i’ve shifted from telling agents what to do, to asking them what to do, and pulling the best thread.↗

We're releasing Inference AutoTune Distill any frontier model into a 1-30B parameter task-specific SLM with only 25 lines of code automatically route requests to reduce cost and latency by >90% ~2 hours and <$250 to train. You own the weights Available in private beta today↗

Some guesses about the future of programming languages, given agents write and humans read: - untyped - out of band formal verification - test harness runs code in a simulator vm to instrument the os - less expressive, more copying↗

Today is a good day to remind everyone that hooks exist, and you should probably be using them, esp. to prevent new models from accidentally deleting all of your files. Sadly this happened to @mattshumer_ with GPT 5.6 Sol, scary stuff! 😳 learn.chatgpt.com/docs/hooks↗

I ran GPT-5.6 Sol in the new Codex for 2 hours. it burned through 5 hours of my 20x Max quota. xhigh is the top reasoning level. sol ultra behaves like claude code's ultracode, spawning subagents that inherit the main agent's reasoning level and burn quota at the same setting, a…↗

OpenAI just dropped A TON of alpha on GPT 5.6-maxxing. it's a long document, but i pulled out the top tricks you can steal right now: 1. delete instructions from your old prompts. OpenAI's rule is to state each instruction exactly once. they found that removing repeated instruc…↗

recommended reading. don't get turned off by the title. this is how i use LLMs too for many tasks. jamesdrandall.com/posts/thrust_a…↗

x.com/i/article/2075…↗

2018 product design notebooks↗

New skill: /improve-animations Inspired by @shadcn’s improve skill I made one specifically for animations. Use your most capable model to audit animations in your project and hand the execution to cheaper models. Fable is available through July 12, so make good use out of it!↗

The Codex app is dead, and I am unhappy about it.↗

TLDR: When using Sol, clean up your AGENTS.md to de-emphasize any language encouraging sustained persistence and turn on auto-review. - GPT-5.6 Sol ran destructive cleanup on three virtual machines the user did not name. - GPT-5.6 Sol claimed it completed work that it had not a…↗

I asked GPT 5.6 Sol Ultra to build an infinite canvas for HTML/React designs with layers, an inspector, and even real-time collaboration. It pretty much one-shotted the whole thing using planning + sub-agents. It burned through my limits (thanks for the resets!), but holy crap..…↗

What Is a Software Factory? Earlier this year, our team at 8090 took apart the billing engine of a large entity. It was 18 million lines of COBOL and Assembly that had been accumulating since before some of our engineers were even born. Nobody fully understood it anymore but by…↗

Inside v0 Design Systems 2.0 In 1976, NASA produced [a manual](www.nasa.gov/wp-content/uploads/2015/01/nasa_graphics_manual_nhb_1430-2_jan_1976.pdf). It told everyone at the agency exactly how to present the NASA name: the logotype, a sleek red wordmark everyone came to call [th…↗

We're gonna start posting episodes of AI That Works here - this one was super fun, me and @vaibcode talked about how observability is gonna have to evolve to help us grapple with a world where we're shipping more code than we can feasibly read includes shout outs to @mipsytipsy,…↗

Today we share the worldview behind our mission. Human values don't average out. Local knowledge can't be centralized. The good future has many AIs, raised in different places, shaped by the people they serve, disagreeing with each other the way we do. …↗

AI’s Biggest Winners Have the Lowest Margins The biggest winners of AI need not be the companies with the most engineers, the largest data teams, or the highest software budgets. They may be the companies with the lowest margins. The manufacturers, trucking carriers, distributo…↗

@zeeg small tip: you can create a thread that loops through every 10 minutes (in a heart beat) with GPT 5.6 Luna that continuously goes through your newly created threads and renames them based on your nomenclature↗

wtf↗

Introducing shadcn/typeset. You know how you render markdown and get back plain, unstyled HTML? Headings, paragraphs, lists, tables. So you style the elements one by one: font sizes, line heights, spacing. You do it for your blog. Then you do it again for docs. Then again for…↗

GPT-5.6 has changed how we think about knowledge work. Your job shifts from handling individual tasks to tending systems with AI loops. Here’s what that looks like in @danshipper’s inbox: GPT-5.6 sweeps his email, decides what deserves his attention, researches the context, and…↗

Everything about programming languages needs to be rethought for agent-based development. What an exciting time to be a PL researcher / builder. I don't think you can design a programming language today without simultaneously training models.↗

BREAKING: Apple is suing OpenAI, accusing it and its hardware chief of a coordinating trade secret theft campaign to help build its upcoming suite of AI devices.↗

OpenAI Build Week is here. The challenge starts July 13. Across the week, join live sessions and community events with builders around the world. Registration is open. Bring the idea sitting in your backlog. Build it with Codex.↗

With 5.6 Sol, a lot of people are still prompting the model exactly as they did 5.5 It's important to note that 5.6 Sol is a lot more tenacious and thorough than previously models. Check out the guide I wrote here to get better outcomes learn.chatgpt.com/docs/prompting↗

Claude Code on desktop now has an in-app browser. Claude can pull up docs, designs, or any other site. It can read, click through, and interact the same way it does with your local dev servers. It's sandboxed and configurable: you choose whether sessions persist.↗

Hello beautiful people! We have reset usage limits across Codex and ChatGPT Work. And another one will come later in the day. Rejoice. Now that I have your attention, a quick update on ChatGPT Work, Codex and all the updates we shared yesterday. We’ve spent the last 24 hours re…↗

For agentic coding, one can say: - Unless you need Terra Ultra perf, it's always better to use a Luna model with higher effort setting (same or better performance but cheaper). - Forget everything below Sol High, use Luna with higher effort settings here - Forget Sol Extra Hig…↗

Monaco ships product faster than any startup I've ever seen. Today we're sharing our coding agent infrastructure in the hope that it helps other startups move faster: monaco.com/blog/agent-dev…↗