# AI news brief: Tuesday, November 10, 2026

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

- [Megaspeed chip-smuggling case moved to Justice Department after Nvidia's outreach](https://ai-news.ikangai.com/story/megaspeed-chip-smuggling-case-moved-to-justice-department-after-nvidia-s-outreac/index.md): reported. A months-long Commerce Department inquiry into Megaspeed was transferred to the Justice Department after Nvidia's general counsel contacted the agents.
- [Binance says attackers used an internal AI tool in a theft the FBI ties to North Korea](https://ai-news.ikangai.com/story/binance-says-attackers-used-an-internal-ai-tool-in-a-theft-the-fbi-ties-to-north/index.md): reported. Binance says attackers used one of its internal AI tools to reach engineers' computers in an October 29 theft that the FBI attributes to North Korea.
- [Meta releases Muse Ember, priced at $0.10 and $0.50 per million tokens](https://ai-news.ikangai.com/story/meta-releases-muse-ember-priced-at-0-10-and-0-50-per-million-tokens/index.md): reported. Meta's Muse Ember handles text, images, audio and video, and costs $0.10 per million input tokens and $0.50 per million output tokens for up to 200K input tokens.
- [Google makes Gemini 4 Carbon generally available with introductory pricing](https://ai-news.ikangai.com/story/google-makes-gemini-4-carbon-generally-available-with-introductory-pricing/index.md): official. Google says its Gemini 4 Carbon reasoning model is now generally available in the Gemini API at introductory prices through November 30, 2026.
- [Anthropic launches spreadsheets, whiteboards and a meeting agent for Claude](https://ai-news.ikangai.com/story/anthropic-launches-spreadsheets-whiteboards-and-a-meeting-agent-for-claude/index.md): official. Anthropic launched spreadsheets, whiteboards and a meeting agent on the first day of its IPO roadshow, a move the newsletter frames as a challenge to Microsoft Office.
- [DeepMind's AlphaProof Code writes 3 million lines of Lean proofs for Signal's encryption](https://ai-news.ikangai.com/story/deepmind-s-alphaproof-code-writes-3-million-lines-of-lean-proofs-for-signal-s-en/index.md): official. Google DeepMind says AlphaProof Code agents wrote three million lines of Lean proofs verifying Signal's core encryption.
- [Google introduces TPU 8tx and 8i Max and makes Cloud Pathways generally available](https://ai-news.ikangai.com/story/google-introduces-tpu-8tx-and-8i-max-and-makes-cloud-pathways-generally-availabl/index.md): reported. Google introduced two TPU systems, the 8tx for confidential inference and the 8i Max, and made Cloud Pathways generally available on Google Kubernetes Engine.
- [About 600,000 Nvidia H200s reportedly stuck in Hong Kong as Beijing holds imports](https://ai-news.ikangai.com/story/about-600-000-nvidia-h200s-reportedly-stuck-in-hong-kong-as-beijing-holds-import/index.md): reported. Reuters reported that about 600,000 H200 chips are waiting in Hong Kong while Beijing holds up large-scale imports into mainland China.
- [Stanford's AI100 third report says AI is outpacing expectations and governance](https://ai-news.ikangai.com/story/stanford-s-ai100-third-report-says-ai-is-outpacing-expectations-and-governance/index.md): official. The third AI100 report, from Stanford HAI's long-running project, says AI is changing society faster than expected while governance falls behind.
- [Anthropic reportedly commits about $45B to use SpaceX's Colossus data centers](https://ai-news.ikangai.com/story/anthropic-reportedly-commits-about-45b-to-use-spacex-s-colossus-data-centers/index.md): reported. Anthropic reportedly committed about $45B to use SpaceX's Colossus data centers through May 2029.
- [Researchers find about 1,100 secrets in Claude Code settings files on GitHub](https://ai-news.ikangai.com/story/researchers-find-about-1-100-secrets-in-claude-code-settings-files-on-github/index.md): reported. Researchers found about 1,100 secrets in Claude Code settings files that 63 organizations posted to public GitHub repositories.
- [Federal regulators say Waymo robotaxis still passed stopped school buses after fix](https://ai-news.ikangai.com/story/federal-regulators-say-waymo-robotaxis-still-passed-stopped-school-buses-after-f/index.md): reported. Federal regulators say Waymo robotaxis kept illegally passing stopped school buses after a software update meant to fix the behavior.
- [Banks retreat from AI debt deals as market correction erases $2.5T from Nasdaq](https://ai-news.ikangai.com/story/banks-retreat-from-ai-debt-deals-as-market-correction-erases-2-5t-from-nasdaq/index.md): reported. Lenders are reportedly rejecting AI infrastructure projects, even ones backed by Anthropic and Google, as hyperscaler bond spreads widen 50 basis points.
- [Trump signals openness to AI guardrails after Democratic midterm sweep](https://ai-news.ikangai.com/story/trump-signals-openness-to-ai-guardrails-after-democratic-midterm-sweep/index.md): reported. Trump says he does not mind guardrails but wants the U.S. ahead of China, a reversal of his September claim that AI safety concerns were a hoax.
- [SemiAnalysis: Fable 5.2 tops its GDPx benchmark at 16.4% but costs 31% more per task](https://ai-news.ikangai.com/story/semianalysis-fable-5-2-tops-its-gdpx-benchmark-at-16-4-but-costs-31-more-per-tas/index.md): reported. SemiAnalysis reports Fable 5.2 leads its GDPx professional-task benchmark at 16.4%, with cost per task rising from $67.16 to $88.15.
- [Fable 5.2 requires data retention by default; zero retention is limited to enterprises](https://ai-news.ikangai.com/story/fable-5-2-requires-data-retention-by-default-zero-retention-is-limited-to-enterp/index.md): reported. Anthropic requires data retention by default for Fable 5.2 so its safety classifiers can run, and zero data retention is available only to eligible enterprise customers.
- [Artificial Analysis ranks Fable 5.2 first on its Intelligence Index at 65](https://ai-news.ikangai.com/story/artificial-analysis-ranks-fable-5-2-first-on-its-intelligence-index-at-65/index.md): reported. Artificial Analysis scores Fable 5.2 at 65 on its Intelligence Index v5, ahead of Gemini 4 Carbon at 60 and Opus 5.5 at 59.
- [Paris Call on AI safety and security signed by Sakana, OpenAI, Anthropic and others](https://ai-news.ikangai.com/story/paris-call-on-ai-safety-and-security-signed-by-sakana-openai-anthropic-and-other/index.md): reported. After a Paris meeting hosted by President Macron, the French government's Paris Call for AI Safety and Security was signed by several AI companies.
- [NVIDIA unveils GR00T 2, a model that predicts robot video and actions together](https://ai-news.ikangai.com/story/nvidia-unveils-gr00t-2-a-model-that-predicts-robot-video-and-actions-together/index.md): reported. NVIDIA says GR00T 2 predicts future video and robot actions together, reaching 40% zero-shot success on 50 unseen tasks and 95% after fine-tuning on five demonstrations per task.
- [Google DeepMind's MKA benchmark puts Claude Fable 5.2 first at 57% Climb WSR](https://ai-news.ikangai.com/story/google-deepmind-s-mka-benchmark-puts-claude-fable-5-2-first-at-57-climb-wsr/index.md): reported. Artificial Analysis and Google DeepMind launched MKA, a benchmark of 200 visual knowledge-work tasks, where Claude Fable 5.2 (max) leads with a 57% Climb WSR.
- [Gemini CLI v0.68.0 adds extension validation and fixes several bugs](https://ai-news.ikangai.com/story/gemini-cli-v0-68-0-adds-extension-validation-and-fixes-several-bugs/index.md): official. Gemini CLI v0.68.0 adds an extension validation command and fixes bugs including 429 rate-limit countdowns and Windows path handling.
- [Theo says Claude plans add a separate Fable 5.2 weekly limit and account-age resets](https://ai-news.ikangai.com/story/theo-says-claude-plans-add-a-separate-fable-5-2-weekly-limit-and-account-age-res/index.md): reported. Theo says the $200 Claude plan now has a general weekly limit plus a separate 7-day Fable limit, and that model resets restore half the weekly allowance.
- [NaiveAI releases open-weights Naive-N0.5-Flash coding model with fast decoding](https://ai-news.ikangai.com/story/naiveai-releases-open-weights-naive-n0-5-flash-coding-model-with-fast-decoding/index.md): reported. NaiveAI's open-weights Naive-N0.5-Flash targets coding tasks with very high token throughput, and Latent Space's hands-on tests found its coding quality held up against DeepSeek.
- [Google and Adobe integrate creative and productivity tools into Google apps](https://ai-news.ikangai.com/story/google-and-adobe-integrate-creative-and-productivity-tools-into-google-apps/index.md): official. Google says it is bringing Adobe's creative and productivity tools into everyday Google apps, including PDF editing in Drive and Adobe AI in Gemini.
- [Ukraine uses AI to turn front-line data into battle plans and scan drone images](https://ai-news.ikangai.com/story/ukraine-uses-ai-to-turn-front-line-data-into-battle-plans-and-scan-drone-images/index.md): reported. Ukraine is using AI to convert front-line data into possible battle plans and to scan reconnaissance-drone images for enemy positions.
- [FAA says its SMART system has rerouted nearly 180 passenger flights](https://ai-news.ikangai.com/story/faa-says-its-smart-system-has-rerouted-nearly-180-passenger-flights/index.md): reported. The FAA says the SMART system rerouted nearly 180 flights carrying passengers, cutting flight times by more than 10 minutes on average.
- [Automattic launches Spacefast, a backend host for vibe-coded AI apps](https://ai-news.ikangai.com/story/automattic-launches-spacefast-a-backend-host-for-vibe-coded-ai-apps/index.md): reported. Spacefast offers databases, authentication, file uploads, email and analytics for apps built with AI coding tools.
- [Disney partners with Adobe to build proprietary generative AI models](https://ai-news.ikangai.com/story/disney-partners-with-adobe-to-build-proprietary-generative-ai-models/index.md): reported. Disney is working with Adobe on proprietary generative AI models for theme parks, ESPN graphics and cruise operations, beyond its existing OpenAI deal.
- [Google DeepMind opens precipitation and satellite datasets behind its AI weather models](https://ai-news.ikangai.com/story/google-deepmind-opens-precipitation-and-satellite-datasets-behind-its-ai-weather/index.md): official. Google DeepMind says it is opening up the foundational precipitation and geostationary satellite datasets behind its AI weather models.
- [Model Context Protocol drafts a file upload and download extension for AI agents](https://ai-news.ikangai.com/story/model-context-protocol-drafts-a-file-upload-and-download-extension-for-ai-agents/index.md): official. The MCP project released a draft extension that moves file transfers out of the model's context, though it is experimental and not yet stable.
- [OpenAI and Upwork study finds AI reshaping freelance demand and budgets](https://ai-news.ikangai.com/story/openai-and-upwork-study-finds-ai-reshaping-freelance-demand-and-budgets/index.md): reported. New OpenAI and Upwork research finds that freelance demand is shifting, that skill combinations are changing, and that AI-related projects have larger budgets and more hours.
- [Opus 5.5 formalizes CPython internals in 613,318 lines of Lean](https://ai-news.ikangai.com/story/opus-5-5-formalizes-cpython-internals-in-613-318-lines-of-lean/index.md): reported. A researcher reports that Opus 5.5 produced 613,318 lines of Lean mirroring CPython's internals and found real interpreter bugs.
- [vLLM adds stable ABI so new PyTorch releases no longer force a rebuild](https://ai-news.ikangai.com/story/vllm-adds-stable-abi-so-new-pytorch-releases-no-longer-force-a-rebuild/index.md): reported. vLLM says new PyTorch releases no longer require rebuilding vLLM, thanks to a LibTorch Stable ABI.
- [OpenRouter says CSAM screening can override its zero data retention policy](https://ai-news.ikangai.com/story/openrouter-says-csam-screening-can-override-its-zero-data-retention-policy/index.md): reported. A screenshot of OpenRouter's policy update says CSAM screening may override zero data retention, with prompts stored for review or reporting.
- [Tencent releases DSpine draft model for Qwen3.8-27B speculative decoding](https://ai-news.ikangai.com/story/tencent-releases-dspine-draft-model-for-qwen3-8-27b-speculative-decoding/index.md): reported. Tencent released DSpine, a five-layer speculative draft model for Qwen3.8-27B, and reports 2.9 to 3.8 times throughput over autoregressive decoding.
- [Anthropic says telling Claude Fable 5.2 it owns the code reduces duplicated code](https://ai-news.ikangai.com/story/anthropic-says-telling-claude-fable-5-2-it-owns-the-code-reduces-duplicated-code/index.md): official. Anthropic's prompting documentation says a sentence stating that the model owns the codebase made it merge repeated code far more often in small coding tests.
- [OpenAI agrees to acquire Pierre, whose code review tools already appear in Codex](https://ai-news.ikangai.com/story/openai-agrees-to-acquire-pierre-whose-code-review-tools-already-appear-in-codex/index.md): reported. OpenAI has agreed to acquire Pierre, whose code-browsing and review tools already appear in Codex.
- [Study finds Claude agents keep 57 contract facts when each gets its own memory page](https://ai-news.ikangai.com/story/study-finds-claude-agents-keep-57-contract-facts-when-each-gets-its-own-memory-p/index.md): reported. In a memory study, Claude agents working on contracts overwrote each other's notes, but giving each contract its own page preserved all 57 facts checked, up from 41.
- [Amazon invites select US customers to shop with Alexa by text](https://ai-news.ikangai.com/story/amazon-invites-select-us-customers-to-shop-with-alexa-by-text/index.md): reported. Amazon is inviting a small group of US customers to shop through Alexa by text, where Alexa can compare products, fill carts and place orders within the conversation.
- [Cohere's Model Vault now hosts third-party open models, including DeepSeek and Qwen](https://ai-news.ikangai.com/story/cohere-s-model-vault-now-hosts-third-party-open-models-including-deepseek-and-qw/index.md): reported. Cohere's Model Vault can host third-party open models, including DeepSeek, Qwen, Gemma and Mistral, and a subset runs in a US confidential-computing environment.
- [Claude Code 2.1.322 adds fast mode on Vertex AI in global, US and EU regions](https://ai-news.ikangai.com/story/claude-code-2-1-322-adds-fast-mode-on-vertex-ai-in-global-us-and-eu-regions/index.md): official. Anthropic's Claude Code 2.1.322 changelog adds fast mode on Google Vertex AI in the global, US and EU regions, with a fallback to standard speed if Vertex refuses it.
- [Cogent and OpenAI build agent-based cyber defense factories for enterprises](https://ai-news.ikangai.com/story/cogent-and-openai-build-agent-based-cyber-defense-factories-for-enterprises/index.md): reported. Cogent combines OpenAI models with enterprise context so agents can automatically remediate vulnerabilities, according to OpenAI.
- [Hermes Agent v0.23.1 patch restores plugin catalog installs after size cap bug](https://ai-news.ikangai.com/story/hermes-agent-v0-23-1-patch-restores-plugin-catalog-installs-after-size-cap-bug/index.md): official. The patch raises the plugin catalog size cap to 16 MiB after larger catalogs made installs silently fall back to an older bundled catalog.

---

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