Roast your API Spec using LLMs. Take your Open API spec and ask a language model to roast it. You can get some tangible feedback on how to improve your design, e.g. based on RESTful principles. #LLM #API

Spent about an hour this morning merging 4 PRs of QoL updates via Claude Code Web, only spent a few bucks of the $250 allowance. Not sure I'm able to spent it all before they run out :)

I have this neat @claudeai command that have Claude Code inspect @Cloudflare Workflows using the Wrangler CLI, look for errors and other issues, examine the code and provide a report with findings and recommended fixes. Saves an awful lot of time browser the dashboard.

What is a good RSS reader that also makes it easy to discover "indie bloggers" in the tech, space, energy fields? I miss the glory days of Google Reader.

immensely proud of the team for our best model yet. grateful to be able to work with such a strong team of researchers who are always curious and willing to explore the untrodden path

Millions of views on me not being able to get an EU GPU If this was about Google @OfficialLoganK would be in the replies promising a fix If it was about xAI @elonmusk would see it himself probably Meanwhile the only thing the @EU_Commission has posted is to be cautious of fake news! The EU doesn't even monitor their mentions on here and they definitely don't reply nor does the EU's president @vonderleyen Which is funny because the biggest companies in AI here do, they reply on here to hear feedback and improve their trillion dollar companies It's just crazy to me how disconnected the EU is from European citizens now: there's no dialogue, it's just a one way street of imposing rules and regulation and burning billions of € of European people's money with little checks or balances It's very frustrating but Europe is our homeland so we should keep fighting for it

OpenAI eating lunch from a dozen startups and veterans.

Why is it always dinner reservations and flights? No, thank you very much!

Nice one @claudeai ⏺ The type inference issue is strange. Let me try a pragmatic fix by adding a proper type guard with Array.isArray() which TypeScript will recognize:

The choice of AI code model is similar to that of search engine; once you find something that feels right, any attempt with a new model feels off, you don't trust it as much. Of course, we have many more options with code models than we've ever had with search engines.

just use grep

We’re thrilled to announce our $23M Series A led by @Sequoia - joined by @patrickc, @collision, and @david_helgason. Honored to welcome @roelofbotha to our board. We’ve entered a new era of manufacturing. Reusable rockets. Humanoid robots. Self-driving cars. Humanity is reinventing how it builds. We started by building hybrid rocket engines. Along the way, we built tools that made us move faster - design, build, test, iterate. Those tools became @buildonfloweng. Today, Flow has become the default platform for next-gen hardware teams, powering the systems that will define our future.

The @github notification system has become buggy. Unread notifications, but none to be seen when filtering the inbox. Increase of SPAM makes it a lot worse.

Experimenting with node-based canvas for generating images.

Cooking some really nice code reviews and tracking reports in markdown

Dette er indiskutabelt!

Who is the Xerox PARC of AI?

Act now fightchatcontrol.eu/#contact…

Referring to code agents as "junior engineers" is my go-to

Loving the new Agent Window in @cursor_ai

It has been a con job from day one. Germany’s opposing position was only to assure false hope and keep citizens on the back of their feet.