# AI news brief: Friday, November 6, 2026

38 stories, most important first. Status is how solid each one is: official, confirmed, reported or rumor.

- [METR says AI speeds up Anthropic's R&D, AECI growing 1.9x faster since Oct 2025](https://ai-news.ikangai.com/story/metr-says-ai-speeds-up-anthropic-s-r-and-d-aeci-growing-1-9x-faster-since-oct-20/index.md): reported. METR's redacted risk report estimates that AECI scores for Anthropic models have grown 1.9x faster since October 2025, with a 95% interval of 1.5 to 2.8x.
- [Midterm results put AI regulation and data center limits on the agenda](https://ai-news.ikangai.com/story/midterm-results-put-ai-regulation-and-data-center-limits-on-the-agenda/index.md): reported. Democratic control of Congress signals a shift away from hands-off AI regulation, and incoming lawmakers promise hearings and an AI development pause.
- [Thinking Machines releases Inkling 2 Small; open weights promised next week](https://ai-news.ikangai.com/story/thinking-machines-releases-inkling-2-small-open-weights-promised-next-week/index.md): reported. Thinking Machines has launched Inkling 2 Small, a 594B-total, 17B-active multimodal MoE model available on Tinker, with Apache 2.0 weights promised after safety testing.
- [OpenAI reportedly building a ChatGPT tool that lets businesses track their mentions](https://ai-news.ikangai.com/story/openai-reportedly-building-a-chatgpt-tool-that-lets-businesses-track-their-menti/index.md): official. A tool reportedly spotted would let businesses claim a profile and count how often ChatGPT mentions them, without showing the prompts behind those mentions.
- [Anthropic adds personal finance to Claude for paid US subscribers](https://ai-news.ikangai.com/story/anthropic-adds-personal-finance-to-claude-for-paid-us-subscribers/index.md): reported. Paid US subscribers can link bank accounts through Plaid and ask Claude about balances, spending and bills in plain language.
- [Moonshot AI reportedly sought Nvidia Blackwell GPUs via European data centers](https://ai-news.ikangai.com/story/moonshot-ai-reportedly-sought-nvidia-blackwell-gpus-via-european-data-centers/index.md): rumor. Sources told CNBC that Moonshot AI tried to lease access to Nvidia Blackwell clusters in Europe after U.S. officials raised concerns about remote access.
- [Google releases Gemini Omni 1.2 Flash with cheaper, longer video clips](https://ai-news.ikangai.com/story/google-releases-gemini-omni-1-2-flash-with-cheaper-longer-video-clips/index.md): official. Gemini Omni 1.2 Flash is generally available in the Gemini API, with 3 to 15 second clips, conversational editing and a 30% lower price.
- [Prospective Anthropic IPO buyers say the $1.5 trillion target must come down](https://ai-news.ikangai.com/story/prospective-anthropic-ipo-buyers-say-the-1-5-trillion-target-must-come-down/index.md): rumor. Three prospective buyers say Anthropic's $1.5 trillion IPO target has to come down, and the filing still does not show how fast revenue grew in October.
- [Nasdaq 100 falls over 6% as doubts grow over returns on AI investment](https://ai-news.ikangai.com/story/nasdaq-100-falls-over-6-as-doubts-grow-over-returns-on-ai-investment/index.md): reported. The Nasdaq 100 dropped more than 6% in one day as investors questioned whether AI spending will deliver expected returns.
- [OpenAI pledges to cover data center power costs and publish energy and water data](https://ai-news.ikangai.com/story/openai-pledges-to-cover-data-center-power-costs-and-publish-energy-and-water-dat/index.md): reported. OpenAI has made new pledges on data center infrastructure and power costs, including publishing energy and water data and skipping NDAs with officials.
- [California AG Bonta subpoenas OpenAI over cybersecurity incidents](https://ai-news.ikangai.com/story/california-ag-bonta-subpoenas-openai-over-cybersecurity-incidents/index.md): reported. California Attorney General Bonta has subpoenaed OpenAI over cybersecurity incidents, a move the newsletter links to rising state enforcement against frontier labs.
- [Brockman says OpenAI could be IPO-ready in 2027](https://ai-news.ikangai.com/story/brockman-says-openai-could-be-ipo-ready-in-2027/index.md): reported. OpenAI president Greg Brockman says the company could be ready for an IPO in 2027, citing enterprise growth and business confidence.
- [opencode v2.2.6 adds shared MCP connections and Azure support for Claude models](https://ai-news.ikangai.com/story/opencode-v2-2-6-adds-shared-mcp-connections-and-azure-support-for-claude-models/index.md): official. opencode v2.2.6 adds opt-in MCP connection pooling, a worktree option for scripted runs and Azure Messages support for Foundry Claude models.
- [Anthropic details layered AI code review process for Claude Code, as relayed by Theo](https://ai-news.ikangai.com/story/anthropic-details-layered-ai-code-review-process-for-claude-code-as-relayed-by-t/index.md): reported. Anthropic's engineering article describes AI review on every pull request, risk-based human approval and staged rollouts with fast rollback.
- [OpenAI built an internal code repository to replace Git for agent-driven commits](https://ai-news.ikangai.com/story/openai-built-an-internal-code-repository-to-replace-git-for-agent-driven-commits/index.md): reported. OpenAI's CTO of Applications, V. J. Raji, says AI agents now push code faster than Git was designed to handle, so OpenAI built its own system.
- [OpenAI's CTO says ChatGPT has 1.2 billion users and voice is key to the next billion](https://ai-news.ikangai.com/story/openai-s-cto-says-chatgpt-has-1-2-billion-users-and-voice-is-key-to-the-next-bil/index.md): reported. V. J. Raji says ChatGPT has 1.2 billion users and that voice, not typing, is the most likely way to reach the next billion people.
- [OpenAI's ads team reportedly reached $1B annual revenue within about a year](https://ai-news.ikangai.com/story/openai-s-ads-team-reportedly-reached-1b-annual-revenue-within-about-a-year/index.md): reported. According to V. J. Raji, OpenAI's advertising team of about 50 people went from zero code to a $1 billion annual revenue run rate in roughly seven to eight months.
- [Aina opens $399 preorders for AI ring with screen, joystick and mic](https://ai-news.ikangai.com/story/aina-opens-399-preorders-for-ai-ring-with-screen-joystick-and-mic/index.md): reported. Aina has opened preorders for an AI ring with a screen, joystick and microphone at $399, with shipping in January 2027 at $499 standard.
- [Meta rolls out video chat with Muse AI avatar to US users](https://ai-news.ikangai.com/story/meta-rolls-out-video-chat-with-muse-ai-avatar-to-us-users/index.md): reported. US users can now voice and video call an animated AI character that responds with gestures and expressions.
- [Dropbox to acquire Tigris Data for $350M to back AI storage service](https://ai-news.ikangai.com/story/dropbox-to-acquire-tigris-data-for-350m-to-back-ai-storage-service/index.md): reported. Dropbox is acquiring Tigris Data for $350M, and the object storage platform will underpin DBX Agent Cloud, which launches in 2027.
- [Hugging Face publishes free ~70k-word handbook on Mixture-of-Experts training](https://ai-news.ikangai.com/story/hugging-face-publishes-free-70k-word-handbook-on-mixture-of-experts-training/index.md): reported. Hugging Face released a free handbook on training mixture-of-experts models, citing about 800k GPU-hours of experiments behind the guidance.
- [Ant Ling releases Ling-3.2-flash multimodal model with 560B total parameters](https://ai-news.ikangai.com/story/ant-ling-releases-ling-3-2-flash-multimodal-model-with-560b-total-parameters/index.md): reported. Ant Ling's Ling-3.2-flash has 560B total and 25B active parameters and, the company says, beats Opus 5 on two benchmarks.
- [Marin finishes main pretraining of 535B Mochi model on 18T tokens](https://ai-news.ikangai.com/story/marin-finishes-main-pretraining-of-535b-mochi-model-on-18t-tokens/index.md): reported. Marin says its Mochi 535B-A23B model finished main pretraining on 18T tokens, with long-context training and fine-tuning next.
- [Z.ai's Jie Tang outlines training data pipeline for upcoming GLM-5.5 models](https://ai-news.ikangai.com/story/z-ai-s-jie-tang-outlines-training-data-pipeline-for-upcoming-glm-5-5-models/index.md): reported. Jie Tang described the pretraining pipeline for GLM-5.5 variants labeled 32B, 122B and 256B, according to a post relaying a Tsinghua lecture.
- [Report: Nvidia RTX Spark firmware tools limit GPU speed on native Linux](https://ai-news.ikangai.com/story/report-nvidia-rtx-spark-firmware-tools-limit-gpu-speed-on-native-linux/index.md): reported. A MindStudio report says RTX Spark laptops are capped on native Linux without a Windows-only firmware component, cutting GPT-OSS-120B speed from about 65 to 14 tokens per second.
- [FLUX 3 Image Klein open-weights model ranks #2 among open image models](https://ai-news.ikangai.com/story/flux-3-image-klein-open-weights-model-ranks-2-among-open-image-models/index.md): reported. The 5.6B open-weights FLUX 3 Image Klein ranks second among open models on Artificial Analysis's AA-Image-T2I, a short distance behind Qwen-Image-2.1.
- [DeepSeek took 53% of OpenRouter's 27T daily tokens, a post reports](https://ai-news.ikangai.com/story/deepseek-took-53-of-openrouter-s-27t-daily-tokens-a-post-reports/index.md): reported. A post reports DeepSeek accounted for 53% of all tokens on OpenRouter in one day, with open-weight labs together above 75%.
- [Viggle Remake open weights released for recasting and editing whole videos](https://ai-news.ikangai.com/story/viggle-remake-open-weights-released-for-recasting-and-editing-whole-videos/index.md): reported. Viggle released open weights on Hugging Face for its Remake model, which recasts people or edits existing videos while keeping source motion and soundtrack.
- [OpenAI partners with Hacktron to find vulnerabilities in its systems](https://ai-news.ikangai.com/story/openai-partners-with-hacktron-to-find-vulnerabilities-in-its-systems/index.md): reported. OpenAI says it is working with Hacktron to find vulnerabilities, test its third-party service connections and strengthen its defenses using frontier AI.
- [Codex CLI 0.167.0 adds agent command center, backend switching and Vertex AI setup](https://ai-news.ikangai.com/story/codex-cli-0-167-0-adds-agent-command-center-backend-switching-and-vertex-ai-setu/index.md): official. The 0.167.0 release of OpenAI's Codex CLI adds unread-result markers, switching between connected backends and an opt-in Google Cloud Vertex AI setup wizard.
- [OpenAI introduces BioBasicsBench, a 100-task benchmark for AI agents in wet-lab science](https://ai-news.ikangai.com/story/openai-introduces-biobasicsbench-a-100-task-benchmark-for-ai-agents-in-wet-lab-s/index.md): reported. BioBasicsBench is a 100-task benchmark that tests how well AI agents interpret raw experimental data and handle routine wet-lab science.
- [Anthropic details 'dreaming' method for organizing Claude's agent memory](https://ai-news.ikangai.com/story/anthropic-details-dreaming-method-for-organizing-claude-s-agent-memory/index.md): official. Anthropic says offline review of agent memory notes can fix errors, build wikis and promote key notes into future sessions, based on its own tests.
- [Google deprecates gemini-embedding-2-preview, routing requests to gemini-embedding-2](https://ai-news.ikangai.com/story/google-deprecates-gemini-embedding-2-preview-routing-requests-to-gemini-embeddin/index.md): official. Google is retiring the gemini-embedding-2-preview model, and from November 6, 2026, its requests route to gemini-embedding-2 at the same rates.
- [Claude Code 2.1.319 adds background MCP tasks and a deny-instead-of-prompt setting](https://ai-news.ikangai.com/story/claude-code-2-1-319-adds-background-mcp-tasks-and-a-deny-instead-of-prompt-setti/index.md): official. Anthropic's 2.1.319 release lets Claude Code run MCP server tasks in the background and can automatically deny permission prompts that would otherwise pause a turn.
- [Claude Code auto mode now refuses or asks when its server-side check is unreachable](https://ai-news.ikangai.com/story/claude-code-auto-mode-now-refuses-or-asks-when-its-server-side-check-is-unreacha/index.md): official. Claude Code 2.1.319 removes the fallback to the built-in classifier when telemetry or feature flags are off, including behind LLM gateways.
- [Claude Code 2.1.319 fixes a Bash permission bypass and a worktree sandbox escape](https://ai-news.ikangai.com/story/claude-code-2-1-319-fixes-a-bash-permission-bypass-and-a-worktree-sandbox-escape/index.md): official. The 2.1.319 changelog lists fixes for a Bash permission bypass using a crafted environment prefix and a sandbox escape when creating a worktree.
- [Anthropic's Claude apps gateway gets pooled budget rules, login notices and JSON logs](https://ai-news.ikangai.com/story/anthropic-s-claude-apps-gateway-gets-pooled-budget-rules-login-notices-and-json/index.md): official. Claude Code 2.1.319 adds admin controls for pooled group budgets, an operator login message and JSON log output to the Claude apps gateway.
- [Artificial Analysis launches VIALS-AA; Gemini 4 Carbon (high) leads at 64%](https://ai-news.ikangai.com/story/artificial-analysis-launches-vials-aa-gemini-4-carbon-high-leads-at-64/index.md): official. Artificial Analysis launched VIALS-AA, a benchmark for scientific image reasoning, where Gemini 4 Carbon (high) leads with 64% of tasks answered correctly in all three repeats.

---

AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it.
