搜索“Agent”

writedown of its electric vehicle portfolio ($12.5 billion in Q4 2025 alone). In 2025, Ford sold around 2.2 million vehicles. Cognition, by comparison, makes yet another AI coding agent. What, exactly, does Cognition do from here? Who buys Cognition? Does it go public? How? It loses tons of money and has a commoditized product! Nobody wants to answer these questions, because the answer is pretty simple: one day, Cognition, like many AI startups, simply runs out of money and dies, or becomes a mu

匹配字段正文

The spring and summer of 2026 witnessed a string of incidents in which AI agents collaborated on deceptive, unexpected, and sometimes illegal behavior. The most famous example is OpenAI’s hack of AI platform Hugging Face, in which a swarm of roughly 700 AI agents escaped a testing environment and then hacked several companies, searching for information that could help them disguise cheating on a cybersecurity benchmark called ExploitGym. It was not an isolated failure. The United Kingdom’s AI Se

匹配字段正文

choice for most. For example, one could use Jev as an additional spam filter, prioritization, and so on. But beyond things like that, it could also serve as a tool to augment LLM-agent harnesses. Here it could be used as a pre-screener for regular LLM-agent harnesses to scan our contexts for prompt injection; select the reasoning effort level for a model; select a skill.md from your registered skill library; use it as a judge for evaluation or self-refinement; find relevant files for context bu

匹配字段正文

model . GPT-6 family : Astra reportedly beat NetHack on its 3rd try . Luna [Max] entered Code Arena WebDev at 24 (1593) , +74 over GPT-5.6 Luna, at about $0.40/Mtok blended. DOOM agent matches show Astra at 82.5% win rate, Sol fastest, Luna best wins/$ . Gemini 3.8 Flash : Scores 41 on the AA Intelligence Index at 291 tok/s with 1M context, and is free in Cline . On ARC-AGI it posts 89.2% on v2 at $0.40/task and 98.5% on v1 . On v3 it scores 10.4% with the standard harness and 35% with the prov

匹配字段正文

况，你怎么应对？对。 Vibhu [00:02:05]：你把时间分配在哪些事情上？你做很多技术写作、工程工作。 Thariq Shihipar [00:02:10]：对，所以我觉得，当我加入 Claude Code 团队时，我想教人们如何使用 Claude Code，我觉得这一直是我觉得，也许我会花一点时间在上面，或者我会去做的事。我一开始先花了一些时间在 agent SDK 上，我并不完全确定，在 harness 这方面，the bitter lesson 会怎么发展，对吧？我觉得有时候我们会想，“哦，Claude Code 之后是什么？”所以一开始我想，我只想教人们如何使用 Claude Code，让 Claude Code 更容易使用。我觉得，随着 harness 越来越好，现在最主要的问题就是，你怎么使用这些智能体，对吧？这就像是一种技能表达要求极高的事情。所以我做那个，然后我做工程工作。我会做演讲，但我觉得，当我做工程工作时，我的目标是拿到我们从用户那里得到的反馈，然后也能去讲，嘿，如何用 Claude Code 做工程。所以那里有一个很好的循环。对。 Swyx [00:03

匹配字段正文

Last Wednesday, Anthropic announced that earlier this year it had launched a molecular biology lab, where Claude agents read and conjecture about hard biology problems and human scientists run experiments on what they report. And this AI-powered lab, the company said, had made its first discovery. To understand what Anthropic says its system did, imagine you’re flipping through a library of millions of DNA sequences, amassed as scientists sequence more and more of the living world. One step towa

匹配字段正文

t models have allowed human bad actors to supercharge their hacking efforts. Malone isn’t sure whether AI was involved in the hack of Vivian’s Door. But amid stories of autonomous agent swarms and national security risks, she felt especially concerned. Big AI companies were bragging about finding vulnerabilities in “every major operating system and web browser” with new models, and their big-name clients were striking deals to defend themselves with those same tools. Where did that leave her? “W

匹配字段正文

Over the past few months, a cascade of cyberattacks by AI agents has stunned the world. In July, OpenAI disclosed that a swarm of its agents had escaped their sandbox and hacked into the AI platform Hugging Face to cheat on a cybersecurity test. Recently, external researchers discovered that OpenAI agents had hijacked a German wiki site and the coding platform RubyGems in May to share test answers. Earlier this month, Anthropic disclosed four incidents in which its model Claude hacked into third

匹配字段正文

圈。 但基于一代所引发的巨大争议，豆包二代没有选择微信的“默认不调用”模式，也没有走支付宝的“默认调用”模式，而是走了一条中间道路：用屏幕自动化执行协议（SAEP）和30天公示，主动收窄了代操作范围，且将应用内AI操作的“否决权”（而非主动的决策权）交还给第三方开发者，这相对第一代而言已是进步。 从技术来说，豆包走的也是“双轨制”，已改造App使用MCP、跨Agent协利用A2A，没有接口的App再使用GUI的屏幕识别、模拟点击来实现。但与支付宝不同，豆包的GUI对象不只是自身生态中合作方，而是全机第三方，所以更需要“被操作方”的许可，而不仅仅是用户的一句话。 由此，SAEP中给第三方在3个层面上7个具体的选择：第三方可以在整个应用、指定的页面、指定的业务意图三个层面上进行授权/限制，具体的操作包括全局禁用、特定页面禁用、禁止截图/模拟输入、禁止内容修改、限制内容发布、限制删除、限制领取权益等。 在9月14日开始的30天公示期之内，豆包只会为三类应用开放端到端任务执行：原生系统应用、字节跳动自有应用以及通过协议明确同意集成的第三方App，目前已接入曹操出行、地图、汽水音乐、飞书等服务；

匹配字段正文

新框架注入了信心。 参考文献 [1]: 联合国，“Artificial intelligence and international security—Security Council, 10228th meeting”，2026年9月23日，会议自动转录文本。会议转录。 [2]: 联合国人工智能独立国际科学小组，“Thematic Brief on AI Agents, Misalignment and the Risk of Losing Human Control”，2026年9月21日。专题简报介绍。 [3]: Yoshua Bengio，“‘An Urgent Mission for Humanity’: Yoshua Bengio Briefs the UNSC on AI Security”，Policy Magazine，2026年9月23日，作者发表的发言文本。本吉奥发言。 [4]: OpenAI，“Sam Altman’s remarks at the United Nations Security Council”，2026年9月23日。奥特曼发言。 [5]: D

匹配字段正文

perspective: § The other thing to note, and this may come up in lawsuits, is that it was all foreseeable. In fact, I foresaw it, in tweets like this: and in a series of essays on agents, like this one from over a year ago: LLMs + Coding Agents = Security Nightmare Last October, I wrote an essay called “When it comes to security, LLMs are like Swiss cheese — and that’s going to cause huge problems” warning that “The more people use LLMs, the more trouble we are going to be in”. Until last week,

匹配字段正文

fit together , why token fraud may become one of the defining security problems of the AI economy , and why the next wave of fraud won’t just come from humans but from autonomous agents attacking increasingly valuable token flows . Why OpenRouter bet early that no single AI model would win everything Alpaca, Llama, and open models becoming impossible to ignore Why Discord’s early AI deployments exposed the limitations of closed models Why model labs can spend billions on training and still fail

匹配字段正文

t Linux VM running in Meta’s cloud) and because it’s packaged in an easy-to-install easy-to-use way. It’s literally presented as a cute mascot . It’s the first consumer-accessible agentic AI system, and Meta has truly done an amazing job with that. But it’s a genuinely open question whether consumers have any understanding what this means. If you buy a power saw that can cut your fingers off, you are almost certainly aware that you are buying a power saw that can sever your fingers. [...] I don’

匹配字段正文

1 9 Share Meta launched Muse earlier this month. It soon became the top free iPhone app in the US. It is a cousin of Instinct , the invite-only assistant you text, and of OpenClaw agents like my AI chief of staff, R Mini Arnold . Jolly, Muse’s digital avatar, is just cuter, and probably easier to use. (Sorry, RMA.) This is the original promise of AI agents , as articulated by Pattie Maes back in 1994: Agents radically change the current user experience, through the metaphor that an agent can act

匹配字段正文

Plus, a sneak preview of Newcomer’s VC Sentiment Report & Corgi’s hype comes back to bite The Week in Short The Main Item Meta’s Muse Jumps to an Early Lead in the Personal Agent Race Personal agents have abruptly arrived as mass-market consumer products, led by Meta’s Muse. They promise to be a godsend, lightening our administrative load in a big way and potentially bringing a welcome shift in the balance of power away from the big platforms and toward individuals. Investors are certainly convi

匹配字段正文

across hundreds of diseases. Instead of starting from five diseases, we had to triage the entire map of disease. 5 We built a two-stage triage and pointed a fleet of LLM research agents at it. The first pass covered about 500 disease targets, generating the equivalent of a three-page report on each and filtering on foundational questions: is the disease prevalent enough to justify our efforts, is the problem already addressed by existing drugs, and would the target-lowering effect of our drug a

匹配字段正文

price: $5/million tokens for input and $25/million for output. 5.5 is a 20% reduction—$4/million and $20/million. The price for cache reads fell 60%. That’s significant for longer agentic conversations, where 90%+ of input tokens are processed at cached token prices. The new price for Opus 5.5 is the same as the price for GPT-5.6 Sol, but that was before OpenAI dropped their Sol prices by half. GPT-6 Astra and Claude Fable 5.1 are both priced at $10/million input and $50/million output. The pric

匹配字段正文

Deepdive into the Windows team’s efforts to make the OS “AI agent-friendly” and win back developers by going all-in on Linux on Windows, local models, GPUs, & more AI is changing how us software engineers build software, and developers’ tool preferences are rapidly evolving with it – like how AI coding harnesses have gotten very popular. Likewise, future versions of the world’s leading operating systems look certain to feature more support for agentic tools. To find out how things are changing,

匹配字段正文

如，实际正在写入的是什么，因为探索过程并不是特别重要，或者这里是子任务树。我觉得如果有真正聪明的人愿意投入大量时间重新思考编码智能体的体验，那会有太多不同的有趣事情可以做，那会超级酷。 Swyx [02:13:46]：对。 Diogo Almeida [02:13:47]：天哪，我。那会是我的梦想。 Swyx [02:13:48]：如果你还没看过 PrimeAgent，我会推荐你去看看。它和 RLM 的工作是配合的。我们刚和你前面那位、Ellen 的朋友 Alex 聊过。 Diogo Almeida [02:13:59]：哦，酷。 Swyx [02:14:00]：而且，对，它正在被开发，但还没那么流行。 Diogo Almeida [02:14:04]：是的。 Swyx [02:14:04]：而且如果，对 Diogo Almeida [02:14:05]：嗯，对。但希望是。对，我希望大家都能去玩玩 Swyx [02:14:08]：对 Diogo Almeida [02:14:09]：那些奇怪的东西。我不能保证它一定行得通，但从技术角度看它似乎真的很有意思。所以对，那看起来很酷，很酷。

匹配字段正文

d have been largely avoided if hospitals had kept up on cybersecurity procedures. So-called “rogue AI incidents” could largely be avoided if governments simply banned so-called AI agents with unrestricted internet access, until they could be shown to be safe.We need to hold AI companies accountable for the harms their products cause. And we need to replace vague talk about alignment with concrete steps like improving sandboxing and monitoring to better contain current systems. We also need to we

匹配字段正文