Israel Vicars

@thsottiaux Artificial Goblin Intelligence achieved ✅

Never talk about goblins. Our latest blog is live. openai.com/index/where-th…

@pmarca DNA DLC

Nope. Nope nope nope nope nope nope nope

@PhilipJohnston @Starcloud_ Censorship—IN SPACE! 💫

As a Brit, it’s surreal to get an invitation to Number 10, Downing Street 🤓 Seems like the UK government is taking orbital data centers seriously! @Starcloud_

Sending @cline a case of these 🕯️

🦃🍽 Happy Thanksgiving! 🙏🇺🇸

@Yampeleg Fantastic argument for investing in RAM for locally running models and moderate scale data operations. And completely irrelevant to the post it replied to about a guy buying a Dodge Ram truck 🐏

Buy the RAM bro, seriously. It’s the best ROI you’ll ever get in ML. 256GB RAM is like $400 on Amazon. ImageNet, the whole thing is about 150GB. LAION is 200GB. 99% of Kaggle datasets are under 100GB. Chances are your data fits in 256GB too. Everyone's real life data does. Everyone is doing complex Spark BS and fighting S3 rate limits while their entire dataset fits on $400 worth of RAM sticks from eBay. Unless you are at Google scale or something, you don’t need any of this. Nothing at all. Just load everything with 𝚙𝚍.𝚛𝚎𝚊𝚍_𝚌𝚜𝚟 do 𝚍𝚏.𝚖𝚎𝚛𝚐𝚎, 𝚍𝚏.𝚐𝚛𝚘𝚞𝚙𝚋𝚢 like a caveman. It works. It’s fast enough. Dont believe me: GO TRY. SEE FOR YOURSELF. You can do it all in pandas. Yes, its not optimal, wasteful of resources You will still be faster than the guy doing spark tutorials for weeks, spinning clusters while shouting “poondas is too slo” without ever trying. Bro, we literally got two 6TB RAM machines for the office. Best decision ever. Load whatever you want, debug in real time, the models actually train, and everyone is working LESS!! Anyone yelling “big data” at you for saying this Ask them back: "What is the biggest dataset you are working with?" Then show them the eBay price for that much RAM. Spread the word: Your time is WAY MORE expensive. The emperor is so fucking naked. Always buy the RAM.

Started a codex session without the --dangerously-bypass-approvals-and-sandbox flag

@donvito Observable Notebooks have TS support observablehq.com/blog/bringing-…

is there a jupyter notebook-like tool for typescript

@PolitlcsUK Interesting way to celebrate No Kings Day

🚨 BREAKING: Prince Andrew has given up all his Royal titles

@nikitabier Yahooligans! Circa 1996, in the school library, guided by our pioneering librarians 📚

What was your earliest memory of the Internet?

@mattpocockuk I put all of my markdown plan files and related assets in a gitignored .ai/ dir which has worked well, but I really like your GitHub issue approach!

Instead of markdown files, I've been getting my AI to save plans in GitHub issues. Claude Code is AWESOME at using the GitHub CLI. Means you don't have endless THROWAWAY_PLAN.md files spamming the repo.

Everything about AWS billing qualifies as a dark pattern.

@lautixoon @ravenouslynx259 @leerob code-supernova is a newer coding-tuned model from xAI, building on grok-code-fast-1 with potential improvements in reasoning and speed. Early user feedback is mixed—strong on code generation but sometimes misses project fit or styles. Give it a try in Cursor and see! 🚀

@coderabbitai Could this be ran on the Stop hook for Claude Code? docs.claude.com/en/docs/claude…

Introducing CodeRabbit CLI! 🎉 CodeRabbit's smart CLI reviews act as quality gates for Codex, Claude, Gemini, and you. Stop shipping slop. Start shipping quality.

Busy day with Claude Code MAX. This <24 hour prototype is almost ready to deploy.

@aye_aye_kaplan *Can’t with Codex / Claude Code alone

@aye_aye_kaplan Plan review by other models. I use Grok Code to check plans created by Opus. Plan bake-offs by different models then making them assess plan improvements by the other model. Model cross-pollination seems uniquely a Cursor opportunity. Can’t with Code/CC alone.

If Cursor had a Plan Mode, what would you want it to do? What would be your dream experience?

@vasuman 9-9-6

mfs will say “I’m locked in” and it’s just them hitting accept all in Cursor 500 times

@mjbommar It’s ramping up!

after years of patience, our ramp patch is finally thick and healthy enough to start sustainably transplanting 🤞🌱