OpenAI S-1, Apple Siri AI, Intel 3M Chips, Xiaomi 1T tok/s

Meta, OpenAI Sol, Tencent, Google SensorFM

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Meta, OpenAI Sol, Tencent, Google SensorFM META, OPENAI SOL, TENCENT, GOOGLE SENSORFM Today’s episode follows AI becoming a set of control surfaces: product rollbacks, reasoning throttles, self-improvement workflows, inference economics, geopolitical agent ownership, and boring enterprise plumbing. STORIES * Meta pulls new AI image feature after days of backlash [https://www.bbc.com/news/articles/c2dy6e8klw0o] — consumer AI safety now includes rollback speed, not just reassuring policy language. * OpenAI's GPT-5.6 Sol autonomously post-trained Luna [https://the-decoder.com/openais-gpt-5-6-sol-autonomously-post-trained-the-smaller-luna-model-with-a-fairly-underspecified-prompt] — model development starts to look like supervised automation with benchmarks. * Superhuman competitive programming AI is here [https://news.smol.ai/issues/26-07-10-not-much] — an OpenAI model reportedly dominated an AtCoder exhibition, narrowing another algorithmic coding frontier. * GPT-5.6 Sol reasoning levels [https://the-decoder.com/openai-staffer-maps-out-which-of-gpt-5-6-sols-five-reasoning-levels-fits-which-task-complexity] — intelligence becomes a cost and policy throttle, from Light to multi-agent Ultra modes. * GLM-5.2 on a 25GB-RAM consumer machine [https://news.smol.ai/issues/26-07-10-not-much] — disk-backed expert paging reframes huge open MoE models as memory-hierarchy problems. * Unsloth Qwen3.6 NVFP4 quantization [https://news.smol.ai/issues/26-07-10-not-much] — faster inference economics are being fought in tensor formats, kernels, and memory movement. * Tencent moves to buy majority stake in Manus [https://the-decoder.com/tencent-moves-to-buy-majority-stake-in-manus-after-beijing-forced-meta-to-unwind-its-2-billion-deal] — AI-agent ownership becomes a geopolitical routing decision after Beijing blocked Meta’s deal. * OpenAI kills Atlas and folds it into ChatGPT [https://the-decoder.com/openai-kills-its-atlas-browser-after-just-eight-months-and-folds-everything-into-chatgpt] — agent browsers may become features before they become lasting standalone businesses. * The Fed asks Marc Andreessen about AI and inflation [https://the-decoder.com/the-fed-wants-ai-investor-marc-andreessen-to-help-figure-out-if-ai-can-tame-inflation] — a real macroeconomic question arrives with obvious conflict-of-interest fumes. * Google Research introduces SensorFM [https://www.marktechpost.com/2026/07/10/google-research-introduces-sensorfm-a-wearable-health-foundation-model-pretrained-on-one-trillion-minutes-of-sensor-data] — foundation models move into wearable telemetry and bodily signal representations.

GPT-5.6, Copilot, Meta Muse, China UN

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] * OpenAI GPT-5.6 family: Luna, Terra, Sol [https://simonwillison.net/2026/Jul/9/gpt-5-6] * ChatGPT Work [https://openai.com/index/chatgpt-for-your-most-ambitious-work] * GPT-5.6 in Microsoft 365 Copilot [https://openai.com/index/gpt-5-6-preferred-model-microsoft-365-copilot] * Normal Technology: AI up the stack and enterprise lock-in [https://www.normaltech.ai/p/up-the-stack-how-ais-escape-from] * Meta Muse Spark 1.1 and Model API [https://simonwillison.net/2026/Jul/9/muse-spark-1-1] * Bun rewrite in Rust [https://simonwillison.net/2026/Jul/8/rewriting-bun-in-rust] * Kenton Varda on AI-written change descriptions [https://simonwillison.net/2026/Jul/8/kenton-varda] * Datalab Lift schema-first document extraction [https://www.marktechpost.com/2026/07/09/datalab-lift-vs-the-field-how-a-9b-schema-first-extractor-compares-with-nuextract3-llamaextract-marker-and-docling] * CausalDS: Benchmarking Causal Reasoning in Data-Science Agents [https://huggingface.co/papers/2607.08093] * IdeaGene-Bench: Scientific lineage reasoning [https://huggingface.co/papers/2607.08758] * China at the UN Global Dialogue on AI Governance [https://news.smol.ai/issues/26-07-09-not-much]

Grok, Fable, GPT-Live, Meta: AI Becomes Plumbing

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Grok, Fable, GPT-Live, Meta: AI Becomes Plumbing GROK, FABLE, GPT-LIVE, META: AI BECOMES PLUMBING This episode follows a day where intelligence looks less like one grand oracle and more like routing, orchestration, voice interfaces, robots, sensors, release governance, and agent-data infrastructure. Dreary, yes. Also probably correct. STORIES COVERED * Grok 4.5 is so cheap compared to Fable 5 and GPT-5.5 that benchmark gaps may not matter much [https://the-decoder.com/grok-4-5-is-so-cheap-compared-to-fable-5-and-gpt-5-5-that-benchmark-gaps-may-not-matter-much] — price pressure turns model quality into an economic question. * OpenAI introduces GPT-Live [https://openai.com/index/introducing-gpt-live] — full-duplex voice makes agents listen and respond while deeper reasoning runs behind the curtain. * Mistral enters robotics with Robostral Navigate [https://the-decoder.com/mistral-enters-robotics-with-robostral-navigate-an-8b-model-that-steers-robots-using-just-one-camera] — an 8B model attempts robot navigation from a single camera. * Anthropic turns Fable 5 into a manager [https://the-decoder.com/anthropics-fix-for-fable-5s-high-cost-is-turning-it-into-a-manager-that-delegates-to-sonnet-5] — the expensive model plans, cheaper Sonnet 5 executes. * Claude Fable 5 dominates industry benchmarks at a premium [https://the-decoder.com/anthropics-claude-fable-5-dominates-new-industry-benchmarks-at-a-steep-premium] — strong scores, but with cost that changes the buying decision. * MiniMax plans a 2.7T-parameter open-source model [https://the-decoder.com/chinese-ai-startup-minimax-plans-to-open-source-a-2-7-trillion-parameter-model-later-this-year] — open weights keep scaling into strategically awkward territory. * Meta tests always-on AI glasses [https://the-decoder.com/meta-tests-always-on-ai-glasses-that-capture-your-entire-day] — memory assistance begins looking like ambient surveillance infrastructure. * Meta’s Muse Image raises Instagram photo consent questions [https://the-decoder.com/muse-image-is-technically-impressive-but-metas-use-of-instagram-photos-raises-questions] — agentic image generation meets likeness and opt-out concerns. * OpenAI’s GPT-5.6 launches after a government-forced delay [https://the-decoder.com/openais-gpt-5-6-launches-thursday-after-a-delay-forced-by-the-u-s-government] — frontier release timing becomes a safety-governance problem. * Hugging Face and NVIDIA: Data for Agents [https://huggingface.co/blog/nvidia/open-data-for-agents] — agents need structured data, tools, traces, and evaluation plumbing, not just bigger chat models.

Anthropic, China, Copilot, Gemini Agents

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Anthropic, China, Copilot, Gemini Agents ANTHROPIC, CHINA, COPILOT, GEMINI AGENTS * Claude's hidden inner monologue is now readable thanks to Anthropic's new Jacobian Lens [https://the-decoder.com/claudes-hidden-inner-monologue-is-now-readable-thanks-to-anthropics-new-jacobian-lens] * Copilot goes cheap as Microsoft phases out OpenAI and Anthropic models to cut costs [https://the-decoder.com/copilot-goes-cheap-as-microsoft-phases-out-openai-and-anthropic-models-to-cut-costs] * China eyes export curbs on its top AI models, and Europe is caught in the middle [https://the-decoder.com/china-eyes-export-curbs-on-its-top-ai-models-and-europe-is-caught-in-the-middle] * OpenAI and Anthropic are giving away millions in computing power to attract startups [https://the-decoder.com/openai-and-anthropic-are-giving-away-millions-in-computing-power-to-attract-startups] * Apollo economist warns AI profit gains outside tech could take "well beyond" what Wall Street expects [https://the-decoder.com/apollo-economist-warns-ai-profit-gains-outside-tech-could-take-well-beyond-what-wall-street-expects] * Anthropic's Claude Cowork AI agent is now available on mobile and web [https://the-decoder.com/anthropics-claude-cowork-ai-agent-is-now-available-on-mobile-and-web] * Expanding Managed Agents in Gemini API: background tasks, remote MCP and more [https://blog.google/innovation-and-ai/technology/developers-tools/expanding-managed-agents-gemini-api] * OpenAI Releases GPT-Realtime-2.1 and GPT-Realtime-2.1-mini for Low-Latency Voice Agents in the API [https://www.marktechpost.com/2026/07/06/openai-gpt-realtime-2-1-mini-reasoning-realtime-api] * Nemotron-Labs-Diffusion: A Tri-Mode Language Model Unifying Autoregressive, Diffusion, and Self-Speculation Decoding [https://huggingface.co/papers/2607.05722] * AlayaWorld: Long-Horizon and Playable Video World Generation [https://huggingface.co/papers/2607.06291]

Cloudflare, JADEPUFFER, Nvidia, LeRobot: AI Meets Its Plumbing

Send us Fan Mail [https://www.buzzsprout.com/2614078/fan_mail/new] Today’s episode: AI is moving from model-launch theater into the operating environment around it — permissions, tools, attackers, hardware bottlenecks, provenance, and physical-world constraints. Dreary, but at least the dashboards are cheerful enough for all of us. * Cloudflare splits AI bot controls [https://the-decoder.com/cloudflare-replaces-its-blanket-ai-bot-block-with-granular-controls-for-search-training-and-agent-crawlers] into search, training, and agent crawler categories, turning web access into a more explicit policy layer. * Sysdig describes JADEPUFFER [https://the-decoder.com/jadepuffer-is-the-first-agentic-ransomware-operation-and-it-exposes-old-security-sins-at-machine-speed] as agentic ransomware, showing how old credential, permission, and backup failures become worse at machine speed. * Epoch Capabilities Index coverage [https://the-decoder.com/gpt-4s-dominance-lasted-a-year-while-todays-top-models-barely-survive-seven-weeks-at-the-top] suggests frontier-model leadership now changes hands rapidly while capability margins shrink. * Tencent Hy3 [https://simonwillison.net/2026/Jul/6/hy3] adds a large Apache-licensed mixture-of-experts model to the Chinese open-model race. * Nvidia’s reported Kyber NVL144 delay [https://the-decoder.com/nvidias-kyber-nvl144-reportedly-pushed-back-more-than-a-year-asian-suppliers-drop] exposes circuit boards, racks, and manufacturing as hard AI scaling constraints. * Import AI on Fable writing GPU kernels [https://jack-clark.net/2026/07/06/import-ai-464-fables-writes-gpu-kernels-ai-automation-and-analog-computation] and Hugging Face Kernels updates [https://huggingface.co/blog/revamped-kernels] show that low-level optimization is becoming part of the AI automation loop. * China pushes ByteDance and Alibaba to shut down humanlike AI companion personas [https://the-decoder.com/china-forces-its-biggest-ai-platforms-to-shut-down-humanlike-chatbot-personas], making simulated intimacy a direct regulatory object. * Amazon closes Mechanical Turk to new customers [https://the-decoder.com/amazon-sunsets-mechanical-turk-the-original-artificial-artificial-intelligence], marking a quiet shift in the human labor layer behind automation. * Sean Goedecke on C2PA [https://seangoedecke.com/c2pa-only-works-if-everything-is-signed] argues provenance only works if signing becomes nearly universal. * LeRobot v0.6.0 [https://huggingface.co/blog/lerobot-release-v060], ChinaTalk on Unitree [https://www.chinatalk.media/p/the-robots-are-here], dense spatial perception research [https://huggingface.co/papers/2607.05247], Zhipu ZCode [https://the-decoder.com/zhipu-ai-launches-zcode-to-challenge-claude-code-and-openai-codex-at-a-fraction-of-the-cost], and small models in unreliable-network settings [https://spectrum.ieee.org/small-language-models-ai-pharmaceuticals] all point to AI value moving into deployment shape: local, cheap, evaluated, and physically competent enough not to fall over immediately.