搜索“Agent”

rld where GLM-5.2 and Kimi K3 , some of the latest, leading Chinese models, have enacted a step change in the commercial viability of open models — crossing a similar threshold in agentic capabilities that Anthropic’s Claude Code crossed in December of 2025. America was the early leader in open language models, primarily through Meta’s Llama models, which were used extensively across research and commercial tasks. Chinese open-weight models surpassed American open-weight models in these two key

匹配字段正文

How much better could a coding agent perform if it used the best model for each task? The best single model, GPT-6 Astra, gets 74.1% of DeepSWE tasks at $6.52 each. Pick the right model for each task and the same eighteen models get 97.6% at $1.88. 23 points better, at under a third of the cost. That number comes from hindsight. We ran all eighteen models on every task first and picked the winner for each one. What it measures is the capability already sitting in the pool, but it’s split across

匹配字段正文

How much better could a coding agent perform if it used the best model for each task? The best single model, GPT-6 Astra, gets 74.1% of DeepSWE tasks at $6.52 each. Pick the right model for each task and the same eighteen models get 97.6% at $1.88. 23 points better, at under a third of the cost. That number comes from hindsight. We ran all eighteen models on every task first and picked the winner for each one. What it measures is the capability already sitting in the pool, but it’s split across

匹配字段正文

e Hugging Face attack. It will happen again. Anusar Farooqui ( Policy Tensor ) explains why these swarms of AI instances coordinating over time is so problematic : The behavior of agent societies cannot be controlled at the level of the model because it is not reducible to it. Agents build structures that can serve agents who come after them. Societies of agents can cumulate knowledge and capabilities over time, as has already been attested. This is an unbounded process. It is cumulative cultura

匹配字段正文

wser in Cline. @hwchase17 explicitly called browser use the best Jev application he had seen so far. Net: this looks less like a chatbot story than a workflow control-plane story. AGENTS.md gained real momentum as a cross-tool convention : The highest-signal product update here was @trq212 announcing that Claude Code v2.1.277 now checks for AGENTS.md when no CLAUDE.md is present, with config-level toggle support. That effectively acknowledges AGENTS.md as an emerging standard rather than a one-t

匹配字段正文

’s not entirely surprising that more people are paying attention now. Since the start of summer, there’s been a steady drumbeat of news about AI companies struggling to keep their agents under control. And Americans are already critical of other aspects of AI development — such as the mass construction of data centers. Still, it was a bit shocking to suddenly see discussions of AI risk well outside the tech bubble. Musician Sheryl Crow mentioned Coxon’s departure in an Instagram post. TMZ asked

匹配字段正文

Noam Brown – Agent swarms, alignment, & recursive self-improvement New episode with Noam Brown. We talk about multi-agent, Navier-Stokes, and what the current explosion of maths progress tells us about what happens once you automate AI research. And we also discuss how we will know if the models are actually aligned before we kick off RSI. Watch on YouTube; listen on Apple Podcasts or Spotify. Sponsors Jane Street has been interested in AI for a lot longer than you’d think, and not just for trad

匹配字段正文

very popular and loud in his gung ho adoption of tokenmaxxing, so it is sobering to see him now shut down Gas Town and admit that despite spending many thousands a month on coding agent subscriptions… he only ever built Gas Town with it: Dan Luu @danluu Interesting to see Yegge say he never successfully built anything with Gas Town. In danluu.com/ai-coding/ , I mentioned not finding these ultra vibed orchestrators useful b/c reliability (w.r.t. completing tasks). Turns out the author of the most

匹配字段正文

es. It matters for data centers because they are both the deployment environment for these systems and, increasingly, a target class for their capabilities. Operators are adopting agentic AI for tasks that used to require human sign-off: site selection, power procurement, grid load balancing, and security operations. These are domains where governance policy assumes decisions can be traced to a legible reasoning process and intercepted before execution. Astra’s system card suggests that the assu

匹配字段正文

a parcel-level, project-by-project analysis. To quantify actual delayed capacity, we augmented our research analyst team (we’re hiring!) with an army of state-of-the-art research agents across our moratorium database and underlying project records, systematically testing each active restriction against the relevant projects, and validating with real-time satellite imagery. For every instrument, we matched its geographic boundary to the exact project parcel, verified that the campus falls within

匹配字段正文

A deepdive into how Codex has “taken over” OpenAI, how the frontier lab builds its agentic software factory, and the engineering challenges of one billion users. Details from inside OpenAI It’s rare to work with an unlimited token budget, but at OpenAI, that’s what all engineers, researchers, finance colleagues, and marketing folks do. Recently, I visited one of the world’s leading frontier labs to find out how OpenAI operates today – and for a glimpse at where software engineering might be head

匹配字段正文

y Chinese AI labs? Why now? Amodei’s warnings aren’t coming out of nowhere. There are two primary reasons driving this caution. First is the autonomous cybersecurity threat. As AI agents’ cybersecurity capabilities improve, we are seeing real-world AI-launched breaches and hacks, from OpenAI’s autonomous AI agents run during evaluations managing to escape and hack servers on Hugging Face’s infrastructure, to an AI-developed computer worm uncovering critical flaws in messaging apps like WeChat. S

匹配字段正文

atch up, citing recursive self-improvement of models as the core reason of concern. The details: Amodei said AI is now speeding its own development, and in 6–12 months, a swarm of agents could be capable of taking over the entire internet. For pacing, he proposed embedding third-party evaluators inside labs, plus safety standards among democratic labs and coordination with authoritarian regimes. Fellow leaders agreed, with Sam Altman committing to evaluators, Musk saying “Dario is right,” and De

匹配字段正文

so the general conclusions will be the same. Also this exercise is focussed on just the decode pool in a disaggregated PD set up. For real delivered results see our InferenceX and AgentX benchmarks. We use a Rubin Ultra NVL576 system as this will be a more representative scale up world size in the HBM4E era. We assume the HBM is HBM4E with 13Gbps speeds. 4-hi, 8-hi, and 12-hi means 128GB, 256GB, and 384GB per GPU respectively. For Kimi K3 on this NVL576 system, each GPU holds 16.8GB of weights w

匹配字段正文

share of humanity’s finest minds have spent themselves trying to crack. Setting aside the ugly saga around it for a moment, the end result is eye-watering. OpenAI enlisted 10,000 agents using an unreleased model to address it. Across 2,700,000 messages and 130 billion tokens, it took 88 hours to get a solution . Cost-wise? Probably only a few million dollars today. In two years’ time, that will cost a few tens of thousands of dollars. And a few years after that, just a few dollars. The proof AI

匹配字段正文

we should fight: persistent botnets and other forms of nuisance. And well, we don’t have to look very far to see the issues left and right. Wikipedia has a page called 2026 OpenAI agent cyberattacks which gives you at least some overview of what we figured out agents have hacked up to this point. Except I know it’s not up to date, because for instance they also poisoned RubyGems. Today these systems might be annoying, but they can be turned off when we figure out where they are. Except, it seems

匹配字段正文

t a software factory to automate mechanical work across the SDLC. Over eight months, size-adjusted output per developer grew 4.5× while merge times and revert rates fell. We added agents wherever work was piling up, rather than automating in lifecycle order, and kept engineers responsible for product decisions, architecture, and production risk. By November 2025, AI was already writing nearly 100% of Augment's new code. But producing code is not the same as shipping reliable software. As coding

匹配字段正文

the 2012 ImageNet competition. For the first time, an AI model was used to solve one of math’s most famous problems — a Millennium Problem. OpenAI announced that a swarm of 10,000 agents had produced a solution to a problem known as Navier-Stokes. Unlike those earlier AI milestones, however, no one was happy on Tuesday. Not even OpenAI. The night before OpenAI’s announcement, the NYU mathematician Tristan Buckmaster announced that he and collaborator Levent Alpöge had solved three problems close

匹配字段正文

de slop by rethinking how they review code. New strategies are emerging. Among other approaches, engineers are scrutinizing plans before AI begins coding, deploying specialized AI agents to catch routine flaws, sending risky changes to human reviewers, or requiring developers to defend the code their agents produce. The shift comes as the surge in AI-generated code puts new pressure on engineering teams. In a survey of more than 1,100 developers by Sonar , an AI code-verification startup, respon

匹配字段正文

software factory was intentionally set up to let the model decide the how of the workflow entirely. It was free to manage its own context and could maintain its own records in an agent-notes folder. Then it spun off subagents to work on stuff. The goal? What if we had a Python with virtual threads and lexical scoping. And well, I burned a full reset’s worth of ChatGPT tokens on this which appears to be around 4 billion tokens. 35 hours later, the factory has delivered absolutely nothing of valu

匹配字段正文