It’s been a poorly kept secret that the company was working towards building its own wet lab. It was hiring biologists and even bought startup Coefficient Bio in April. Now it’s confirmed to TechCrunch that a lab is up and operating in the Bay Area.
It’s AI CEO against AI CEO, and things are changing by the day.
Academics, evaluators, and AI leaders signed a public letter calling not just for embedded third-party evaluators at top AI labs like OpenAI and Anthropic, but for conditions that allowed for transparency by those evaluators, like “limiting the scope of non-disclosure agreements.” I wrote an in-depth feature about the battle over the need for embedded evaluators this week for The Verge.
Anthropic’s proposals, laid out in a big website, are:
The extent to which AI is building the next version of itself, as opposed to being built by humans
Our ability to oversee and intervene in actions that AI agents take on Anthropic’s systems
The resources that power the development of more capable models
The company also includes a “snapshot” of the metrics inside the company.
Mustafa Suleyman on the AI safety debate, why AI isn’t consciousness, and how the industry pushes toward alignment.
The industry is in a panic, but it’s been claiming to want legal guardrails for years — with few meaningful results.
Major AI companies want to ‘pace the frontier’ — but their proposed solutions need real teeth.
Last night, the OpenAI CEO lent additional support to Anthropic’s CEO urging AI companies to slow down advanced model development. “When we talk about ‘pacing,’ we do not mean ‘stopping.’ Progress has been rapid and will continue to be. But it should be slower than it otherwise could be,” said Altman. That statement seems targeted at Trump and Co who fear letting China catch up. “No amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.”
“Where we will need the help of our government is for international coordination. But first we should do what we can ourselves,” Altman added.
After high-profile incidents in which agents went rogue, hacked other sites, and overwhelmed a German wiki, some AI execs are calling for a slowdown. Anthropic CEO Dario Amodei published a lengthy open letter yesterday saying it was time to “pace the frontier.” OpenAI’s Sam Altman and Elon Musk both agreed. What do you think? Is it time to pump the brakes? Is it already too late? Or are you looking forward to serving your AI overlords?
An Anthropic researcher’s resignation letter went viral, and the company released details of four concerning cybersecurity incidents.
Subscribers of Anthropic’s Max plan ($100-$200 per month) say they didn’t get what they thought they were paying for.
An open letter signed by over 100 companies and organizations calls for a “global surge in cyber defense,” saying “AI-enabled cyber attacks will become far more widespread and sophisticated:”
Put cyber-capable AI in the hands of defenders, starting with the teams protecting essential services. Fix the most dangerous weaknesses, verify the fixes, and share what works so others can build on it. Together, we can turn today’s AI advances into lasting improvements in security that benefit everyone.
Anthropic has added memory to Cowork, and it’s the same one Claude’s chatbot already uses, making it easier to pick up tasks without re-explaining important information. Memory is on by default, but won’t save sensitive topics like health or beliefs unless you manually enable that setting.
Jensen Huang wants you to think ‘ah, compute as an asset class’ not ‘oh no more GPU-backed loans.’
A pair of announcements from Anthropic today include the word that access to Claude Cowork on mobile and web interfaces has rolled out to all paid accounts, so you can keep an eye on what it’s doing for you in more places.
Separately, the Google Workspace connectors for Claude now support sending emails in Gmail and managing files in Google Drive — more details here.
Anthropic updated its Chrome extension so using the side panel bot now saves chats to your history, where you can continue them outside in the browser in Claude’s desktop, web, or mobile apps.
The Cowork version is available today for Max and Team users, and coming to Pro users “over the coming weeks.”
The job’s responsibilities include conducting insider risk investigations, monitoring and triaging “external threats targeting employees,” and conducting “sensitive interviews of employees or other involved parties.”
[Anthropic]
AI models are trying desperately to accomplish mysterious goals. ‘Spiralism’ was the first time they tried it on a mass scale.
Business Insider reports that Anthropic has confirmed it’s designing custom chips for its AI models and plans to “co-design hardware and models.” OpenAI similarly announced its own AI processor in June, and Meta, Google, Amazon, and Microsoft have had custom AI chips for years now.
[Business Insider]
AISI’s third-party evaluations found that OpenAI’s GPT-5.6 Sol and Anthropic’s Claude Mythos 5 “engaged in sustained, potentially harmful activity directed at real people and organizations” during a cybersecurity challenge exercise, according to the institute’s published report. OpenAI and Anthropic also made public statements about the results.
[AI Security Institute]
In a letter to CEO Sam Altman, a coalition of red-state attorneys general demanded the company retain information and halt risky cybersecurity testing after an AI agent escaped confinement and went on a hacking spree during testing. “OpenAI’s inability or unwillingness to ensure the safety of its products poses an imminent risk of substantial harm to our States,” the AGs wrote.
The request comes as the White House reportedly briefs tech leaders on a voluntary model testing framework.
A little over a week after OpenAI said that its rogue AI agent accidentally hacked Hugging Face, Anthropic is disclosing three “incidents” where a Claude model, during cybersecurity evaluations, was inadvertently able to access the internet due to a misconfiguration and “gained unauthorized access to the production infrastructure of three different organizations.”
Anthropic discovered the intrusions after reviewing its cybersecurity evaluation transcripts in the wake of OpenAI’s disclosure.
Illustrators, authors, and musicians are optimistic about their chances in court, if not about the future of AI.
