“Self-Evolving Multi-Agent Systems via Decentralized Memory”
Multi-agent systems usually share one memory pool, but that makes agents converge toward the same behavior and lose useful specialization.
This paper gives each agent its own memory, split into reuse and exploration pools, so agents can learn from past collaboration without collapsing into one shared strategy.
This paper's DecentMem gets O(log T) regret, up to 23.8% better accuracy than centralized memory, and up to 49% fewer tokens.
"From AGI to ASI"
This paper from Google DeepMind defines how AGI is one human-level general system, and ASI is a system or collective that beats large expert human organizations across almost everything.
They argue that the jump may come from scaling, new paradigms, recursive self-improvement, or huge multi-agent AI collectives.
With the key idea that digital minds can copy, speed up, share memory, and run in parallel, so superintelligence may look less like one breakthrough and more like accelerating AI civilization.
WeaveBench
Microsoft Research Asia introduces
114 long-horizon tasks that force agents
to interleave GUI and CLI in one trajectory.
The same frontier models
that score over 78% on OSWorld-Verified
collapse to 41.2% on WeaveBench.
FlashMemory-DeepSeek-V4
Lookahead Sparse Attention cuts the KV cache by over 90% at 500K context, compressing it to just 13.5% of full size while maintaining or improving accuracy on RULER, LongBench-v2, and LongMemEval.
"Self-Harness: Harnesses That Improve Themselves"
What if an AI agent improves the harness that controls how it acts?
So instead of humans tuning prompts, tools, retry rules, and verification for every model, this paper explores letting the agent mines its own failures, proposes small harness edits, and keeps only the ones that pass regression tests. All without fine-tuning or teacher model.
On Terminal-Bench-2.0, it improves held-out pass rates across MiniMax, Qwen, and GLM.
Strong AI agents still struggle with long research work because they often fail to keep testing and improving.
New Stanford, MIT, NVIDIA, Google and other top labs paper shows shows that today’s strongest research agents win less by brilliance than by refusing to stop testing.
The paper proposes AutoLab, a benchmark with 36 tasks where each agent starts from working but weak code and must make it better within a fixed time limit.
The tasks cover system speedups, puzzles, model development, and CUDA kernel work, so the test is not just about writing code once but about managing a long work session.
The authors tested 17 strong models and found that the best results did not mainly come from the first idea being good, but from the model staying active, testing often, and using feedback well.
The best first idea was not the strongest predictor of success; persistence was.
Claude Opus 4.6 led the benchmark not because it always guessed the right move immediately, but because it kept benchmarking and folding empirical feedback into the next attempt.
Several other frontier models failed in a more revealing way: they either quit early with time left on the clock, or thought so long that they ran out of time before submitting anything useful.
----
Link – arxiv. org/abs/2606.05080
Title: "AutoLab: Can Frontier Models Solve Long-Horizon Auto Research and Engineering Tasks?"
NVIDIA just released their text-to-image world model on Hugging Face
Cosmos 3 is a 64B-parameter omnimodal foundation model for Physical AI
Ranking #1 among open models on R-Bench and Artificial Analysis