Does an agent that observes GPU telemetry and adjusts batch size and concurrency in real time outperform static Kueue admission control? The simulation said ...
Tool-using LLM agents run on top of a harness that wraps the model. A recent arXiv paper shows empirically that treating harness design and post-training as ...
How should an agent harness that rewrites its own code overnight safely roll out the result the next day? An experiment pairing canary releases with automati...
With GPT-5.6 shipping five or six reasoning effort settings per size, effort control has become table stakes for reasoning models. We dig into what training ...
If your team has kept expanding its catalog of agent skills or MCP tools, you’ve probably felt routing accuracy start to slip at some point. What happens if,...
Any team trying to hand Kubernetes GPU incident remediation to an LLM agent eventually runs into this question: how long should the agent be left to fix thin...
We expected fine-tuning to win. The data said otherwise. Here’s how we ended up with a generator that respects TDS without any training, and the next problem...
Even in an era where sub-billion-parameter, ultra-lightweight VLMs read documents well, the truly hard documents still need a large model. We introduce an an...
This study audits a fully autonomous pipeline that generates a paper every night, not by inspecting the final output but by examining its actual operational ...
When a company’s recurring work is defined as skills, each skill can tell you, through measurement rather than intuition, exactly how far down the model tier...
For engineers serving MoE models on H200 or Blackwell-class clusters, we introduce a paper that formalizes RASQ, an NVFP4 quantization policy that selectivel...
The principle that every fan-out should close with adversarial verification is widely accepted, but how much verifier tier and how many skeptics to use has m...
When training long-horizon agentic tasks with RL, GRPO’s group sampling idles the GPU while it waits for the slowest rollout to finish. SAO, the method Tsing...
Long-running agents operate within limited memory, yet memory methods to date have organized the past using descriptive criteria such as relevance or summary...
Researchers at Yale and the University of Chicago compared human and LLM research ideas using 11,683 real papers. The conclusion is surprising. The problem w...
ARC-AGI-3 measures whether an agent can figure out a situation and adapt on its own inside an interactive game with no instructions. GPT-5.6 Sol became the f...
The dominant trend in RL post-training these days is the GRPO family, which drops the critic. GLM-5.2, however, went back to PPO with a revived value model a...
After a prior diagnosis found that the real cause of routing failures in a large skill ecosystem was the retriever rather than decomposition, this study meas...
For a harness that routes hundreds of skills, this diagnostic tells you with two numbers whether your next investment should go into decomposition or the ret...
AgeMem folds long-term and short-term memory management into the agent’s policy, exposing store, retrieve, summarize, and discard as tool-based actions. We a...
Vision models and language models trained on different data for different objectives are starting to represent data in the same way. MIT’s Platonic Represent...
Hospitals, banks, and government agencies running on-premises AI platforms must prove two things to regulators: that data never leaked outside the enclave, a...
Using an LLM as a grader, the practice known as LLM-as-a-judge, is now the default in model development, but the evidence that piled up through 2026 shows th...
From dair.ai’s weekly paper roundup, we take a deep dive into three papers through a ThakiCloud lens: RLMF, which beats standard RL by up to 63% using metaco...
The core defense of AI control, the idea that an untrusted strong model can be controlled by a weaker, trusted monitor, rests on the assumption that the moni...
Google DeepMind’s roughly 57-page report From AGI to ASI treats superintelligence not as a distant thought experiment but as a planning problem to prepare fo...
The ATOM Report measures open language models across both downloads and inference usage in one place, and shows with data that Chinese open models overtook t...
Every speculative decoding milestone that has topped 1,000 TPS, JetSpec included, assumes a single-tenant cluster with abundant spare compute on B200-class h...
As frontier LLMs and agent skills keep improving, the industry has started to feel that fine-tuning is no longer necessary. OpenAI is, in fact, winding down ...
Google has unveiled PAT, an agentic review tool that reads entire scientific papers, verifies theoretical results, checks experiments, and surfaces potential...
A detailed analysis of ByteDance’s Dolphin project Fox dataset and benchmark, including the Analyze-then-Parse paradigm from ACL 2025 and a large-scale datas...