Seth Karten
@sethkarten
Prime Agent | Continual Harness, Agent Bazaar | Research @PrimeIntellect | PhD @Princeton | NSF GRFP
623 Following 4.2K Followers
Forget P(doom). What is your P(success)?
What is the probability AI and humanity succeed beyond your wildest dreams?
Over two weeks, Prime Agent orchestrated a swarm of over 2,000 agents to rewrite itself in Rust.
With 10,000+ sandboxes, 200B+ GLM-5.3 tokens, and 16,000 agent-to-agent messages, this was our largest test yet of Prime Agent’s multi-agent capabilities and our infrastructure powering these agent swarms.
Prime Agent now reaches usable input ~13× faster and uses 83% less startup memory than before.
Show more
We rewrote Prime Agent in Rust with 5x faster cold start, 8x less memory, 2x streaming throughput, and full behavioral parity.
Prime Agent is a harness designed from the ground up for long-running agents. The RLM architecture allows many agents to coordinate across a session that outlives any single context window, and is what enabled 2,000+ agents to rewrite the harness they were running on.
Show more
prime agent rewrote itself in rust , all powered by glm5.3 and our own inference, spawning thousands of subagent that all communicate with each other
🦋🦋
This has been insanely fun to work on for the past couple weeks!
This rewrite tested the limits of all our products: thousands of subagents via Prime Agent, billions of tokens via Prime Inference, and hundreds of sandboxes via Prime Sandboxes!
It's incredibly rewarding to see our full stack come together, and Prime Agent is now faster than ever :)
Show more
Agent swarms turn software work into dependency graphs and state machines.
We limit-tested this by having Prime Agent automate its own Rust rewrite.
Now enjoy one of the fastest CLI agents :)
Over two weeks, Prime Agent orchestrated a swarm of over 2,000 agents to rewrite itself in Rust.
With 10,000+ sandboxes, 200B+ GLM-5.3 tokens, and 16,000 agent-to-agent messages, this was our largest test yet of Prime Agent’s multi-agent capabilities and our infrastructure powering these agent swarms.
Prime Agent now reaches usable input ~13× faster and uses 83% less startup memory than before.
Show more
That’s a wrap on our poster presentation at @COLM_conf 🎉
Great experience presenting our work on Automatic generation of high-performance RL envs. Truly grateful to have worked with @sethkarten!
Lots of interesting conversations, thoughtful questions, and exciting ideas!
Show more
Continual Harness mentioned in the State of AI report 2026 👀
welcome to the 9th annual @stateofai report!
ai building better ai, world models, physical ai, geopolitics, cyber defense and 12-month predictions.
read it at stateof(dot)ai or watch my editor’s cut here.
a few of this year’s biggest findings follow:
Show more
Scaling inference is one of the largest burgeoning research areas.
more computation. more memory. more tokens.
new blog post: why taking inference scaling seriously leads to swarms.
contrary to popular belief, I don't think multi-agent systems are a new scaling axis. they're just structured inference scaling.
Show more
Attending COLM 2026?
Come say hi to talk about self-improving agents and economic multi-agent alignment
Looking forward to chatting with old and new friends!
While durable harnesses have been kinda rage for the past few days, I've been experiencing durability since I switched to Prime Agent as my daily driver a couple of months ago. This fantastic talk lays out the thought process and the design behind it.
Show more
Looking forward to giving a talk on Self-Improving Agents via Prime Agent and the Continual Harness tomorrow at this COLM 2026 Social at 1PM
hosting the Self-Improving Agents social at @COLM_conf tomorrow
- 4 talks: @sethkarten @yoonholeee @qinyuan_ye and Noah Provenzano
- panel: @HuaxiuYaoML @ZhiruoW @jasonwu0731
my 4th one after ICLR, ICML and NeurIPS, and I still love hosting these.
1-2:30pm, Franciscan D. come by if you're at COLM
Show more
Excited to be at COLM 2026 next week! If you’re around, definitely come say hi. Lots of interesting conversations to be had around self-improving agents, and multi-agent systems
Attending COLM 2026?
Come say hi to talk about self-improving agents and economic multi-agent alignment
Looking forward to chatting with old and new friends!
Come join us after COLM on Wednesday to see what Prime Intellect's researchers are working on
we're organizing a small event at our office next wednesday during COLM, if your loss curve looks like this, join us :)
Introducing Prime Inference
Own your intelligence
Own your inference
Win
Immediately available to use prime inference in prime agent!
Introducing Prime Inference:
We've served trillions of tokens for RL and dedicated customer deployments
To own your intelligence, you need to own your inference
Unpacking our inference stack
Show more
@kaimcpheeters @sethkarten step one: have all the harnesses compete
step two: see prime agent come out on top
step three: profit.
prime agent v0.9.5
we fixed a lot of bugs and, of course, we had prime agent feature its favorite updates
it picked our perf work. then it created the video itself.
Prime Agent reached 20k stars on GitHub!
Thanks for being such an amazing community 🦋
