Daniel Auras @rasdani_

one of the craziest bets of prime intellect was swinging for everything from the start. we operate datacenters, we build our own trainers and envs and sandboxes, we have a platform for large enterprises and solo devs, we train huge models, we publish. ~40 people btw. hiring, DM

43 39 1K 85K 352

View Details

Daniel Auras @rasdani_

3 days ago

@willccbb

1 2 19 704 1

View Details

will brown @willccbb

3 days ago

something has definitely shifted in the past few weeks. seeing a huge uptick in large enterprises wanting to secure compute and post-train their own models in house, frequently on top of GLM-5.2. everyone is starting to understand how open source wins.

95 193 3K 255K 631

View Details

Daniel Auras @rasdani_

3 days ago

@DhruvAhuja2003 facts

0 0 0 298 0

View Details

Daniel Auras @rasdani_

3 days ago

let me correct the typo in the first sentence "Misanthropic executed the largest known distillation attack on humanity to date."

Sam @Discoplomacy

3 days ago

Anthropic’s letter accusing Alibaba of distillation.

68 51 326 544K 246

1 2 37 2K 5

View Details

Daniel Auras @rasdani_

3 days ago

Claude @claudeai

5 days ago

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

2K 2K 28K 19.8M 12K

0 0 3 312 0

View Details

Daniel Auras @rasdani_

4 days ago

@xeophon whack-a-mole mentioned 👀

0 0 1 59 0

View Details

Johannes Hagemann @johannes_hage

5 days ago

large-scale agentic RL is one of the most exciting systems challenges today building an efficient stack means optimizing many components, both individually and as one co-designed system: training, inference, request routing, weight broadcasts, in-flight weight updates, environment execution, sandboxes and more. we're hiring engineers who want to push these systems at massive scale: primeintellect.ai/careers

Prime Intellect @PrimeIntellect

5 days ago

prime-rl is fully open source, and we're hiring systems engineers to take it further. Read the full prime-rl performance deep dive: primeintellect.ai/blog/rl-at-1t-…

2 7 105 18K 49

5 10 160 12K 57

View Details

Daniel Auras @rasdani_

5 days ago

@PrimeIntellect 🐐🐐🐐

0 0 6 397 0

View Details

Matej Sirovatka @m_sirovatka

5 days ago

They distilled me into a blogpost today 🫡 was a lot of fun optimizing the whole stack over the last few months, still a long way to go but the whole team cooked on this release 🚀

Prime Intellect @PrimeIntellect

5 days ago

Today we're releasing prime-rl v0.6.0 — enabling RL at trillion-parameter MoE scale on agentic workloads at the highest efficiency. We've relentlessly optimized our RL infra. The result: GLM-5 on agentic SWE tasks at 131k context and sub-5-minute step time.

37 95 949 288K 430

6 10 122 7K 25

View Details

samsja @samsja19

5 days ago

prime-rl can now train 1T parameters MoE blazingly fast, under 5 minutes per step, or 1k steps in ~3 days To achieve this we shipped in our latest prime-rl 0.6.0: * inference: wide-ep, fp8 inference, llm-d router, mooncake, kv cache cpu offloading * training: fsdp2, deep-ep expert parallelism, dsa cp, fp8 training, router replay * agentic rollout: we rewrote the core of our rollout orchestrator for better scalability support for glm5, kimi, nemotron, ..., prime-rl is open source but also end to end optimized to run on our dedicated RL infra and compute layer

Prime Intellect @PrimeIntellect

5 days ago

37 95 949 288K 430

18 37 417 85K 179

View Details

Prime Intellect @PrimeIntellect

5 days ago

37 95 949 288K 430

View Details

Daniel Auras @rasdani_

a week ago

> Open weights are now our default. We’ll keep building toward the frontier and releasing increasingly capable models in the open. more open weights poolside doing the right thing in critical times love it!

Poolside @poolsideai

a week ago

Today we’re releasing the weights for Laguna M.1, our most capable model to date, with a 256K context length. Both base and post-trained checkpoints are now available on Hugging Face under Apache 2.0.