Harshith

had just shipped a repo for training speculative decoding heads to speed up inference of llms by ~3x. get any base model, train a few speculative heads, see the difference in throughput. 🧵on more details. 1/n

@Shloak1712 it's hitting the inflection soon.

Oh my my, Robotics and Physical AI companies are raising left and right and becoming Unicrons at Series A. I love it!! Lets solve hardware, data and deployments at light speed!!! I can’t anymore catchup with the raises, like 4 raises announcement today itself i guess.

@asharoraa @PrimeIntellect congrats ash!!

Some news: A little over a year ago, I moved from London to San Francisco. 🇺🇸 In Oct, I joined @PrimeIntellect. Nine months later, we have gone from 0 to $100M ARR, and raised our $130M series A. I’ve now lived and worked across 5 countries as an adult, and nowhere comes close to the pace, ambition and optimism of the US. Every day, you’re surrounded by people trying to build something that changes the world. The US is truly the greatest country in the world 🇺🇸 After years in VC, I joined a seed startup called Prime Intellect as Head of Applied GTM to help shape our product strategy, commercialization, revenue and applied AI offering across post-training and RL. My friends know I spent most of 2024 building an RL thesis, which inevitably led me to Prime Intellect. Leaving VC was hard. There are few jobs more meaningful than being trusted to back someone else’s life’s work. But at some point, I realized I didn’t just want to be close to the arena. I wanted to be in it. This felt like a once-in-a-generation opportunity to build. Today we are 40+ FTE, $100M+ ARR and infinite passion to do more for our mission! It has been the most intense, humbling and rewarding chapter of my career. Every day, I get to work alongside an extraordinary team building the infra for the next generation of AI from post training and RL to the full-stack that ambitious teams need to build, improve and own their intelligence. Our mission is simple: make frontier AI infra and open superintelligence accessible to every ambitious team, so people can own their intelligence. We are still at day one. I genuinely believe the next decade of AI will be built very differently from the last. Open source will dominate. Post-training will become the way companies make AI actually work for them. And the next era will belong to teams that own their own agents, models, data and intelligence. Moving up the stack is how you win. We’re hiring across all teams. If you want to work on one of the hardest and most important problems in AI, come build with us.

@nileshrthr @heyclicky t a l k t o m o r e c u s t o m e r s

guy selling post trained oss models: if you don’t own your models you don’t own anything guy selling harness on top of models: if you rely on one model you are cooked guy selling tokens: instead of manually prompting your models just put them in a loop where they prompt themselves

@adarshxs lesgo!

Bangalore's AI infra lore. Soon.

@alokbishoyi97 yo yo

creating a high signal whatsapp community of super cracked indian / desi researchers and engineers , purely centred around post training comment or DM to get added

@_sarthak4 im the guy in the yellow hood, opus is the pipe guy lmao

me to my codebase every few days

@itzzme_pk @xai hey let's chat! shoot a dm

hi @xai connect me with folks who are in the domain of ai inference :)

every infra piece you need to know to do RL on GLM-5 primeintellect.ai/blog/rl-at-1t-…

@miten Miten we agreed on 10x.

@kunalb11 congrats @kunalb11, finally Meta's AI will be a smooth part of life.

It’s been a minute. 2015–2018 - Exited FreeCharge. Spent time learning and investing. - Pondered about: Why can't trust be rewarded? Started with $1M of personal capital. - Launched CRED to reward people for paying credit card bills on time. 2019–2025 - Built a system run by a team that values ownership, judgment, and craft. - Grew from 0 to 17M members by aligning incentives with behaviour. - Built several products during COVID lockdowns. - Raised $900M+ from global investors. Did 4 ESOP buybacks. - Made Indiranagar and IPL ads slightly more interesting. - Received a full stack of regulatory licences. - Lost 35 kilos. - Scaled from 0 to ~$325M ( ~₹3,200 crore) in annual revenue across payments, lending, insurance, commerce, wealth, and credit cards. 2026 - First profitable quarter (yet occasionally asked what our business model is) - Raised another $900M from Meta in primary and secondary capital. - Announcing our 5th ESOP buyback. Today CRED is ready for its next phase. I am stepping back and @miten steps in as interim CEO, partnered with an incredibly talented team. He has been heading strategy and finance and suffering me since 2020. I’m stepping away from the operating role and will continue as a shareholder. My commitment doesn’t change. Just the role. Extremely grateful to our members, partners, regulators, and investors who made this possible. And to our board, Shailendra, Micky, Saurabh for their extraordinary conviction. Team CRED, I’ll still expect you to be a 10x version of yourselves. As for me, I’ll be joining Meta to lead WhatsApp globally. Meta comes in as a minority investor in CRED. No access to member data. While it’s come very far, the delta between WhatsApp today and its full potential is massive. I look forward to working with Mark, Chris, and the leadership across Meta for the next step in WhatsApp’s journey. Will, thank you for scaling something the world relies on quietly, and for making this transition smooth. Onwards.

@tensorqt this is such a good reminder of the next really helpful tools will never look like the previous gen. i think one thing @tensorqt said that really stuck is, we all had flywheel running in mind, but none took it to a good abstraction. nice stuff.

your research doesn't have to be ugly

insannnee. this is good. stateful dag + sandboxes with right target persona should unlock a lot of things. ml research is a step closer to being solved.

Speculation Is All You Need. In this blog post, we announce the co-release (w/ Z Lab) of six more state-of-the-art DFlash speculators for @Alibaba_Qwen 3.x. Over 1k output tps for 3.5 122B-A10B on a B200. Read the blog for why we're all-in on spec dec. modal.com/blog/spec-is-a…

Speculative decoding is the closest you can get to a free lunch method. It is beautiful and astounding. I am surprised that it does not play a greater role in "AI".

@Mantthanjain hey manthan, had to dm you regarding an interesting opportunity - can you open dms? or shot a dm for me.

Am i the only one who use to call it Pepsi stick back in childhood?

@raghavmundara05 @jaixbhatia v v v interesting

Most companies fail. People rarely do. That one line is why I've spent the last 3 months building Preflop, a new financial instrument that lets you invest in a person's trajectory. A person raises capital by issuing tokens against their own future earnings. They pay a share of their income for a set period, then convert to a small permanent stake that trades continuously. Once the income share phase ends, the payments stop and the tokens reprice as the individuals’s career unfolds. Issuers raise without debt or dilution. And backers get in before the world catches up, converting early conviction into shared success. We've built markets for companies, real estate, royalties, and revenue. Never for the thing that generates all of them. The math holds (all 350 pages of it). The pricing engine, raise mechanism, and a dozen other specs are already live. After 947 LOIs and 86 meetings across 5 cities, I realized 30 minute conversations don’t do justice to the depth here. So I'm building the rest in public, and iterating with your feedback. Live at app.preflop.co

Sandboxes are all the rage (Modal, E2B, AWS, ..). Most AI teams pay a >4x markup to run sandboxes on someone else's machines. Introducing SkyPilot Sandboxes — Run BYOC sandboxes on your own clusters. • 50,000+ sandboxes on a single cluster • Sub-second launches with warm pools • Great for RL rollout (keep sandbox clusters close to GPUs) Benchmark shows @skypilot_org Sandboxes are 4-10x cheaper than Modal at lower latency. Full results in blog.

🚀 We're excited to be a Day-0 launch partner for NVIDIA Nemotron 3 Ultra. Deploy NVIDIA's latest open model for agentic AI on Simplismart Our optimizations we deliver higher throughput than TensorRT + MTP + NVFP4. Read more: lnkd.in/g7BrrNhy #NVIDIA #Nemotron

@kts_o7 @akshat_bhatia02 @NishantKus38420 gg!!

Earned our first contract at #Artifylabs Worked with a stealth startup from the US to build their MVP for angel raising. @akshat_bhatia02 , @NishantKus38420 and me build things for anyone who needs quality software. Check us out at linkedin.com/company/artify…