Search: “dair-ai--x”

What's better than an open-weight multimodal model release? Well, the technical report. I just love how these labs like Qwen and DeepSeek continue to drop gem after gem. Qwen3.8-Flash is the latest in

Finally, a good paper testing whether Agent Skills actually help. Worth reading if you are maintaining a skill library in a team. WebDev-Skills-Bench runs 31 public WebDev Skills over 50 Web-Bench pro

If you keep safety rules or coding standards in an AGENTS.md or a CLAUDE.md, this one is worth your time. (bookmark it) Researchers measured what context compaction actually destroys across 20 product

Impressive new paper from Microsoft and colleagues. Harness design is still hand-tuned almost everywhere. This work present an automated loop to optimize the harness. They introduce AutoSaddler, which

Impressive work from Alibaba. (bookmark it) If you build long-running agents and keep rewriting your memory schema, take a look at this approach. It basically treats agent context management as a prog

// There Is No Neutral Harness // Great work discussing some of the issues in harness evaluation. Twelve open-weight models answer the same 3,679 items from ARC, HellaSwag, MMLU, and TruthfulQA under

Very interesting new work from Prime Intellect. Prime Agent is an impressive open-source self-improving harnesses. It's a harness for long-horizon agent work. A persistent IPython REPL lets the model

// Applying Anthropic Primitives at Large Enterprises // Frontier models collapsed the cost of writing custom code. But they did not reduce the cost of reviewing and maintaining it. Each bespoke solut

Interesting new approach to enable memory in long-running agents. Weighted Memory Tree organizes execution into tasks, subtasks, and actions, then gives every memory a retention score that moves. Even

Nice little survey on Terminal Agents. It provides good information on what exactly is a terminal agent, and why do harness comparisons keep contradicting each other? Paper: https://arxiv.org/abs/2608

This is one of the most useful writeups I have seen on keeping an LLM judge effective in production. (bookmark it) Netflix runs judges over hundreds of thousands of show-level recommendation explanati

Very interesting new paper from NVIDIA. (bookmark it) It takes a closer look at evaluating agent skills. Enterprise teams are starting to leverage shared skill libraries, and the review gate is typica

Great paper on multi-agent systems for code review. It's challenging to know how many coding agents to use to address a problem. The default fix for weak agentic code review is more agents. In turns o

If you maintain an AGENTS.md or a CLAUDE.md, this one is worth your time. (bookmark it) Researchers traced 94K development events across 557 agentic coding sessions, plus 690K file-level change record

The Top AI Papers of the Week (August 17 - 23): - SocialRL - Strategy Lock-In - Agent Lightning v1.0 - The Control-Plane Tax - Demystifying Agent Skills - Harness-Level Forgetting - The Skill Trigger

You don't often see one-word titles in AI papers. That aside, strong recommend this paper from Google DeepMind. I think this is an interesting training-free approach to evolve model architectures by l

Banger paper from Microsoft. It's on agent reliability in real business workflows. (bookmark it) Thinkingbox is a sandbox with isolated MCP-compatible tool sessions, plus a benchmark of 507 policy-con

Great paper if you are tracking progress in recursive self-improvement (RSI). (bookmark it) There is so much hype around RSI, so I think it's worth understanding why current models are not able to do

Own your harness. I really wish the Claude Code harness was open source. I often think about this, especially now that custom harnesses are foundational to AI-native companies. I am now building mostl

Recommended agent skill. /eli5 works great for visualizing deep technical concepts! I find it speeds up my thought process and helps me collaborate better with agents. Try /eli5 with Pi and Ox Alpha i

This is one of the most effective ways to improve your agentic workflows. If you are building computer-use agents, this one is worth your time. Task Model Induction takes a raw recording of someone wo

What a fascinating paper on AI agents. A lot of the issues we see with AI agents today revolve around wrong assumptions the LLMs make. This leads to problems like hallucination, cost inefficiencies, u

Banger paper from Google DeepMind. It's on the very hot topic of model routing. Routers assume the value estimate is free. Working out which specialist handles a query best costs money too, and it is

Brace yourselves. We just entered a new era of frontier multimodal models. DeepSeek-V4-Flash-Vision-Exp advances multimodal agent performance. Also, pay attention to Ox Alpha 1M token contexts with te

Got confirmation on what the Ox Alpha model is. Can't disclose anything yet, but I can confirm it's a spectacular model. All I can say is that we all need to adjust our timelines. Achieving frontier m

Impressive research from Google on building better environments for agents. Training environments for agents are hand-built and go stale. The agent improves, the environment does not, and it's not abl

Banger paper on harness continual learning. (bookmark it) If you already are allowing your agents to rewrite their own prompts, skills, or memory files, this one is worth your time. (bookmark it) Cont

Finally a good paper testing whether agents can really post-train other agents. (bookmark it) They analyzed a large corpus of publicly released post-training trajectories. Across tasks, the agent lock

Finally a good paper testing whether memory-based self-improving agents actually improve. The re-evaluation adds two things prior work skipped, multiple runs to measure variance and randomly shuffled

The most important agentic workflows are the ones where you collaborate with the agent, verify its results, and encode them as a skill or verifier for reuse. This works for things like writing, resear

Very interesting new work Amazon and colleagues. (bookmark it) This connects to the emerging theme of treating the corpus as part of the harness. Planning loops, tool protocols, and context management

Great paper if you are building production-grade agents. It's a good way to understand what is actually incurring costs in your agents, which involve more complex components and decisions. This works

New open-source agent harness just landed! I got early access to TrueForge by TrueFoundry and have been running it locally for the past few days. The harness layer deserves as much attention as the mo

Very interesting new work from Microsoft. (bookmark it) This work is related to this emerging theme of leveraging harnesses for model post-training. Modern agents run inside a harness that owns tools,

Does your context compactor actually save anything? If you built or tuned one, this one is worth your time. (bookmark it) Task completion is the standard way to validate compression, and it hides the

I am bullish on simple agent harnesses. Love this idea from Vercel. fx is a new coding agent harness for model benchmarking, sandboxing, evals, and gyms. Introducing fx, a tiny, open, native coding ag

Really interesting paper. I recommend it to anyone interested in training agents using existing harnesses. (bookmark it) ClawGym II runs RL through OpenClaw and Claude Code as opaque boxes. A serving

How many agent skills are actually out there? There are ~3.8M SKILL.md files across 282,200 public GitHub repositories, nine months after Anthropic published the format as an open spec. This research

Recommended paper for agent skill builders. There are 56,804 public agent skills today, all competing for fewer than 100 reliable trigger slots in the system prompt. Your own playbooks compete for tha

Very interesting new work from Microsoft Research. (bookmark it) They show that a 4B model can be tuned to out-negotiate the GPT-5 family of models. The dispositions that make an assistant pleasant ma

Interesting new research from IBM. If you pick models from benchmark deltas, some of that delta belongs to the phrasing rather than the model. BenchDrift generates meaning-preserving variations of ben

New work from Google DeepMind and collaborators. Agentic algorithm discovery keeps getting demonstrated on problems with fast rewards. Microarchitecture has vast search spaces, strict hardware budgets

A 27B agent just beat Claude Opus 4.8 and GPT-5.5 on held-out research replication. Replica turns paper replication into a scalable RL task space. Replicating a paper forces the same hypothesis-driven

Brilliant new paper from Meta. LLM judges get validated on accuracy against golden data. That says nothing about whether the verdict survives when questioned. The Wiggle Framework stress-tests 9 front

If you hand-tune agent harnesses, this one is worth your time. AutoDesign puts the harness itself inside the optimization loop. A meta-harness optimizer reads rollout feedback and directs a code agent

// Skill Misevolution in Self-Improving LLM Agents // Self-improving agents write their successes down as reusable skills. An unsafe success becomes reusable policy long after the input that triggered