Jed

The faster people can get to something 80% done and unstable, the more they are going to crave stuff that's already 100% done and reliable.

Introducing Flue — The First Agent Harness Framework Flue is a TypeScript framework for building the next generation of agents, designed around a built-in agent harness. Flue is like Claude Code, but 100% headless and programmable. There's no baked in assumption like requiring a human operator to function. No TUI. No GUI. Just TypeScript. But using Flue feels like using Claude Code. The agents you build act autonomously to solve problems and complete tasks. They require very little code to run. Most of the "logic" lives in Markdown: skills and context and AGENTS.md. Flue is like Astro or Next.js for agents (not surprising, given my background 🙃). It's not another AI SDK. It's a proper runtime-agnostic framework. Write once, build, and deploy your agents anywhere (Node.js, Cloudflare, GitHub Actions, GitLab CI/CD, etc). We originally built Flue to power AI workflows inside of the Astro GitHub repo. But then @_bgiori got his hands on it, and we realized that every agent needs a framework like Flue, not just us. Check it out! It's early, but I'm curious to hear what people think. Are agents ready for their library -> framework moment?

This is the the quote I've been citing a lot recently.

Time spent with children spikes for a few years in your thirties and early forties before falling off a cliff. It’s a devastatingly short window. Blink and you’ll miss it. Slow down. Play Monopoly. Read them books on repeat. The time you spend with your children is never wasted.

It's wild to think about how massive 1M token context windows in LLMs really are That's roughly equivalent to: - The complete works of Shakespeare - 11 hours of audio - A 5-minute session fixing some TypeScript issue

Unpopular opinion: "AI makes everyone a developer" is true the same way "cameras makes everyone a photographer"

I heard rumors that Claude Mythos can fetch data in React without useEffect

To quote from my keynote at Vercel's internal offsite: Software is free as in puppies. It will pee in your bedroom and eat your furniture. The weight of every line of code is real. We will need to maintain it. We will need to port it. It goes into the context window. And somebody in this room will get paged at 2am because it did something unexpected

everyone is trying to build async agents that work when they sleep but all they really need are Australians

I’ve been following Ghost pretty much since it launched, and could not be happier for them about the $10M ARR milestone 👏 The coolest thing is they cracked this while empowering their users to build sustainable revenue on top of their platform. So much to love here.

Despite the doom and gloom you hear on X the developer job market is fine. In fact dev jobs grew 10% over the last year while the overall market declined. Here's what I'm seeing on the ground as someone who recently went through interview loops with several companies: - There are a ton of early to medium stage startups hiring very aggressively. - AI-native people are in very high demand - If you're interviewing, just aim for earlier stage companies - If you're in SF or NYC you're probably going to have a much easier time - Larger / legacy companies are the only ones doing layoffs, and it's only a few of them. My advice: stay ahead of the curve on AI tooling and be able to demonstrate what you know pointedly.

If you thought your company's edge was "how fast you ship", you're in for a rude awakening. Everyone can ship fast now. Obviously, not everyone can ship tastefully, with quality and restraint in mind. That's the new edge.

The speed Anthropic are shipping what everyone wants *in* Claude is the best ad for how fast you can ship things *with* Claude

This is insane, but it also highlights how well spec'd and documented next.js is. Specs are the new source code.

@wesbos probably the most insane thing I've read this week

cloudflare ported Next.js to vite in a week for $1,100 I don't believe it, but excited to try. It's a drop-in replacement, let me know how it works on your app.

Meanwhile, I can totally see why if you have the money to burn, paying more for faster inference is worth it. I found myself avoiding smaller tasks with Gemini because it's just. so. slow. (to be fair I'm paying Anthropic way more than I'm paying Google for this so ymmv)

Thinking about the trade-off between model speed and skill today. I threw the same (design / front-end) task at Gemini 3.1 and Opus 4.6 today, same source context and prompt. Gemini did it better, but Opus did it faster, which meant tighter iterations and a better result.

One of the things I really enjoy about my OpenClaw setup is smoothly moving from desktop to phone, it’s always there. This is a big upgrade for Claude Code.

It’s only February… this year is going to be wild

The world of software leaps forward right now _by the weekend_