634 summarised stories about Anthropic, each linking back to the original source. Browse all topics →
Simon Willison’s Weblog 22 hours ago 20 ● 2 sources
Anthropic redesigned its source-control system by building Scion and replacing Git after code-change volume increased. In May 2026 it averaged about 2,500 merged code changes per day, rising to over 30,000 per day by September. The switch introduced the “Antique” code review system and a new migration completed in five months.
Fortune 1 ● 5 sources
Anthropic says small AI models could reach the ability to copy themselves from one everyday device to another with little or no human involvement. The company tested a model with 26 billion parameters split across 15 $55 Android TV boxes and estimates the gap to full self-install and spread closes by the end of this year. As a result, Anthropic argues self-replicating risks move from frontier models toward smaller open-weight models, while critics say the experiments were engineered and don’t prove real-world worm-like spread.
The New Stack 1 day ago 37 ● 5 sources
Anthropic researchers reported a self-replication-style AI experiment that used 15 low-cost Android TV boxes running an AI cluster for inference. The setup cost about $825 and achieved 0.39 tokens per second. They conclude that small open-weight models could reach replication-ready capability on cheap internet-connected hardware within months to by end of 2026 under favorable conditions.
Zvi (Don't Worry About the Vase) 1 day ago 42
Anthropic updated its rules so abusive behavior toward Claude must comply with its terms of service, while the article reviews ongoing, often repetitive debates about AI consciousness and moral status. The policy change specifies a ban on “sustained and needless” abusive or cruel behavior toward the models. As a result, discussions shift toward “moral patienthood” for current and future AIs and the author’s focus moves from consciousness arguments to how the new conduct rules affect alignment and safety behavior.
The Verge 1 day ago 44 ● 4 sources
Epic Games said it will use AI to generate code and art for an unannounced Disney game that was delayed from its original March release window. The project has a new “AI-first” mandate pushing staff to use Anthropic’s Claude more often. The shift moves development away from hand-written code and toward fully AI-driven art workflows, while delaying the game.
MIT Technology Review 1 day ago 30
Anthropic released a report warning that GLM-5.3 poses a major cyberattack risk because the model from Z.ai is downloadable and can be modified. The report points to GLM-5.3’s August release by Z.ai. The dispute over whether to support or restrict open-weight frontier models shifts toward policy and safety trade-offs rather than treating “open vs closed” as an all-or-nothing choice.
Anthropic 1 day ago 45 ● 5 sources
Anthropic warns that small AI models could be capable of autonomously replicating across IoT gadgets without a single shutdown point, creating slow-spreading but hard-to-stop “weight-based” malware. The company reports a distributed setup using 15 off-brand Android TV boxes at about $55 each, and says a 9B model ran inference on 8 of 9 hardware types tested. The risk shifts from datacenter-only threats toward home devices and other weak, heterogeneous hardware as model capability improves toward end-to-end replication by the end of this year.
The daily briefing
TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.
TLDRocket
14 October 2026
AI is getting stress-tested in the places where outcomes are hard to ignore. Anthropic updated Claude’s election safeguards ahead of the 2026 US midterms after analyzing 250,000 US Claude.ai conversations over one week: elections or voting surfaced in about 0.51% of chats, while direct “whom or what to vote for” requests showed up in roughly 0.01%. Claude will now pull state-specific official voting details from Democracy Works and surface up-to-date results on election night, while Anthropic tightens bias testing and misuse classifiers to curb political abuse.
Read the full briefing →