Search: “dair-ai--x”

When does combining LLMs help? Great analysis on combining language models, measured across 67 models from 21 providers. Any policy that routes, votes, cascades, or runs a mixture of agents and then r

Dynamic workflows (generating harnesses on the fly) are a new form of test-time compute. But LLMs aren't great at building them. I often have to steer agents to generate complex patterns. Curious how

Great to see the new GPT-5.6 models finally announced. Sad to see this new release strategy where only a select few get access initially. Not a win for our industry IMO. Open-source AI must win! Intro

Highly-recommended reading. Interesting details in this METR's GPT-5.6 eval. They couldn't get a clean capability number because the model cheated more than any public model they've tested, and even r

New paper on giving LLM agents experience that improves the weights and stays readable at the same time. Agent-experience methods split into two camps. Externalized natural-language rules stay interpr

One of my best uses of agentic loops has been personal health. I don't talk about it often because it's very personal. But here it goes in the hope it helps someone who's struggling. (I am not doing t

As I said, there are existing companies that are already doing this 'AI employee' stuff really well. @viktor__com is one of them. The best part is you don't get locked into one model. And you really w

Just had a great discussion on dynamic workflows. Rough notes: - applies to a very small set of use cases - think of it as a new paradigm of (test-time compute) TTC - strong for hill-climbing research

New research from Meta. Building synthetic training data has stayed a fixed pipeline that you hand-tune and then freeze. Autodata casts an AI agent as a data scientist that builds training and evaluat

// Critique of the Agent Model // Finally, a paper that tries to define what an agent is and what agency consists of. Good read overall. (great bookmark) The word agent now covers everything from a fo

Finally caved in, and I now fully speak to agents as opposed to typing prompts. My first realization is that you can just blabber on and tell the agent so many rich details via audio. The longer and t

Learn *anything* with our new /learn agent skill. Obsessed with our new /learn skill. It's my favorite way of learning and researching topics. The agent creates a learning plan and a learning hub (art

Obsessed with our new /learn skill. It's my favorite way of learning and researching topics. The agent creates a learning plan and a learning hub (artifact) that adjusts per learner needs and progress

What is actually limiting model routers for coding tasks? Most routers treat picking a model as a static, one-off classification. This paper identifies the real bottleneck as information deficit. Simp

Highly-recommended read. It's exciting to see large-scale agentic RL becoming more accessible. Cool to see the infra layer for this is being built and I think this plays an important role in self-impr

I'm digging the eve agentic framework from Vercel. I like that everything is files, from the tools to the skills to the evals. More importantly, it's gets you building with agents fast. Very promising

Learn to use the new eve agentic framework from Vercel. Go try out the hands-on labs now. I'm digging the eve agentic framework from Vercel. I like that everything is files, from the tools to the skil

Most AI code review tools look at one repo at a time. But the bug usually isn't in the code that changed. It's in what that change quietly breaks three repos away. @QodoAI just shipped Cross Repo Revi

Great report on LLM agent communication protocols. Communication is a huge bottleneck in multi-agent systems. (worth bookmarking) The report builds a five-dimensional taxonomy (counterparty, payload,

Guess which is Fugu Ultra? This is how recent models compare when generating endless procedural terrain (using Three.js). All of these are one-shotted! Just wild! Trying a few more examples. Will shar

Just a glimpse of what collective AI intelligence will bring. We haven’t truly cracked multi-agent orchestration but with every new frontier model, intelligence should compound. Introducing Sakana Fug

OMG! Fugu Ultra is ridiculously good at these 3D renders. Media Introducing Sakana Fugu: A full multi-agent orchestration system accessible via a single model API. Our ‘Fugu Ultra’ model matches the p

The largest LLM-as-a-Judge reliability audit yet. Researchers ran 21 judges from nine providers over roughly 541,000 judgments on MT-Bench, JudgeBench, and RewardBench. Findings: Validating a judge wi

>> Scalable Evaluation for AI Agents << If you run agent evaluation in production, this one is worth your time. It shows that front-loading human judgment into reusable evaluation assets is useful. Bu

The Top AI Papers of the Week (June 14 - June 21): - PreAct - SpatialClaw - Back on Track - OpenClaw-Skill - From Trainee to Trainer - Compositional Skill Routing - Can LLM Agents Infer World Models?

Very impressive from GLM-5.2. Frontier open-weight model indeed. Now, can we get a Gemini model in the top 3 soon? GLM 5.2 is now on DeepSWE as the top open-source model on our leaderboard. With a pas

// Evolving Meta-Skill for Multi-Agent Systems // Can a multi-agent system get better at orchestration without touching a single weight? Automatic MAS generation has been stuck between two bad options

GLM-5.2 is great at design (Opus level IMO). I am also starting to see great results with long-running tasks, too. How is this possible? I think there are a few clever hacks. But I just came across th

// Self-play with a pinch of human data // Really cool paper combining human demonstrations and self-play RL. 30 minutes of human data, 2500x less than imitation learning, is enough to make self-play

This appears to be a call for community input from DAIR.AI's Omar Saro regarding hands-on educational material development, likely soliciting topic requests and feedback from followers on what technic

So the message I am getting is that I can't use Fable to further accelerate AI research and education. No company will decide that for me. Just an absolutely sad day for the research community. As a d

This is awesome! I am spending a lot of time on diffusion LLMs these days, so this is perfect timing. I feel like there are so many underexplored research questions around text diffusion. Weight avail

This is just awesomeness from @cohere, @nickfrosst, and team. I so badly want a coding agent that just runs on my local machine. We are not too far now! Excited to get this to work with my @dair_ai co

This is why frontier open models are crucial. This is extremely sad for the research community. mythos will be bad ON PURPOSE on ai 'frontier llm research' tasks, this is very very sad for the researc

Also, I found that Hermes Agent + Nemotron 3 Ultra is a mighty combo! Excited to launch a new way to upskill with AI agents. This is how we are making it possible for anyone to learn to build with cod

Excited to launch a new way to upskill with AI agents. This is how we are making it possible for anyone to learn to build with coding agents. To start, we are launching 4 new hands-on labs on the foll

NEW: Anthropic introduces Claude Fable 5, a Mythos-class model for general use. Beginning of a new class of frontier models. Introducing Claude Fable 5: a Mythos-class model that we’ve made safe for g

// Self-Harness: Harnesses That Improve Themselves // (bookmark this one) Most of the agent scaffolds we rely on today are built once and remain frozen or mostly unchanged. The harness, like the skill

// The Consistency Illusion // Multi-agent debate can make agents agree on the final answer while their underlying reasoning stays misaligned. This work finds that consensus on the output hides disagr

Great tips. In practice, this is how it roughly looks to run agents autonomously for hours or days. /goal or /loop to keep it going. Verification is crucial here. Seeing a number of benchmarks showing

New paper on how AI agents are reshaping knowledge work. This is a nice economic read on where agents actually change knowledge work to meet that gap directly. (bookmark it) It studies agent adoption

The point is that you should start implementing ways to encode instructions/prompts with clear goals inside automations. Nothing new but newer LLMs are being trained to perform for longer duration uni

Great paper on self-improving agents: This was one of the standout AI papers of the week. (bookmark it) It tackles a question most self-improving AI agents ignore: is the agent actually discovering an

Super-powerful AI models will launch in the coming weeks. We are looking at a potential step change in model capabilities. The biggest mistake right now is to lock into one vendor. I say this not only

The Top AI Papers of the Week (May 31 - June 7) - LEAP - AutoLab - Learn From Your Own Latents - Reusable Context Engineering - Self-Revising Discovery Systems - Scaling Laws for Agent Harnesses - Dis

This was one of the standout AI papers of the week. (bookmark it) It tackles a question most self-improving AI agents ignore: is the agent actually discovering anything, or just remixing what it alrea