Apple began developer testing of Qwen inside Apple Intelligence in China this week, ahead of a planned December rollout with iOS 27.3, bringing Qwen into Siri, Writing Tools and Visual Intelligence on an opt-in basis. Tencent's November 4 Hy4 release adds visual reasoning and computer use, while Baidu, Moonshot, Ant and Meituan also published or launched new models.
Every held its first conference, Thesis: 2027, on Thursday at Pioneer Works in Brooklyn, drawing 400 attendees in person and thousands more online across 25 sessions. The event centered on Dan Shipper's question of what great human work looks like after automation, with talks spanning frontier labs, tech companies, startups, schools and art studios. Paid subscribers can watch every session on Every TV, and recordings will roll out on YouTube over the coming weeks.
Chang Che argues that AI has shifted cybersecurity from an era of zero-day scarcity to zero-day abundance, and that China's top-down system is poorly suited to the fast, decentralized patching this requires. One week after the PHP zero-day was disclosed, the US had patched over a quarter of the 14 million exposed services while China had patched 4%, per Censys data. Che traces this to China's 2021 law barring researchers from publicizing or selling zero-days abroad, which cut the incentives behind Western bug bounty ecosystems, and to Nicole Perlroth's view that AI models make hoarded exploits obsolete, spelling the death of NOBUS.
Exponential View's Sunday edition #609 argues tech was effectively on the ballot this week, citing Texas Governor-Elect Gina Hinojosa's claim that "99.9% of us" are pitted against "15 billionaire tech guys who are controlling government." It also questions whether AI's productivity payoff has arrived, noting US productivity grew 3.6% last quarter but calling the data too noisy to attribute to AI, and argues Europe is financing the tech race it is losing.
SemiAnalysis toured Envision's Galaxy campus in Ulanqab, Inner Mongolia, planned for up to 800MW of IT capacity, with its first 120MW phase delivered within a year at about a fifth of US costs. The campus runs a live 5MW 800V DC pilot using solid-state transformers that convert 10kV AC straight to 800V DC for racks, and drops the central chiller by sharing one water loop between liquid- and air-cooled servers. Envision supplies over 60% of the first two phases' power from its own 420MW wind farms plus 4-hour battery storage, and the full campus is expected to be completed by the end of 2027.
Artificial Analysis released Decision Index v2, with 45% of the Index now built on private datasets, up from 25% in v1. Drex Max v1.1 from Nace AI narrowly leads decision models at 73, within 1 point of GLiDE v1.2 and Mammoth 27B. The new Decision Score rewards both correctness and calibrated confidence, while AA-Paperclip and AA-Symmetry add private document and consistency tests.
The bottleneck to AI-driven progress may not be wet-lab or compute limits but human dignity, as AI researchers and entrepreneurs resist a world where machine intelligence makes their contributions superfluous. The essay argues their value function — contribution relative to the most competitive replacement — would push them to handicap AI, keeping humans in the loop on non-verifiable questions like resource distribution and ethics.
GitHub published a developer's guide to agentic engineering, walking through how Copilot supports the Plan, Build, Verify, and Observe stages of software work. The guide cites GitHub's 1.2 billion commits in one week this October, over 400% more than a year earlier, and a Faros AI study finding pull requests spent 441% longer in review as teams moved from low to high AI adoption. It details features including Copilot in Slack or Teams, Plan Mode, GitHub Agentic Workflows, cloud sandboxes, session continuity, shared memory, Copilot Dreaming, and Copilot Fusion.
The Pragmatic Summit is returning as a one-day event hosted on 10 February in San Francisco, almost exactly one year after the first Summit. The first few speakers are confirmed, with familiar faces from the newsletter and podcast.
Cerebral Valley AI Summit has added five speakers for next week's event, including Perplexity CEO Aravind Srinivas, Vercel CEO Guillermo Rauch, Applied Intuition CEO Qasar Younis, Flapping Airplanes CEO Benjamin Spector and General Catalyst CEO Hemant Taneja. They join a lineup featuring Databricks' Ali Ghodsi and Palo Alto Networks' Nikesh Arora, with Bessemer's Byron Deeter and Talia Goldberg revealing the AI 100 onstage. Tickets run from $199 founder passes to $2,999, and the summit is nearly sold out.
Meta introduced spaces in its Muse app, an interactive way for users to build a more personalized Muse by describing an idea such as a gift tracker, daily word game or DIY planning board. Users can adjust each space with Muse, place it in the sidebar or main tabs, hide or schedule it, and connect apps so details like concert dates and price changes update in the background. Spaces are available everywhere Muse is available starting today, and Meta says it is exploring sharing options and rolling out specialized model versions trained with learnings from Muse Wonder.
Asaf Ezra argues that AI has made exploits cheap and fast while patching still takes weeks to years, so the gap between exploit and patch will not close by patching harder. He cites the October PHP bugs that were exploited four days before disclosure and an Anthropic report of an unreleased model producing a Linux kernel exploit in under an hour, then ranks four off-device defenses by reach: removing exposure from the open internet, segmentation, network-level virtual patching and eliminating bug classes. He also argues backups are now a primary target, citing Sophos data that 94% of ransomware victims said attackers tried to compromise their backups and 57% of those attempts succeeded.
Ed Zitron argues that US midterm election results have exposed a vulnerability in the AI bubble: voters in multiple states and municipalities backed restrictions on data center construction. He cites 33 ballot measures to restrict data centers, 21 Ohio municipalities voting for restrictions, and 12 states where the winning or leading gubernatorial candidate called for a pause on data center development. He contends that NVIDIA's growth and the growth stories of AI labs depend on an unimpeded data center buildout, and that the industry assumed it could expand without the consent of the communities affected.
Democrats' midterm sweep is being read as a wake-up call for Silicon Valley, with AI infrastructure and chip stocks sliding as investors price in slower data center buildouts. Texas governor-elect Gina Hinojosa made data center opposition and Elon Musk central to her campaign, and new comptroller Sarah Eckhardt will review JETI subsidy applications. Tesla and SpaceX shares fell 11% and 9% on Thursday.
Apple has become the only Mag7 company to beat a major index year-to-date, as its capex stays below pre-AI-spree levels while the four main hyperscalers raised spending roughly 6x-9x since 2023. Roblox said over 80% of its top 100 games by engagement used AI tools in the past month, with daily asset uploads tripling month-over-month.
Idealab founder and chairman Bill Gross joined Exponential View to discuss building businesses in the AI age and his newly founded Idealab.ai, which uses AI to explore potential business ideas, evaluate them against real-world problems, and help founders pressure-test and refine them. Gross stressed iteration and speed, joking he should have named the company "Iterate Lab."
Ai2 released ToyBlocks, an open simulation framework built around a robot arm and blocks on a tabletop, as the first of its FoundationWorlds for controlled robot learning and evaluation. The framework auto-generates skills and demonstrations via RL and motion planning in Isaac Lab, reaching up to 24x real-time on a single GPU, and ships with 85,336 long-horizon demonstrations covering nearly 1,600 hours and 144 million frames. In its experiments, ManiFlow's atomic-skill success was 80.0% in simulation and 81.1% on a real Franka FR3, while the Molmo 2 plus ManiFlow system succeeded in 55.5% of simulated and 50.0% of real-world episodes.
Markets fell $2.5 trillion yesterday, and a Bank of England study found AI is again failing to deliver the expected productivity gains. Gary Marcus frames the drop and the study as parts of the same bad news.
After Democrats won the midterms, over $2.5 trillion was wiped off the Nasdaq as investors fled the AI trade, and US voters passed 43 of 44 local measures to slow data centers down. The Lanarkshire deal, where DataVita pays up to $100,000 per megawatt per year into a community fund, shows how grid-queue leverage can be turned into local benefits. Oracle's New Mexico package worked out to a one-off $84,000 per megawatt, nearly 18 times stingier.
Zyphra released MAIA in beta, a proactive multiplayer AI agent that remembers shared team context, takes action across connected tools, and runs as a standalone Mac desktop app and inside Slack, free during beta. MAIA keeps persistent memory of people, decisions and preferences with sources, executes multi-step tasks in plain language, and can schedule recurring work. It follows existing tool permissions, confirms before irreversible actions, and lets users view and delete memories, with private conversations kept private unless shared.
MIT Technology Review tested Tau Robotics' humanoid cleaning service in San Francisco, where robots are remotely piloted by humans rather than fully autonomous. The robots collect training data for future autonomous models, with pricing at $60 per weekly cleaning with two robots or $100 for a one-time visit. The robot cleaned a kitchen and toilet but struggled with stairs and repeatedly stalled during the first visit.
Ai2 released ToyBlocks, an open simulation framework built around a robot arm and blocks on a tabletop, as the first of its FoundationWorlds for controlled robot-learning research. ToyBlocks auto-generates tasks and demonstrations via RL and motion planning in Isaac Lab, reaching up to 24 times real-time data generation on a single GPU, and ships with 85,336 long-horizon demonstrations covering nearly 1,600 hours and 144 million frames. In real-world tests on a Franka FR3, ManiFlow's atomic-skill success was 80.0% in simulation and 81.1% on hardware, while the Molmo 2 plus ManiFlow system succeeded in 55.5% of simulated and 50.0% of real episodes.
US voters punished Republicans over data centers and AI accelerationism, with bans or restrictions passed in at least 38 jurisdictions and pro-AI super PACs largely failing. Platformer argues that regulating AI cannot follow the social media playbook, which produced no comprehensive privacy law, no Section 230 reform and no new antitrust legislation despite years of hearings.
vLLM has migrated its CUDA build to the LibTorch Stable ABI, and as of v0.33.0 it publishes LibTorch ABI Stable CUDA wheels, so a new PyTorch wheel can be imported without rebuilding vLLM. The team moved in-tree sources and seven LibTorch-dependent dependencies, including FlashMLA, DeepGEMM and FlashAttention-2/3, onto stable APIs, using torch-abi-audit in CI to confirm no unstable symbols remain. The post notes tradeoffs such as a limited stable API subset and small host-side overhead, and says CPU and ROCm builds are next.
OpenClaw rebuilt how it discovers models so new ones like GPT-6.1 Sol appear without shipping a release, pulling lists from a self-updating catalog at catalog.openclaw.ai (rebuilt every four hours, about 1,000 models), each account's own provider model list, and a handshake with a newer Codex on your PATH. The catalog switches live on a running Gateway, and model lists no longer shrink after provider timeouts or catalog swaps.
NVIDIA Research and Inferact's REINFORCE Zero paper shows that aligning rollout and trainer kernels lets unclipped token-level importance sampling correct the remaining weight lag, training stably for 20,000 updates at policy lag 64 and reaching 2.40x the training-step throughput of synchronous training with aligned kernels. With standard kernels, the median rollout-to-learner KL rises 83.33x across the run and online AIME24 pass@1 falls from 34.27% to 13.96%, while a sequence gate stabilizes training but rejects 24.48% of sequences by the last 1,000 updates. The zero-mismatch setup now covers DeepSeek-V4.1-Flash, Kimi K3.1 on release day, and DeepSeek-V4-Flash and Kimi K3.1 on AMD MI355X, with code in Inferact/zero-mismatch-rl and NVIDIA-NeMo/labs-molt.
Why it matters: NVIDIA Research and Inferact split the async RL rollout-learner gap into weight lag and kernel mismatch, showing that aligning kernels lets unclipped importance sampling hold stable training at policy lag 64.
VeRL-Omni v0.3.0 was released with asynchronous and decoupled RL execution paths, the V1 trainer as the default, and expanded post-training coverage across image, video, audio, and multimodal understanding. The release adds fully async Qwen3-Omni text and audio training, separate-async Omni and diffusion training with a replay buffer, and new recipes including MiniMax H3 with DiffusionNFT and MiniCPM-o 4.5 GSPO. In a MiniCPM-o 4.5 AVQA experiment on 4×H800 80 GB, separate-async 3+1 was about 9% slower per prompt than colocated 4+0 at natural response length, roughly tied at 512 tokens, and about 6.7% faster at 1024 tokens.
Sierra, the customer experience AI platform, has been inducted into JPMorganChase's Hall of Innovation, with the award presented to CEO Bret Taylor at J.P. Morgan's 18th annual Technology Innovation Symposium in Palo Alto on October 29. The Hall of Innovation, established in 2011, recognizes privately held technology companies that have made a significant impact at JPMorganChase through innovation and close collaboration. Chase CIO Gill Haus cited Sierra as a strong collaborator in exploring agentic AI for banking, while Taylor said agents built on Sierra can be deployed across web, mobile, chat, email and voice to resolve complex tasks.
Sierra released Accelerators, prebuilt packages for high-volume workflows like order tracking, appointment scheduling, claims intake and ID verification, available now in Agent Studio. Each Accelerator ships with basic integrations and guardrails, and Ghostwriter adapts it to a company's own policies; updates from Sierra carry over while customer changes are tracked separately. Teams will soon be able to build and share their own Accelerators internally.
Newcomer and Felicis released Emergent, a list of 29 seed-stage or earlier startups nominated by current and former founder speakers from the Cerebral Valley AI Summits as having generational potential. Each company is capped at a seed round and paired with a blurb from its nominator. The largest cluster builds tools to run and control agents, with hardware, open source, and AI-native services also prominent.
SemiAnalysis released the first long-context, multi-turn agentic inference results on InferenceX/AgentX for Google TPUv7 Ironwood across four models, finding it competitive on performance per dollar or state-of-the-art in several cases. On GLM-5.2 in FP8, TPUv7 delivered 1.2x to 2.7x the tokens per dollar of B200, B300 and MI355X from 60 to 300 tok/s/user, while FP4 GPUs still led on GLM-5.3 because TPUv7 lacks native FP4 support until TPUv8 in Q1 2027. The piece also argues NVIDIA's bureaucracy slows driver and ecosystem contributions compared with AMD and Google.
Meta is rolling out real-time voice conversations and expressive avatars in the Muse app on mobile, web and Mac, starting in the US. Users tap the phone icon to talk, can interrupt responses, and choose from hundreds of voices plus voice-only, avatar or full-screen modes. Voice is coming soon to AI Glasses and Muse Charm.
Why it matters: Meta's Muse app now supports real-time voice conversations with expressive avatars, showing how voice and interruption handling are moving into consumer AI assistants.
Gary Marcus said AI companies would no longer get a free ride, and the market split sharply: chipmakers and neoclouds like CoreWeave were hit hard while Amazon, Apple and Google bounced. He now describes the follow-on session as a slaughter with almost no company exempt, after investors rotated toward larger tech firms better able to weather a cash crunch.
Google DeepMind rolled out Gemini Omni 1.2 Flash, now generally available across the Gemini API and Gemini for business, with native 15-second video generation, multi-turn conversational editing and stronger reference following. The price drops 30% from Gemini Omni 1.1 to $0.07 per second for 720p video, and users can draft at 360p or 720p before upscaling to 1080p or 4K. Clips can be set from 3 to 15 seconds in one-second increments, and outputs carry an imperceptible SynthID watermark.
Why it matters: Gemini Omni 1.2 Flash raises native clip length to 15 seconds and cuts the price 30% to $0.07 per second of 720p video, with multi-turn conversational editing now generally available.
OpenClaw has released Multiplayer, a free and open-source feature that lets multiple people connect to the same OpenClaw agent and workspace, with prompts appearing live across laptops and requests coordinated through a shared queue. In demos at GitHub Universe in San Francisco, two attendees edited the same 3D scene by typing prompts on separate machines without restating context. The OpenClaw Foundation says it has had 50+ people collaborating in the same shared environment and is continuing to expand the feature.
Why it matters: OpenClaw 把多人共享同一 Agent 与上下文做成免费开源功能，为团队协作中反复重建上下文的问题提供了一种可迁移的做法。
Google DeepMind rolled out Gemini Omni 1.2 Flash and made it generally available across the Gemini API and Gemini for business. The model generates and edits native clips up to 15 seconds, up from 10 seconds in Omni 1.1, supports multi-turn conversational editing, and holds character, product, and style references more consistently; 720p video now costs $0.07 per second, 30% less than Omni 1.1. Outputs carry a SynthID watermark, and the model is rolling out in Flow, Google Vids, the Gemini app, and Google Ads.
Why it matters: Gemini Omni 1.2 Flash cuts the price 30% to $0.07 per second for 720p video while extending native clips to 15 seconds, giving teams producing video at scale a cheaper option to compare against prior versions.
Google Research published a Nature study on AI-PPG-ACC-HTN, a multimodal system that detects signs of hypertension from smartwatch PPG and accelerometer data without cuff calibration. Validated against 24-hour ambulatory blood pressure monitoring in 196 participants across five US sites, it reached 65.8% sensitivity and 90.0% specificity for initial screening, and 78.4% sensitivity and 90.9% specificity for confirmatory testing. The authors describe the work as early feasibility research, not a replacement for clinical arm cuffs or a standalone diagnosis.
Why it matters: Google's cuffless smartwatch hypertension model is benchmarked against 24-hour ambulatory monitoring, showing how passive multi-day sensor aggregation could support opportunistic screening.
) references. Agents acted on the bug ID in 6% of encounters, and 83% of those actions were unprompted by the developer; 41% of unprompted detours chased already closed tickets while 57% hit still-open bugs and expanded scope. Controlling for developer, codebase complexity, task type, model tier and other factors, unprompted bug detours added 70% more steps, 77% more tokens and 105% more wait time per session.
Why it matters: Google's large-scale trace analysis shows stale TODO bug references pull coding agents into unprompted detours, and its cleanup agents removed thousands of obsolete references with over 95% reviewer acceptance.
Anthropic outlined how it designs memory for Claude agents, defining agent memory as knowledge retained across sessions and separating it from weight changes, in-context learning and compaction. 5 matched it with memory. 7 wrote 413 files across 225 folders and earned one badge.
Why it matters: Anthropic details how Claude's agent memory works, comparing models' note-taking on CL-Bench and Pokémon Red and showing how offline dreaming corrects and organizes stored notes.
Anthropic research argues that AI systems can now derive equations like the Hodgkin-Huxley model in an afternoon, but fitting experimental data is not the same as understanding. The work breaks scientific modeling into building, constraining, and interpreting models, and notes that AI has made the most progress on constraining them, where Bayesian inference is the standard method for mapping all plausible parameter sets rather than a single fit.