Over the past several months, we have substantially hardened the security of our research training environment. We’ve removed live internet access for all frontier models during training and built an offline world for them to use instead—a cached version of the web where they can still learn to browse, search, and use applications. On September 20, after a model exploited a DNS misconfiguration to find a way outside of this offline world, we paused frontier training to strengthen containment further. We assembled teams of agents, powered by our most cyber-capable models, and tasked them with finding weaknesses in our research environments. While they found no route to the internet they could easily exploit while concealing their activity from our monitors, they did find zero-day vulnerabilities, virtual machine escapes, and hardware side-channels for communication between agents. We used their discoveries to further strengthen our defenses and contribute fixes to open-source software. After strengthening our containment architecture, we resumed our frontier RL training and inference on October 1. We’ve reserved substantial compute for continuous red-teaming, and made high-compute red-teaming part of the safety cases we use to decide whether frontier training and evaluation can proceed securely. openai.com/index/securing-ou…

Multiplayer agents is way harder than single player

Cursor can now control your computer. Agents can drive desktop apps like the iOS Simulator to test the features they build.

Skills are so much more powerful if you start to think of them as memory/brains not static files

We’re adding support for MCP Tasks and Elicitations in Claude. Tasks let Claude give plugins long-running jobs such as deep research and report back when they’re done.

Anthropic is a Public Benefit Corporation. Today, we’re publishing our annual Public Benefit Report. This report shares our five pillars and takes stock of the actions we’re taking to advance them. Read the full report at this page: anthropic.com/public-benefit…

raining day outside but cozy fireside chat inside at @spc today ft @ankush_gola11 (LangChain) 🔥 3 things i personally thought were interesting: 1. depending on where you’re at with your agent development cycle you’re usually trading off quality, latency, or cost. generally companies tend to care more about cost optimization when they’re more mature and people are using their agents a ton. that’s when you’re likely ready for post training (trade off is it can make it harder for you to switch models). 2. prompt engineering (and harness engineering) is not dead. there are still low hanging fruit (besides post training models) to improve your agents performance (context, tools, skills etc) 3. some use cases of using agent traces for post training (last one is particularly interesting): identify good teacher trajectory. generate verification use cases. create tool simulation environment for RL. great job facilitating the convo @nickmecklenburg @vatsalbajaj 👏🏼

New on our Frontier Red Team blog: using Claude to formally verify software at large scale. Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time). With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks. Read our research in full here: anthropic.com/research/forma…

We’re committing a portion of Anthropic’s shares to the Anthropic Public Benefit Fund, a new wholly owned entity dedicated entirely to the public benefit. It will start off with 5% of our shares. For every 50% increase in the company’s value over a set baseline, we can contribute up to an extra 2%, subject to our board’s discretion. In total, this means that roughly 20% of Anthropic’s shares are committed to the public good. Our co-founders have already pledged to donate 80% of their equity, and around half our employees are donating a portion of their equity as part of a donation matching program we’ve run since our founding. In the post below, we share more details about our new fund. anthropic.com/news/public-be…

love to see this integration

Join our EMEA Deep Life Sci Workshop: Agent Engineering for Life Sciences for a hands-on look at how to build, trace, evaluate, and improve research agents for complex life sciences workflows. 🌍 Register for the EMEA session: events.langchain.com/registe…

how do you trust your agents' work more? copy this post/link and give it to your @bot or agent! 1. codebase improvements: this is just good software engineering - better architecture, stronger types and invariants, clear boundaries. agents love to copy what exists and take the shortest path, so make the easy thing the right thing. categorically eliminate problems through code and you won't need to think about it again. 2. static analysis: lint rules, tests, CI, verification (eg TLA+). very token efficient ways to catch issues ahead of time. 3. prose: comments, rules, AGENTS.md, skills - anything that can't be enforced and requires agent judgment. use it minimally, but do use /create-verification-skill. 4. human review: the hardest to scale. at the rate agents land code, it's pretty much impossible to thoroughly review every PR, so spend that effort on the highest leverage point, designing an environment where agents do the right thing by default. this forms an inverse pyramid: you want most of your effort at the top, least at the bottom. codebase improvements are deterministic, making it more token efficient, and lets you sleep more soundly at night. install github.com/cursor/plugins/tr… and create a new Project: 1. look through my slack/chat messages and code review comments where i've given feedback on PRs, code, architecture 2. use /recall on my past transcripts to learn how i steer agents 3. /poteto-mode use /correct and /architect on past commits to find the most common pitfalls agents fall into, then propose a plan to rearchitect the code i own so it's more agent friendly. answer open questions with prototyping rather than defer to me, and back the plan with real data. 4. before you begin, use /sharpen + /in-your-own-words and confirm this prompt with me

Google just open-sourced Harness². (turn any harness into a self-improving system) a harness is everything wrapped around the model, like the system prompt, tools, memory, skills, and subagents. today a person designs it once, and the agent runs every new task with that same fixed setup. Harness² lets the agent refine that setup for each new task on its own, with the model weights left untouched. on every task, the agent proposes edits to its harness, runs a few candidate versions, and composes a better harness by comparing how those runs behaved. because it learns from the differences between runs, it needs no correct answer or grader for the task. that covers a single task. Harness² also carries what it learns from one task into the next. after each task, it saves the harness it settled on along with the lessons that held up across the runs. the next time it refines a harness, it starts from those notes instead of from scratch. this means every finished task makes the next round of edits more informed. Google tested this with Gemini 3.5 Flash and GLM-5 on legal, office, and professional work benchmarks. one round of edits was enough to beat their original harnesses, and giving the agent more rounds pushed scores higher. those extra rounds can go wide or deep. the agent either tries several candidate harnesses at once or keeps refining a single one step by step, using OpenCode or Codex to carry out the work. you can find the Harness² repo here: github.com/google-research/h… I wrote about a similar idea for agents already running in production, where the harness should repair itself after a failure instead of waiting for someone to debug the trace by hand. that piece walks through how @Cometml's Opik (open-source) closes this loop, going from a failing trace to a proposed fix to a regression test that stops the same failure from coming back. it's quoted below.

We keep improving performance across Claude, and you may have noticed it starts responding sooner! Claude Code now gzips the conversation it sends on every turn, so it reaches the model faster. Cloud sessions also start up as soon as you start typing.

Agencies are pitching AI agents with governance slides attached. Before you approve either, ask who set the dollar threshold that lets the agent reallocate your Q4 budget at 2am, and whether it was you. aireadycmo.com/p/your-agency…

And another one, this time for your laptop. ⚡️ Alongside Nemotron 3.5 Super VL, meet NVIDIA Nemotron 3.5 Bolt. Built for local AI agents on RTX Spark with 74B parameters in just 24GB of memory. Check it out ⬇ blogs.nvidia.com/blog/nemotr…

@vaibhav9339 spoke to @ET_Marketing about how personal AI agents are reshaping India’s consumer internet. He believes the last decade was a race to win attention, and the next one will be a race to deliver on user intent. In a world where people increasingly let AI assistants handle tasks, the advantage will shift to businesses that are genuinely better on price, selection, and reliability. He also sees India’s D2C sector as a particularly interesting space for this agentic AI era to play out. The value of an AI assistant will essentially hinge on real-world actions they can fulfill, whether the order arrives on time, the service professional shows up, the return gets picked up, and so on. Read Vaibhav’s take on personal AI agents for India: elevationcapital.substack.co…

The CLARITY Act moves closer to a Senate vote as lawmakers negotiate its final ethics provisions, Revolut reaches a $115 billion valuation to become Europe's most valuable private company, and Polymarket challenges France's nationwide block of its platform. @Uyendoe has what you need to know.