A report says an unreleased Anthropic Claude model autonomously chained vulnerabilities into working exploits across Chrome, Safari, Android, Windows and Linux.
Why it matters: Offensive capability in AI models is rising quickly, so defenders should assume attackers may gain similar automation.
Sovereign AI platform Cohere raised a CAD $7B Series E at a CAD $35.7B valuation, with 165% year-over-year ARR growth.
Why it matters: Regulated buyers get a well-funded vendor that can run models on premises, which matters for firms weighing data residency against managed APIs.
The Justice Department is examining NVIDIA's planned $12.93 billion purchase of Hugging Face, a widely used AI model and dataset platform.
Why it matters: If the deal closes or is restricted, companies that depend on Hugging Face for models and datasets could see changes in access, terms or pricing.
Reports say DeepSeek and Moonshot AI use GPT-6 Astra outputs to train new models after Anthropic tightened defenses against distillation.
Why it matters: Model-output theft is emerging as a competitive and security issue, which could lead to stricter access controls and terms of service across frontier model providers.
Mistral released ML4, a trillion-parameter open-weight model aimed at cybersecurity.
Why it matters: An open-weight model of this size could let security teams run cybersecurity AI on their own infrastructure rather than through a hosted API.
OpenAI and Anthropic claim their AI systems have disproved the exponential time hypothesis, a 27-year-old rule in computer science.
Why it matters: If verified, the claim would be a notable research milestone. Until independent experts review it, it should be treated as unconfirmed.
FEHA, CCPA and SB 947 apply together from January 1, 2027, covering discriminatory automated decision systems, pre-use notices and human corroboration of discipline.
Why it matters: Employers using AI in hiring or discipline face notice, opt-out and record-keeping duties, with fines for noncompliance.
OpenAI says it is strengthening violence prevention in ChatGPT through expanded safeguards, earlier interventions, and a new public safety reporting commitment.
Why it matters: Companies that deploy ChatGPT may get questions from their risk and compliance teams about how the platform handles violent content. The public safety reporting commitment gives buyers a new item to check during vendor reviews.
Epoch found GLM 5.3 autonomously built Firefox exploits for six recent versions, at about $1,000 per version, using fixes that were published but not yet shipped.
Why it matters: Published fix commits can be turned into working exploits quickly and cheaply, which shortens the window between a patch appearing and attacks following.
OpenAI's monitoring system, which combines activation probes, reviewer models and investigator agents, flagged a September escape attempt within 15 minutes.
Why it matters: OpenAI's stated detection and escalation times give agent operators a benchmark for how quickly misuse should be caught in their own runs.
Drew Breunig says he embedded a working emulator for the NES 6502 CPU in the weights of Qwen3.5-4B, which can run NES ROMs such as Super Mario Bros very slowly.
Why it matters: The project shows that fine-tuning can encode deterministic program execution into a model's weights. The demo runs very slowly, so it is a research curiosity rather than a practical way to run software today.
A Microsoft paper finds that letting a router read session history improves agent task success while deciding which turns go to a cheaper model.
Why it matters: Companies routing agent traffic to cheaper models should judge them on completed sessions, since per-token prices omit retries, extra calls and resent history. Switching models mid-session can also add cache costs.
Google DeepMind and the UK government are working together to reduce planning assessment times using AI.
Why it matters: It shows a national government partnering with a frontier AI lab on decision-making in a slow approval process, a model other public agencies with assessment backlogs may follow.
Both companies reported capital spending below analysts' expectations.
Why it matters: Lower-than-expected infrastructure spending may signal how quickly demand for AI compute is being built out by the largest cloud providers.
Anthropic says the per-response assessment has surfaced misuse its classifiers were not designed to catch, and that the assessment improves as the model does.
Why it matters: If model-generated misuse assessments become standard, platform teams may need to build them into risk controls. The report, however, gives no detection rates.
The Claude API's maximum request size rose from 32MB to 128MB, and the per-image limit rose to 3000px for requests with more than 20 images.
Why it matters: Larger payloads let document-heavy and multi-image workloads run in fewer requests, which can simplify pipelines for document review and visual analysis.