. It's The Future of Business. Most marketing teams run ~30 experiments a year. The next generation will run 36,500+. Easily. They'll run experiments while they sleep. Current marketing teams run 20-30 experiments a year. Maybe…↗

Can Alpha School Succeed Where Success Academy Stalled? I’m sitting a few feet from one of the nicest billionaires I’ve ever met. He wants to change how every kid in America learns. Joe Liemandt is wearing a black ALPHA baseball cap. Expensive yet understated polo. He’s smiling…↗

x.com/i/article/2026…↗

Lessons from Building Claude Code: Prompt Caching Is Everything It is often said in engineering that "Cache Rules Everything Around Me", and the same rule holds for agents. Long running agentic products like Claude Code are made feasible by prompt caching which allows us to reu…↗

Skill Graphs > SKILL.md people underestimate the power of structured knowledge. it enables entirely new kinds of applications right now people write skills that capture one aspect of something. a skill for summarizing, a skill for code review and so on. (often) one file with on…↗

I gave OpenClaw a name, its own computer, and told it to figure things out. Here's the 32-hour recap I didn't give an AI agent access to my computer. I didn't give it access to my email, GitHub, browser sessions, or files. I gave it its own computer, accounts, and workspace. A…↗

The vibe coding trap There's a popular narrative on X right now that AI can build software so quickly and cheaply that SaaS is dead (or will be soon). Why pay for Linear when AI can build a project tracker in an afternoon? Why pay Stripe $30k/year when you can vibe code your ow…↗

x.com/i/article/2020…↗

x.com/i/article/2019…↗

x.com/i/article/2019…↗

x.com/i/article/2018…↗

x.com/i/article/2017…↗

x.com/i/article/2015…↗

x.com/i/article/2013…↗

x.com/i/article/2011…↗

Who is Cowork for, exactly? I've been playing with [Cowork](claude.com/blog/cowork-research-preview) today, and I have thoughts. # First: what is Cowork? Pitched as "Claude Code" for non-technical tasks or knowledge work. [Embedded Tweet: x.com/i/status/2010805682434666759]…↗

Comparing Codex CLI vs Claude Code side-by-side It's been almost a month [since I last tested the Codex CLI vs Claude Code](x.com/iannuttall/status/1953833034794651649) (a lifetime for AI models and tools) I've been afk over the summer and a lot of chatter is about how much Cod…↗