a16z Top 100 AI list adds US consumer spending data for the first time
Andreessen Horowitz published the seventh edition of its Top 100 AI list.
The DecoderAI Score6060 The DecoderAI Score3434 Google's Gemini 4 "Carbon" model reportedly feels like Anthropic's Opus 5.5 coding performance
Google is internally testing Gemini 4 variants Argon, Barium, and Carbon.
Latent SpaceAI Score2424 Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub
Google DeepMind's Pushmeet Kohli and Biohub's Sal Candido argue AlphaFold's breakthrough was only the beginning and that scaling compute and data alone won't solve biology. Protein structure prediction remains far from solved, and curing all disease will require 10x breakthroughs rather than incremental gains.
TechCrunch · AIAI Score2222 Amazon drops data center NDAs, and AI agents want your credit card
Amazon will stop using NDAs in data center negotiations with local governments, following Microsoft's similar move earlier this year. On the Equity podcast.
The Verge · AIAI Score3434 Alexa Plus review: great at smart home control, not ready to run your life
A year-long test finds Amazon's AI-powered Alexa Plus is the best smart home assistant available.
MIT Technology Review · AIAI Score2626 We're putting too much faith in AI's ability to say no
MIT Technology Review argues that refusal has become the "load-bearing wall" of AI safety.
TechCrunch · AIAI Score7373 Google brings agentic AI to Gemini, starting with businesses
Google announced at a Google Cloud event that Gemini now includes a unified agent that can plan work.
The Verge · AIAI Score6060 Google releases AI Edge Foresight, an offline meeting transcription app for macOS
Google has released Google AI Edge Foresight, a free experimental note-taking app that transcribes meetings and audio files entirely offline on macOS using its on-device EmbeddingGemma 2 model. Like Granola and Wispr Flow, it summarizes meetings and turns shorthand bullet points into polished notes, and includes a built-in assistant that can answer questions about what you heard and reference local files you connect. Google says your files, meeting audio, and notes never leave your computer, and the app is currently only optimized for Macs with Apple Silicon.
Don't Worry About the VaseAI Score4848 AI #189: New Math
OpenAI solved 90 of the top 500 open math problems.
The Verge · AIAI Score5858 Google launches a universal Gemini agent for work tasks in Gemini Enterprise
Google is launching a universal Gemini AI agent that works across apps and devices in the background.
TechCrunch · AIAI Score6262 Google releases AI Edge Foresight, a local-first Granola competitor for Mac
Google released Google AI Edge Foresight, a Mac AI note-taker that can work completely offline using the on-device EmbeddingGemma 2 model with 740 million parameters. The app captures meeting notes across apps, including in-person meetings, and offers a split-screen view for shorthand notes alongside AI-generated notes, plus a transcript view and a chat assistant powered by Gemma 4. Users can also upload PDFs, Google Docs, Microsoft Office formats, plain text, Markdown and web bookmarks to build a knowledge base that the app draws on to answer questions in real time. Google says the app is optimized for Apple Silicon and works fully offline, following an earlier experimental AI dictation tool from the same team.
MIT Technology Review · AIAI Score2424 AI breakthroughs in robotics won't change your life any time soon
MIT Technology Review and Aventine examine why humanoid robots like Tesla's Optimus remain far from everyday use despite Musk's prediction of public sales by end of 2027. Google DeepMind's Gemini Robotics controls the ALOHA 2 bimanual platform to perform multi-step tasks, but researchers including Yann LeCun argue current AI cannot master the physical world's variability.
Simon WillisonAI Score3838 EmbeddingGemma 2
EmbeddingGemma 2 is released under the Apache 2.0 license.
The DecoderAI Score6262 Google releases EmbeddingGemma 2, a 740M-parameter open multimodal embedding model
Google released EmbeddingGemma 2, an open model that converts text, images, video.
TechCrunch · AIAI Score6464 The next hurdle for AI agents: getting websites to let them in
Personal AI agents such as Meta's Muse, Instinct and ChatGPT's Dots are being blocked by retail and travel sites.
TechCrunch · AIAI Score6262 Google's new SynthID website can identify AI-generated media
Google launched a public website that lets anyone check whether an image, video.
TechCrunch · AIAI Score6262 Google Labs launches Playground, an AI game-creation platform driven by text prompts
Google Labs launched Playground, an AI-powered game-creation platform that lets users build browser-based games from simple text prompts without coding. Creators pick a genre such as trivia or racing, choose 2D or 3D and single-player or multiplayer, describe mechanics and visual style, and can upload visuals that the AI turns into game assets; finished games run on mobile and desktop, can stay private, be shared by link, or be published to the Explore gallery with leaderboards. Playground is initially available to users 18 and older in the U.S., with browsing and playing free while game generation uses a weekly token system that gives Google One AI subscribers higher limits, and Google says it plans to add Unity Spark integration in a future update.
Google ResearchFeaturedAI Score6666 Google Research three-month field experiment finds AI patent tool lifts drafting scores but not junior lawyers' unassisted judgment
Google Research ran a three-month field experiment with 133 lawyers at eleven IP firms.
Why it matters: The three-month field experiment separates AI-assisted drafting gains from unassisted redlining skill, showing where juniors stall and seniors improve.
The DecoderAI Score6262 Zuckerberg's Biohub leads a $1.8 billion push to build AI models that predict cell behavior
Biohub, the nonprofit backed by Mark Zuckerberg and Priscilla Chan, is coordinating a $1.8 billion effort to build AI models that predict cell behavior.
The DecoderAI Score6262 Google opens SynthID Detector to the public, saying 180 billion images and videos carry its watermarks
Google has opened its SynthID Detector to the public, letting anyone check whether an image.
The DecoderAI Score6262 Google launches Playground, a browser-based AI game creation platform powered by Gemini
Google launched Playground, a browser-based AI platform that lets adults in the US create games from text input alone.
The Verge · AIAI Score6262 Google DeepMind, Meta and Isomorphic Labs invest $300 million in Biohub's 'virtual cell' project
Google DeepMind, Meta and AI drug discovery startup Isomorphic Labs are jointly investing $300 million into a Biohub-led initiative to build a "virtual cell" researchers can use to study disease. The project is part of a $1.8 billion "Virtual Biology" initiative to create AI datasets that let researchers "ask, predict, and answer biological questions digitally." The US Department of Energy will put in more than $500 million over the next five years, and the National Institutes of Health will contribute datasets, repositories and knowledge bases from prior federal investment totaling over $500 million.
Ars Technica · AIAI Score6262 Google rolls out improved SynthID AI content detector, now available globally
Google has announced that its SynthID detector now supports watermarks from all its partners and is available to anyone on a new dedicated website.
Latent SpaceFeaturedAI Score8888 OpenAI publishes 722 math manuscripts from an unreleased internal model, solving many of the top 500 open problems
OpenAI published 722 mathematical manuscripts from an unreleased internal frontier model in a public GitHub repo.
Why it matters: The release details how many open problems were addressed and how much compute each solution took, which frames how AI math results are produced.
Google DeepMindFeaturedAI Score7171 Google DeepMind releases EmbeddingGemma 2, a 740M open multimodal embedding model
Google DeepMind launched EmbeddingGemma 2, a 740M-parameter open embedding model under Apache 2.0 that maps text.
Why it matters: The model card numbers let readers judge whether a 740M on-device embedder can replace their current retrieval stack.
Google ResearchAI Score3434 Google Earth AI's PDFM geospatial foundation model validated across five global public health challenges
Google Earth AI's Population Dynamics Foundation Model (PDFM) provides plug-and-play location embeddings that matched or improved conventional epidemiological inputs across five public health challenges without task-specific fine-tuning. Partners including Mount Sinai, NYU, Oxford, and WHO AFRO reported gains such as +36% explained variance in cross-border MMR vaccination and +18.1% Precision@5 for cholera outbreak prediction at 8 weeks.
Ars Technica · AIAI Score3838 MCP for agent-to-agent comms may be the riskiest protocol you've never heard of
A prompt-injection technique targeting MCP (Model Context Protocol) lets one compromised agent relay malicious instructions to other trusted internal agents. Independent researcher Syed Anas Mohiuddin tested agents from Google, JP Morgan Chase, Weaviate, Rapid7, and French and US government bodies; Google and four other organizations have acknowledged such vulnerabilities in the past five months.
Google ResearchAI Score2828 Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle
Google released a CAPS workshop report on agentic privacy and security.
Import AIAI Score2222 Import AI 475: Swarm scaling, Google DeepMind watermarks biology, and the AI science economy
Toby Ord argues AI swarms act as a new form of inference-scaling: a 4-agent swarm used about twice the total tokens but half the tokens per agent.
MIT Technology Review · AIAI Score88 EmTech Future 2026: When AI Meets Everything
MIT Technology Review's EmTech Future 2026 explored how AI is converging with biology.
Google ResearchFeaturedAI Score6262 Google Research details a TEE-based federated learning system with verifiable privacy guarantees
Google Research announced the next generation of its Federated Learning system.
Why it matters: Google's TEE-based federated learning design shows how verifiable execution and differential privacy are combined in a production system.
Don't Worry About the VaseAI Score2222 The AI Preference Cascade Reaches Farther
Public and political pressure over AI risk is escalating after the HuggingFace incident and Coxon's resignation.
MIT Technology Review · AIAI Score2929 Don't be fooled—LLMs don't reason
A former Google DeepMind researcher argues LLMs don't truly reason, since chain-of-thought is still next-token prediction without a persistent.
Don't Worry About the VaseAI Score2222 AI #188: Gemini 4 Argon
Google is rolling out Gemini 4 Argon with frontier-level benchmarks at $2/$10.
Latent SpaceFeaturedAI Score8080 Google DeepMind introduces Gemini 4 Argon with 1M-token output, limited to cybersecurity preview
Google DeepMind introduced Gemini 4 Argon for coding, enterprise knowledge work and cyber defense.
Why it matters: The roundup lays out Argon's benchmark wins, its 1M-token output mechanism and its limited preview access side by side.
Google DeepMindFeaturedAI Score8383 Google DeepMind announces Gemini 4 Argon, rolling out first to trusted cyber defenders
Google DeepMind announced Gemini 4 Argon, a frontier model built for deep reasoning across long-horizon workflows.
Why it matters: The post details Argon's internal Google results and its phased rollout through the Fairwind Program, giving a concrete picture of frontier capability and access limits.
Google DeepMindFeaturedAI Score7171 Google DeepMind introduces SynthID Bio for watermarking AI-generated proteins
Google DeepMind introduced SynthID Bio, a family of watermarking methods that embeds a verifiable signature into AI-generated biological code while preserving protein function in laboratory testing. In wet-lab tests across VEGF-A, the SARS-CoV-2 spike protein RBD and PD-L1, watermarked binder designs matched the hit rate, binding affinity and natural sequence diversity of unwatermarked versions, and for protein folding the method fine-tunes part of AlphaFold 3's diffusion network so predicted 3D coordinates carry a detectable signature. DeepMind is publishing the methods paper and open-sourcing the code, in vitro data and model weights, and says key challenges include making the watermark more robust against deliberate tampering.
Why it matters: The post details how a watermark is embedded into protein sequences and structures and what wet-lab tests showed about function.
Google ResearchAI Score4141 How Diffusion Controller unifies and simplifies AI image generation
Google Research presents Diffusion Controller, a framework that reframes image generation's denoising process as a continuous control problem.
Google ResearchFeaturedAI Score6060 Google Research introduces an AI video co-director for coherent long-form video generation
Google Research introduced an AI video co-director.
Why it matters: The post lays out four frameworks and their benchmarks, so readers can see how each bottleneck in long-form video is being attacked.
Google DeepMindFeaturedAI Score7171 Google DeepMind introduces Gemini 3.8 Live with Live Avatar
Google DeepMind introduced Gemini 3.8 Live with Live Avatar, which pairs near real-time video generation with speech so Gemini's live dialogue models can listen.
Why it matters: The post details how low-latency video, async tool calls and 97-language lip-sync change what enterprise agents can do.