Shared my first trace from @NanoClaw_AI to @huggingface yesterday. Very cool! By default, all agents should store their traces on HF (in private) so that you can keep a history of them, analyze them,… & share them and post-train better models, harnesses and more. Excited
SOFTWARE
-
Fastest P-Video-Replace model for character replacement, 70% off until Sunday
By
–
P-Video-Replace from @PrunaAI is up on Replicate!
— Replicate (@replicate) 4 juin 2026
This is the fastest model for character replacement in existing video.
And until Sunday, we're giving our community 70% off, making this model only $0.009/s of output video.https://t.co/0vZhSlksom https://t.co/tWLiRDuBSiP-Video-Replace from @PrunaAI is up on Replicate! This is the fastest model for character replacement in existing video. And until Sunday, we're giving our community 70% off, making this model only $0.009/s of output video. http://
replicate.com/prunaai/p-vide
o-replace
… -
Google drops Gemma 4 12B with novel multimodal architecture
By
–
Google just dropped Gemma 4 12B! You can now run it locally on just 8GB RAM using Dynamic GGUF from Unsloth. The architecture is different from any multimodal model before it. No separate vision encoder, no audio encoder. Both flow directly into the LLM backbone. Vision is
-
Pipeline order is a hyperparameter for optimizing LLM execution strategies
By
–
5/5 Takeaway: pipeline order is a hyperparameter. If you're already paying for parallel rollouts, reuse them – they're relevant context, not just candidate answers. Full write-up: [
https://
ai21.com/blog/first-sca
le-then-enrich-how-the-right-execution-strategy-helped-us-reach-state-of-the-art-on-swe-rebench/?utm_source=org-twitter
…] -
AI21 Labs surpasses Claude Code in efficiency and performance with Test Agent.
By
–
4/5 Still came in ~$0.30 under Claude Code’s spend at a similar score. So we added a lightweight Test Agent that writes repo tests and filters failing patches, pushing our final result to 60.9% – surpassing Claude Code (60.9% vs 56.2%) at the same cost.
-
AI21 Labs: ReAct Agent Performance with Enrichment and Scaling Strategies
By
–
2/5 Started with a baseline: classic ReAct agent (GPT-5.2), single Docker-terminal tool. Baselines on the slice: vanilla 53.8%, enrich-only 55.6%, scale-only (n=5 + LLM judge) 55.4%, enrich-then-scale 57.7%.
-
MiniMax M3: Open-weights frontier model challenges closed model dominance
By
–
THE ERA OF RELYING EXCLUSIVELY ON THE 3 MAJOR CLOSED MODELS IS OVER@MiniMax_AI's M3 is officially out 💥💥💥
— Charly Wargnier (@DataChaz) 4 juin 2026
It delivers the exact same capabilities you expect from a frontier model, combining massive leaps forward in a highly cost-efficient, open-weights package.
Here's why… pic.twitter.com/NDUppZzMlqTHE ERA OF RELYING EXCLUSIVELY ON THE 3 MAJOR CLOSED MODELS IS OVER @MiniMax_AI
's M3 is officially out It delivers the exact same capabilities you expect from a frontier model, combining massive leaps forward in a highly cost-efficient, open-weights package. Here's why -
AI21 Labs: Reversing agent pipeline order achieves SOTA results
By
–
1/5 Our latest Labs in Front piece: Agent pipeline order matters. By reversing a common agent recipe – scale first, enrich second – we reached SOTA on a Dec ‘25 to Mar ‘26 slice (123 issues): 60.9%.
-
Claude’s neurosymbolic code useful, but more AI work needed
By
–
now/years. claude code is neurosymbolic and pretty useful in its domain, but there’s lots more to be done (see my 2020 article Next Decade in AI).
-
Launching SynthTraces: generating synthetic coding agent traces
By
–
Today I'm launching a new project called SynthTraces It is a minimal codebase to generate synthetic coding agent session traces using Pi (from @badlogicgames
) I wanted a large number of coding-agent traces, so I built a tiny harness where two models talk to each other: – an