According to @GeminiApp, Google will demo Neural Expressive design with dynamic visuals and mode switching on Discord this Friday.

According to JeffDean, new work shows a single cortical neuron can classify images, recognize words, and solve parity, surpassing perceptron limits.

According to CNBC... Google’s AI project delay hit chip stocks, but hyperscale data center buildout and AI capex plans remain on track, per analyst reports.

According to Eric Topol, frontier LLMs from Google, OpenAI, Anthropic outperformed clinical tools in blinded clinician tests, per Nature Medicine.

According to demishassabis, DiffusionGemma generates text up to 4x faster than Gemma 4 by producing blocks simultaneously, per Google Gemma’s announcement.

According to Sundar Pichai, DiffusionGemma brings text diffusion to Gemma 4, generating blocks of text for up to 4x faster inference.

According to GoogleDeepMind, DiffusionGemma outputs up to 4x faster by generating text blocks simultaneously and self-corrects complex markdown.

According to @NotebookLM, Gemini App Notebooks are fully live in Europe, with web access for Google AI Ultra, Pro, Plus and mobile expansion coming.

According to JeffDean, Google’s Gemini 3.5 Live Translate adds speech to speech in 70+ languages, rolling out in Translate and Google AI Studio Live API.

According to @NotebookLM, Gemini 3.5 and Antigravity now power transparent chat and secure cloud notebooks with 100+ skills for deeper analysis.

According to @NotebookLM, agentic chat, advanced reasoning, and new output formats roll out to Google AI Ultra and Workspace AI Ultra Access users.

According to @NotebookLM, major upgrades add agentic chat, advanced reasoning, and new output formats for Google AI Ultra users.

According to Sawyer Merritt, Google signed a $920M per month deal with SpaceX for 110,000 NVIDIA GPUs from Oct 2026 to Jun 2029, per a new filing.

According to SawyerMerritt, SpaceX will receive $2.17B monthly from Google and Anthropic for compute capacity, implying a $26B annual run rate.

According to SawyerMerritt, SpaceX signed a $920M per month Google Cloud deal for 110,000 Nvidia GPUs, fueling large scale AI training and inference.

According to @NotebookLM, users can now view prompts and sources behind each artifact and quickly iterate to customize outputs.

According to JeffDean, Google’s Gemma 4 12B is a unified multimodal model with open weights that runs on laptops under Apache 2.0.

According to @demishassabis, Gemma 4 tops 150M downloads and adds a 12B model that runs locally on 16GB VRAM under Apache 2.0 for laptop-grade multimodal AI.

According to sundarpichai, Alphabet raised about $45B and plans $40B more via ATM in Q3 to fund AI demand, with $10B from Berkshire Hathaway.

According to @emollick, Stanford found Gemini 2.5 beat professors 75%, was rated less harmful, and newer models perform even better.