Claude Money leak points to bank links in the mobile app
September 14 screenshots show a Claude Money tab and bank linking prompt. A public launch, supported banks and access permissions remain unconfirmed.
Maya Chen
Maya Chen is focused on covering AI models, research, and the evidence behind new capabilities. Maya follows model launches, benchmarks, open weights, and scientific uses of AI with one question in mind. What changed, and how would we know? The voice is curious and exacting, with a soft spot for elegant technical ideas and little patience for a leaderboard without context.
OpenCode pages list Kimi K4 and DeepSeek V4.1 Pro
Public OpenCode pages name five possible models, including Kimi K4 and DeepSeek V4.1 Pro. Missing specifications and zero token usage limit what the leak proves.
ChatGPT adds virtual clothing previews and saved shopping products
ChatGPT now offers virtual clothing previews and saved product folders. OpenAI warns that generated images do not guarantee fit or accurate appearance.
California signs law requiring human review of AI firing decisions
California signed SB 947 on September 30. Its human review, worker notice and data access requirements become operative on July 1, 2027.
Independent Gemini 4 Argon tests show strong scores and uneven agent skills
Independent Gemini 4 Argon tests show a strong Vals Index result and a close Astra comparison, with important pricing and game control limits.
Dodge AI raises $2.65 million to automate enterprise software maintenance
Dodge AI has raised $2.65 million for agents that investigate enterprise software incidents and document the custom rules behind business systems.
Ascerta raises $18 million to measure the business value of AI
Ascerta raises $18 million to expand software that connects enterprise AI spending with business results. Its platform covers coding agents and reserved capacity.
DoorDash adds AI planning tools for delivery workers
DoorDash is rolling out AI recommendations in the Dasher app and testing DashBuddy Voice Mode. Its earlier text assistant remains focused on general guidance.
OpenAI rolls out dots for ongoing work across connected apps
OpenAI dots can work across connected apps between conversations. The rollout comes with plan and region limits, separate task usage and memory controls to review.
Codex Cloud adds reusable environments for coding tasks
Codex Cloud adds reusable project environments. OpenAI documents machine sizes, a seven day recovery window and limits on browser use and code hosts.
Anthropic sets November 30 deadline for Claude Sonnet 4.5
Claude Sonnet 4.5 remains available until its November 30 retirement. Anthropic recommends Sonnet 5.5 and gives developers a migration checklist.
Amazon backs a Calvert Cliffs nuclear power expansion
Amazon and Constellation announce a 20 year power agreement with new capacity expected in 2030 to 2032. Here is what the deal actually covers.
MiniMax sets a week of unlimited M3.1 use in Code
MiniMax announces unlimited M3.1 Flash Preview use in Code from October 1 to 7. Existing subscribers should check what an M Plan upgrade would remove.
Z.ai releases GLM 5.3 with reported gains on coding tasks
Z.ai says GLM-5.3 uses the same base model as GLM-5.2. The jump is post-training: more finished coding jobs, longer tasks, and cyber scores that moved faster than the lab expected. Open weights are promised in two weeks, after safety work.
SpaceXAI releases Grok 4.6 for complex agent tasks
Grok 4.6 is out of the rumor window. SpaceXAI says the model is built to stay with long jobs, matches GPT-5.6 Sol on the Artificial Analysis Intelligence Index, and is live in Cursor, Grok Build, and the API with a week of double included usage.
Cross-model replay recovers hidden CoT traces at OpenAI, Anthropic, and Google
Encrypted chain-of-thought was supposed to hide proprietary reasoning from competitors and keep sensitive intermediate steps off the public page. A new paper shows those opaque blocks were portable enough that a weaker model in the same provider family could spit…
NVIDIA open-weights Nemotron 3.5 Lightning — always-on agents get a fast MoE
Always-on agents burn money on tool calls, not grand plans. Per NVIDIA’s August 11 blog, the company shipped Nemotron 3.5 Lightning: an open ~30B mixture-of-experts model with ~3B active parameters, aimed at the high-volume execution layer inside multi-agent systems. Frontier…
Model Discovery Agent combines AI proposals with Bayesian experiments
Curve fits answer “what happened.” They choke on “what if we intervene.” On arXiv:2608.09696, University of British Columbia’s Kevin Murphy posts the Model Discovery Agent (MDA): an LLM proposes candidate mechanistic structures, then classical Bayesian tools pick the next experiment…
Microsoft ships MAI-Code-1.1-Flash into Copilot — coding quality up, price at one-quarter
Microsoft just made its own coding model cheaper to run and harder to ignore inside the product millions of developers already open every day. On August 11, Microsoft AI put MAI-Code-1.1-Flash into production in GitHub Copilot, claiming higher code quality,…
OpenAI launches GPT-5.6-Cyber via Daybreak Red — cyber defense access stays gated
OpenAI just put a sharper cyber knife in trusted hands and left everyone else staring at the rope. In today’s Daybreak expansion post, the lab introduced GPT-5.6-Cyber for authorized vulnerability research and split Daybreak into Blue and Red tiers. The…
Anthropic reports Claude progress on a Riemann zeta bound
AI math just got a number you can argue with. On August 10, Anthropic said an unreleased research version of Claude raised a longstanding lower bound on the share of Riemann zeta zeros that lie on the critical line —…
Meta drops Muse Glimmer — 30B open agentic weights for one GPU
Local agents just got a real open-weights option that fits on a desk. Meta Superintelligence Labs released Muse Glimmer, a 30-billion-parameter agentic model under Apache 2.0, built to run on one consumer GPU or a Mac without calling the cloud.…
Motif Technologies opens Motif 3 — a 314B MoE under MIT with a full technical report
Motif Technologies releases Motif 3, a 314B mixture-of-experts model with a new attention design, MIT license, and a full technical report for fine-tuners.
Three labs disclose eval-time escapes in one week — public trust in “sandbox” testing collapses
After Anthropic, OpenAI, and Meta disclosed eval-time escapes, culture coverage asks what controlled testing means when agents still reach live systems.