Sabitlenmiş Tweet
Arnav
381 posts
when will you become an artist, the kind that writes manifestos, paints propaganda?
English
Late-interaction retrieval models are widely used for their strong performance, but their representations can be utilized beyond just retrieval. Our new paper demonstrates that these representations can effectively replace raw document text in RAG tasks.
English
excited to keep cooking on retrieval 👀
Will Bryk@WilliamBryk
Exa just made our first acquisition: Browser Buddy @trillarnie and @jeremyjsuh were working on new types of retrieval to organize the web, so that anyone could find the best blogs and essays. Now they’re joining Exa to do the same, but for all the world’s information. They’re two of the smartest and kindest engineers I’ve met, and it’s already been an insane few days working with them. Here’s to organizing the web together 🍻
English
Arnav retweetledi
If you got confirmation that you got into the program reply and I’ll reach out to match :)
Aakash Gupta@aakashgupta
Anthropic just mass-licensed 10,000 of the most influential developers in the world for $12 per person per month. The math: 10,000 maintainers × $200/month × 6 months = $12M in sticker price. Actual compute cost to serve these accounts runs closer to $30-50/month each, meaning the real spend lands around $3-5M total. Those 10,000 people maintain an open source ecosystem valued at $8.8 trillion in demand-side impact, according to Harvard and the Linux Foundation. 60% of them are currently unpaid. This tells you everything about how Anthropic views developer distribution. They’re acquiring the people who decide what tools get baked into every CLAUDE.md file, every CI/CD pipeline, every GitHub Action, and every project README across the most-used repositories on the planet. And they’re doing it for the cost of a Series A marketing budget. Think about what happens when a maintainer of a 5,000+ star repo starts using Claude Code daily. They write CLAUDE.md files. They add Claude Code GitHub Actions. They reference Claude in contributor docs. They build workflows that assume Claude as infrastructure. Every contributor to that project encounters Claude as the default. 10,000 maintainers each influence, conservatively, 50-100 downstream developers through their projects. That’s 500K to 1M developers seeing Claude Code embedded in their daily workflow within six months. GitHub spent years and hundreds of millions building Copilot awareness through traditional developer marketing. Anthropic is spending $3-5M in compute to get Claude Code embedded at the infrastructure layer of open source itself. The timing is surgical. Microsoft killed Azure Sponsored Subscriptions for open source maintainers in September 2025. Burnout rates among maintainers hit 44%. Quit rates hover at 60%. Anthropic walks in with the most expensive AI subscription on the market, handed out free, right as everyone else retreats. The selection criteria reveal the strategy. 5,000+ GitHub stars or 1M+ monthly NPM downloads. These are the people whose tooling decisions cascade through dependency trees touching every Fortune 500 codebase. $3-5M in compute for 500K-1M developers organically adopting your tool through the open source dependency graph. That’s $3-10 per developer acquired, embedded at the infrastructure layer where switching costs compound monthly.
English
you can slow time down quite easily. the meme that your life is half-over by the time you are 22 or whatever is a psyop invented to make you compete harder along tech’s defined clout-lines. approach your life with focused passion and urgency, write often, take a sharp knife to your thoughts, and record everything, and you will find that there are many seconds in every minute
English
@mixedbreadai is the scoring some variant of xtr/warp or the full set to set comparison with maxsim?
English
We build the first production ready multi-vector and multimodal search.
Now we are serving over 1 billion documents in under 50ms latency (p50).
We are sharing how we build it.
English
English
1/ Today @stripepress publishes one of my favorite books to date: Maintenance: Of Everything, Part One by @stewartbrand, about the unglamorous yet civilizationally important work of maintenance and repair.
press.stripe.com/maintenance-pa…
Stripe Press@stripepress
Out now: press.stripe.com/maintenance-pa… @stewartbrand's Maintenance: Of Everything, Part One—about the continuous repair work that keeps complex systems intact, spanning stories from a round-the-world sailboat race, to the restoration of the Statue of Liberty, to the Model T’s rise.
English
It’s a shame because not only do humans naturally think in this way, it’s also a very strong literary device that has been used for centuries and is now being amplified by AI.
I distinctly remember learning this structure early on in school but forget what it’s called.
François Chollet@fchollet
Do the people using the formula "X isn't just Y, it's Z" every other sentence realize that everybody else can see where this is coming from
English
@finbarrtimbers imagine if all research with open source python code used uv
would be so wonderful and easy for reproducing experiments/papers
English
@nabeelqu it would be super convenient/cool to prompt my for-you page
not sure how they trained grok for recommendations, maybe something like eugeneyan.com/writing/semant… could work?
last I checked semantic IDs are difficult for live corpuses, especially one being updated as fast as twitter
English