chirag (@chirag_plusplus) header
chirag (@chirag_plusplus) profile picture

eng @blaxelai (now part of @baseten) prev: cofounder @humanbehaviorai (yc x25), uchicago math

San Francisco, CA
Joined February 2024
Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)Photo by chirag (@chirag_plusplus)
5 months ago, I didn’t know what an MoE layer was, now, I kinda do. Anyways, here’s how @stefanopopoulos and I made the fastest delta weight syncs in the world.
3
2
1
62
6,708
Today, @JudgmentLabs opens for general access. We believe long running agent systems will generate more exhaust data than the world can currently process. Hidden within agent traces are rich records of how to make these systems better: what users ask for, when agents stumble, and how agents interact with the external world. Judgment is the most performant monitoring and improvement layer for long-horizon agents. We’re excited to see what you build!
243
259
125
1,084
1,250,239
Huge congrats to the Base Labs team!
Today we’re releasing two decision models, elephant-1 (9B) and mammoth-1 (27B), trained by the @baselabs team, available for free for one week on our Model APIs. mammoth-1 is #1 on the Jev Decision Index and elephant-1 is tied for #1 on the Artificial Analysis Decision Index. It’s a new model, but more importantly, this is an open-source model you can access, deploy, and post-train on Baseten (or wherever you do your best work). The free week ends 10/27/26 at 11:59 PM PT. Try it and let us know what you think. Try mammoth-1 here: baseten.co/library/mammoth-1… Try elephant-1 here: baseten.co/library/elephant-…
9
349
Our first research summit, All-Gather, is on November 18 at the Exploratorium in SF. Apply and see you there!
Progress that moves the frontier is almost never made alone. We’re thrilled to announce All-Gather, our first-ever research conference that will bring together the brightest minds in research, founders, and engineers to join us for a day of debate and exchange. Come hear Noam Brown, Sarah Guo, Horace He, Bryan Catanzaro, Jerry Tworek, Sabri Eyuboglu, Tuhin Srivastava, and many more debate pre-training, post-training, infrastructure, model performance, and emerging fields. Join us on November 18th in San Francisco. Research has been part of our DNA from our early days of optimizing open-source models through our recent launch of Base Labs. We’ve grown with the open ecosystem, and we’re excited to keep building alongside it. Space is limited, but we’d love you to join us. Apply here: baseten.co/all-gather/
1
2
55
2,503
why does claude stall at every hour and half hour, @anthropic?
3
199
chirag retweeted
New on Blaxel: sandbox archiving 🧊 When you archive a sandbox, only the filesystem is preserved (not the memory), in a colder storage tier. Built for sessions that seldom need to be restored.
1
1
1
8
477
Really cool companies in he new batch!
Good luck F26! iced coffee was on us today. the caliber is honestly insane, these companies look so much further along at day one than our batch was just a few that stood out to me: @Simulithic, Plek AI, @IncTerse, Podesto…
5
604
Me: Can you explain this massive context dump more simply? Claude: Running 5 new commands ????
5
211
big news: @blaxelAI has joined @baseten!! I am more convinced than ever that the agents that’ll make it out on top in this market are the ones that 1) run fast 2) don't break. Joining Baseten is how we enable this for all builders, by colocating inference and agent infra. If you’re already using Blaxel: nothing changes. Blaxel continues and features will keep shipping. More in post 👇
13
4
2
64
5,969
chirag retweeted
Your coding agent should keep working when your terminal disconnects. Blaxel’s official Herdr plugin runs Codex, Claude Code, or OpenCode in a persistent Sandbox. Reconnect later, open the preview, and bring back a checked Git patch. Try it: github.com/blaxel-ai/herdr-b…
4
12
1,041
Update: I left @humanbehaviorai to join Blaxel (@BlaxelAI). At Human Behavior, I realized that long horizon agents will need a new class of cloud for agents to use. I feel I have to be on that frontier. I'm grateful for and bullish on @OfficialAmogh and @skyguy_skyguy, and HB!
11
1
37
6,624
Introducing Real-time RL. In the real world, time isn't free. The environment keeps "moving" even when you're computing your next action. We show how RL agents can learn to adaptively think in real-time games. 1/🧵
17
84
17
719
78,323
chirag retweeted
there is a rhetoric in ai rn that vibing and half-assing is the future of technology. do not fall for this psyop. the future is deep understanding and mastery. always has been
94
865
35
7,576
140,580
What’s with sf parties? Everyones a DJ and nobody’s dancing
3
1
10
1,577
Founder boy and venture girl
2
1
1
33
21,848
She called me dawg is it over chat
13
42
4,097
The GOATS. My bosses sent me cupcakes from across the country for my birthday 🥹 @skyguy_skyguy @OfficialAmogh @chirag_plusplus @humanbehaviorai
4
5
34
23,138
how long until claude starts asking for equity
15
1,078