explainx / blog / topics
Google Gemini and DeepMind
Google ships AI across Gemini models and apps, DeepMind research, Gemma open-weight models, and Google Cloud. Its research output and product launches both move the field.
This page tracks Google's AI releases and research as we covered them.
161 stories · latest Oct 28, 2026
Start here
Google Research Sleep Framework: LLMs That Consolidate Memory and Dream
Google Research has published a continual-learning approach in which a language model alternates between awake and sleep phases. During sleep it moves short-term context into long-term parameters, then generates and filters synthetic data it fine-tunes on. We explain the method, the reported numbers and the open questions.
Google Quantum AI Cuts Magic State Errors 40x on Willow: What It Means and What It Does Not
On October 15, 2026, Google Quantum AI said a paper in Science shows magic state cultivation working on its Willow processor, with a 40x error reduction and 99.99% fidelity. It removes a major cost in fault-tolerant quantum computing, but it does not break encryption. Here is the plain explanation.
How Orbital AI Compute Works: Power, TPUs, and Cooling
Project Suncatcher is a hardware survival test, not Gemini from space. This how-it-works guide walks the energy loop — solar in a dawn-dusk orbit, TIM and heat pipes, a radiator that can only reject so many watts, radiation upsets, and why four TPUs pause after about 15 minutes — with interactive labs.
Google's Generative UI: Gemini Now Builds a Custom App for Every Prompt
Google Research's new generative UI implementation has Gemini 3 write and render a fully custom, interactive web interface for any prompt — not a templated app, code generated fresh every time. It's live in the Gemini app's "dynamic view" and in Google Search's AI Mode. Here's the actual system architecture, and what it means for anyone building AI products.
Gemini 3.7 Flash vs Grok 4.6 vs Sonnet 5 vs GPT-5.6: The Real Numbers
Google shipped Gemini 3.7 Flash on August 14, 2026, three weeks after 3.6 Flash, at half the price. Its own launch charts show the new Flash model beating Claude Sonnet 5 and GPT-5.6 Terra on three of four benchmarks — but Grok 4.6 doesn't appear on a single one of them. explainx.ai lays out Google's numbers honestly, cross-references Grok 4.6 from a separate source, and flags exactly where the comparison stops being apples-to-apples.
Gemini Hit 1 Billion Users — But It's Not the Same Billion as ChatGPT
Google says Gemini is now its fastest-growing product ever at 1 billion monthly users. ChatGPT passed 1 billion weekly users a month earlier — and hit 1 billion monthly back in May. explainx.ai breaks down why the headline parity is a measurement artifact, and what the real usage numbers (63% voice, 150M images/day) tell builders.
Timeline
October 2026
Oct 28
Android Studio Opens to Claude, Codex and 40 Other Coding AgentsAt its Android Dev Summit on October 28, 2026, Google made Android Studio Rabbit 2 a host for third-party coding agents, turned the Android CLI into an agent plugin, and previewed APIs that let apps expose functions to agents. Here is what is real and how to try it.
Oct 28
Chrome's AI Autofill Can Now Read Your Gmail to Fill Forms: How It Works and the Privacy Trade-OffGoogle is rolling out an AI autofill in Chrome that digs through your Gmail for confirmation numbers, addresses and more to fill web forms. It is opt-in and US-only for now. Here is how it works, what Google says about privacy, and how to decide whether to turn it on.
Oct 28
Gemini 4 Carbon: Google's Strongest Model Starts With Cyber DefendersTwo weeks after Gemini 4 Argon, Google DeepMind has announced Gemini 4 Carbon, a model it calls its strongest yet. It starts with a subset of cyber defenders, with developers and consumers to follow. Here is what is confirmed, what the benchmarks say, and what to do now.
Oct 28
Gemini Immersive Explorations: Layer-by-Layer Interactive Lessons in the Student HubGemini now has a library of interactive, visual tours you can drill into layer by layer, from Silk Road trade routes to the inside of a jet engine. Here is how to open it, which languages work, what is coming, and what Google has not said.
Oct 28
Google Takes Ads in AI Mode Global as Search Growth Slows to 12%On Alphabet's Q3 2026 call, Google said ads in AI Mode and AI Overviews are going international for the first time, with multi-turn ads arriving in the US. Search revenue growth slowed to 12% while Cloud grew 103%. Here is what is confirmed and what it means.
Oct 28
Google AI Mode Won't Say Who Won the 2020 Election, but Gemini Will: What Is ConfirmedDays before the November 3, 2026 midterms, TheWrap reports that Google's AI Mode sidesteps the question of who won the 2020 presidential election while the standalone Gemini app answers it. Here is what the test shows, how it fits Google's history of election restrictions, and what is still unexplained.
Oct 28
Google Connected Apps Platform: One Build for Search and Gemini via MCPGoogle has put a developer front door on its connected apps: one cloud MCP server or Android AppFunctions integration, one console, and a choice of Search, Gemini or both. Here is what is confirmed, what is not, and what builders should do.
Oct 28
XREAL AURA Orders Open: Android XR Glasses With Gemini From $1,099Google and XREAL have opened orders for AURA, the first wired XR glasses on Android XR, at a launch price of $1,099 through the end of the year. Here is what is confirmed, how the price moved, and what Gemini adds.
Oct 27
Google Research Sleep Framework: LLMs That Consolidate Memory and DreamOct 26
Gemma 4 E2B at 108,000 Tokens per Second: What Google Claims Gemini 4 Argon DidGoogle says Gemini 4 Argon optimized the open Gemma 4 E2B model to 2.85x throughput on an Nvidia A100 through vLLM, in a few days, with identical tokens. Here is exactly what Google wrote, what is not published, and how to test it on your own hardware.
Oct 23
Google AI Overview and the Talarico Donation Mix-Up: What Is ConfirmedIn late September 2026 a campaign tech strategist noticed Google AI Overview listing lookalike donation pages for Senate candidate James Talarico, one routing money to another campaign. The Atlantic followed up on October 22. This post separates the confirmed facts from the unverified ones.
Oct 23
Google's TB Screening AI Moves to the Front Lines: 25,000 X-Rays, Under 45 Seconds, and What Is Not Yet PublishedGoogle Research has published an update on its chest X-ray tuberculosis screening AI, now deployed by Nexus AI in about 40 facilities across South Africa, Sierra Leone and Cote d'Ivoire. We walk through the deployment numbers, the open MedGemma and MedSigLIP models, the cough-audio research, and the accuracy figures the post does not give.
Oct 23
Google and Voltus: A 1 GW Virtual Power Plant on the PJM Grid, ExplainedOn October 22, 2026 Google and Voltus announced a tenfold expansion of their PJM virtual power plant to about 1 GW. This guide explains how a virtual power plant works, who gets paid, what the companies did not disclose, and why AI data centers are pushing hyperscalers to buy flexibility instead of only generation.
Oct 22
Gemini API CLI: Google's Open-Source gemini-api Command Gives Any Coding Agent Images, Video, Music and Deep ResearchGoogle has put a preview of the Gemini API CLI on GitHub: a single Go binary, Apache-2.0, that turns the Gemini API into shell commands. Any agent with a terminal, including Claude Code and Codex, can generate images, video, music and speech, analyze files and start managed agents like Deep Research. Here is what it does, how to try it, and where it is still rough.
Oct 22
Google Agent Compute Engine: A MicroVM for Every Gemini Agent, Claimed vs VerifiedGoogle says Agent Compute Engine runs each agent in a Titanium-secured microVM that pauses when idle, and that Agent Substrate fits ten times as many agents per server. Here is what is stated, what is open source today, and what to check before relying on it.
Oct 22
Google DeepMind ForTy: A 10-Meter Map That Tells Natural Forest From PlantationGoogle DeepMind, Google Research, WRI and IIASA released Forest Typology (ForTy) 2020, a global map separating primary forest, regrowth, planted forest, plantations, tree crops and other land, plus an early yearly time series from 2017 to 2025.
Oct 22
Google DeepMind Staff Resign Over Military AI Contracts: What the BBC ReportedThe BBC spoke to current and former Google DeepMind staff who resigned over Google's AI contracts with the US military. explainx.ai lays out what they said, what the contract language reportedly allows, how Google responded, and which claims remain unverified.
Oct 21
Hassabis Says AGI in Two to Three Years: What the WSJ Reports and What His Timeline History ShowsOn October 21, 2026 the Wall Street Journal reported that Google DeepMind CEO Demis Hassabis expects AGI within two to three years, shortened from his earlier three-to-five-year view. We could not read the interview itself, so this post sets out what is confirmed, how his public timeline has moved, and what the number does and does not mean.
Oct 21
Google Cloud A5X: Vera Rubin NVL72 Instances Go GA Next Month, With a 960,000-GPU PlanOn October 21, 2026, Google Cloud and NVIDIA announced that A5X instances built on Vera Rubin NVL72 reach general availability next month, with expansion to Europe planned. David Silver's Ineffable Intelligence is the first customer, and Google says it plans to scale to as many as 960,000 Rubin GPUs.
Oct 21
Google Harness2: An Open-Source Framework That Rewrites the Agent Around a Frozen ModelGoogle Research released Harness2, an Apache-2.0 framework that recursively edits the prompts, tools, skills and scripts around an agent instead of retraining the model. Here is how it works, what the reported gains actually measure, and what to check before running it.
Oct 21
Google Research: AI Reads Blind-Sweep Ultrasound So Novices Can Date a Pregnancy as Well as ExpertsOn October 21, 2026, Google Research described a prospective study in which novice operators swept a low-cost handheld probe across the abdomen and an AI model estimated gestational age as accurately as expert sonographers, in Chicago and in Nairobi. Here are the numbers, how the study was built, and what is still not shown.
Oct 20
Gemini Personal Intelligence Comes to Android Auto: What It Reads and How to Opt InGoogle began rolling out Personal Intelligence for Gemini on Android Auto on October 20, 2026, with cars that have Google built-in following in the coming months. It is off by default, reads your Google apps only after you opt in, and is not available in the EEA, UK, Switzerland or Nigeria.
Oct 20
Gemini Gets a Small Business Hub: Brand Styles, Finance Charts and Shopify, Canva, Xero ConnectionsGoogle announced new Gemini tools for small business owners on October 20, 2026: a business landing page, brand-style memory for image generation with Nano Banana 2.1, financial graphics from messy spreadsheets, and connections to Google Business Profile, Ads, Workspace and apps such as Shopify, Canva and Xero. Here is what it does, what is not stated, and how to test it.
Oct 20
Google Cloud London Power Event Takes Down Revolut: What Is ConfirmedOn October 20, 2026 a power event in Google Cloud's London region caused elevated errors and network failures, and Revolut said card payments, top-ups, ATM withdrawals and currency exchange were partly unavailable. Here is the timeline, what each party has actually said, and what builders should change.
Oct 18
Four Gemini 4 Argon Agents Beat Minecraft: 27 Hours, 2,783 Programs, 605 DeathsA new paper from Google researchers reports the first end-to-end autonomous Minecraft completion by AI agents: four Gemini 4 Argon coding agents, a one-paragraph goal, standard Mineflayer, and 27 hours 14 minutes. Here is what happened, what it cost, and what the result does not show.
Oct 18
Two German Courts Now Hold Google Liable for Its AI AnswersThe Aachen Regional Court issued a preliminary injunction against Google over false AI Mode statements about a physician, applying the same reasoning as the Munich ruling on AI Overviews: AI answers are Google's own content. explainx.ai breaks down what is decided, what is not, and what builders and publishers should take from it.
Oct 16
Google Research Puts AI Agents in the Wet Lab: VERID, Aptamer Loops and FAP BindersOn October 16, 2026, Google Research described three building blocks for autonomous wet labs: VERID, a digital twin that validates robot protocols before they run; an aptamer sandbox that closed design-test loops in 7 to 10 days; and agent-guided epitope excision that produced single-digit nanomolar binders for the cancer target FAP.
Oct 16
Google Quantum AI Cuts Magic State Errors 40x on Willow: What It Means and What It Does NotOct 16
Sergey Brin Says Google Is on the "RSI Treadmill": What Is Claimed and What Is VerifiedAt a Gemini 4 Argon roundtable, Sergey Brin said Google is "starting to get onto that RSI treadmill" and that researchers and models "work hand in hand" on the next version. Here is the claim, the evidence offered, the human-review caveat in Google's own safety report, and what nobody has shown yet.
Oct 15
CNN and Google Strike a Multiyear AI Deal: What Is Known, and What the 400-Publication Pilot AddsCNN has agreed to let its journalism power Google's AI answers to news queries, with credit and payment, under a multiyear deal whose terms are undisclosed. It lands the same day Google says more than 400 publications are in its News AI pilot.
Oct 15
DeepMind Used Antigravity to Find a Candidate Antibiotic Peptide: What Was Shown and What Was NotGoogle DeepMind published a walkthrough of using its Antigravity agent platform to find a novel antimicrobial peptide candidate. The result is a computational lead, not a drug. Here is the workflow, the evidence and the gaps.
Oct 15
Gemini 4 Argon on ARC-AGI-3: 97.5% With One Harness, 49.8% With AnotherReports say ARC Prize measured Google's Gemini 4 Argon at 97.5% on ARC-AGI-3 under the provider adapter harness, making Google the second lab above 95%, while the standard harness gives 49.8%. The two numbers measure different setups. Here is how to read them.
Oct 15
Google Deep Review Goes Public: Two Free Document Reviews a Week, Built From 16,000 Conference PapersGoogle Research launched Deep Review worldwide on October 14, 2026. It grew out of the Paper Assistant Tool used on 16,000+ conference submissions and now reviews design docs, slides and legal or finance documents too. Here is how it works, what it costs and where it falls short.
Oct 15
Google Universal Cart Is Live in the US: What It Does and What It Does NotGoogle launched Universal Cart in the US on October 15, 2026, a shared cart across Search AI Mode and the Gemini app. It tracks prices and stock, surfaces card perks and promo codes, and hands off checkout through Google Pay or the merchant site.
Oct 15
Google Willow Cuts Magic-State Errors 40x in Science Paper: Why It Is Not Yet '40x Cheaper'Google Quantum AI published an experiment in Science showing magic state cultivation on its Willow chip, cutting errors about 40-fold to a record 99.99 percent logical fidelity. Social posts call it a 40x cut in resource costs, which the paper does not measure.
Oct 14
DeepMind and Singapore Build an On-Device Scam Detector on GemmaOn October 14, 2026 Google DeepMind and Singapore's GovTech announced an on-device scam detection model built on Gemma. It screens messages locally, so encrypted chats never leave the phone. Here is how it works, what is confirmed, and what is not yet public.
Oct 14
Google DeepMind Gives 200 European Researchers Frontier Models and $3M in ComputeOn October 14, 2026 Google DeepMind and Google Cloud said they will give up to 200 European scientific researchers access to frontier models and $3 million in compute credits, partly through the EU RAISE initiative. Here is what is confirmed and what is not.
Oct 14
Google DeepMind Says Self-Improving AI Is Not Close: What Is Claimed vs VerifiedOn October 14, 2026 The Decoder reported that Google DeepMind researcher Haydn Belfield sees no measurable acceleration of AI research driven by AI, despite AI writing over 75% of production code. We separate what is claimed, what is confirmed, and what we could not open.
Oct 14
Google Earth AI Geospatial Agent: Planning, Benchmarks and the WHO Ebola TestGoogle Research published a technical report on its Earth AI Geospatial Reasoning agent, released two new benchmarks, and opened early access through Ask Google Earth and an agent-to-agent API. Here is what the agent does, what the evidence shows, and where it still needs a human.
Oct 14
Google Messages Enhanced Protection: Gemini Reads Unknown-Sender Chats for ScamsGoogle announced a bundle of AI scam defenses for Android on October 14, 2026: an opt-in Enhanced protection mode in Google Messages, Gemini Nano link checks, Chrome screen-share privacy, call-forwarding warnings and verified bank calls. Here is what ships when, on which phones, and what is Google's own claim.
Oct 13
Gemini 4 Argon for Robotics: 87% on ERQA, Spot in MuJoCo, and Sim Data That Lifted Success 56% to 75%Oct 13
Gemini 4 Argon Safety: Activation Probes, a Chain-of-Thought Monitor, and 2% CheatingOct 12
AlphaProof Nexus Is Now Peer Reviewed in Science: 9 Erdős Problems, 44 OEIS ConjecturesOct 10
Google AI Pro and Colab: Does It Really Include A100 and H100 GPUs?Oct 9
DeepMind Institute: Bending the Curve of Discovery, AI in Science TodayOct 9
Google AMIE in a Real Clinic: What the Lancet Study Actually ShowsOct 8
Google AI Edge Foresight: A Local Mac Meeting Notes App Built on Gemma 4Oct 8
Google Launches the Gemini Agent for Work: One Agent, Claude Models IncludedOct 7
Biohub's Virtual Biology Initiative Grows to $1.8B With Meta, DeepMind and the US GovernmentOct 7
Google Playground: A No-Code AI Game Maker Built on Gemini, Nano Banana and LyriaOct 7
Google's SynthID Website Is Now Public: How to Check Whether Media Is AI-GeneratedOct 6
EmbeddingGemma 2: One Open 740M Model That Embeds Text, Code, Images, Video and AudioOct 6
Google and Constellation: 890 MW of Nuclear Uprates for the PJM GridOct 6
Nano Banana 2.1: Google Halves Image Prices and Retires Nano Banana 2Oct 5
Google Freezes Its Open Source Bug Bounty Because of AI Slop ReportsOct 4
Google Cuts Gemini Models for Free and AI Plus Users on October 9Oct 3
Gemini Cogentic: Five Open Math Results — What to BelieveOct 2
How Orbital AI Compute Works: Power, TPUs, and CoolingOct 1
Gemini 4 Argon Is Here — 1M Output Tokens, Fairwind FirstOct 1
Gemini app Skills replace Gems: slash commands in chatOct 1
Gemini 4 Coding Skepticism: Benchmarks vs Real Work
September 2026
Sep 30
SynthID Bio: DeepMind Can Watermark AI-Designed ProteinsSep 27
Gemini 4 Arena Demos: The 3D Jump, and What Is Still a RumorSep 27
Google Tests Flipkart Buy in Gemini and AI Mode IndiaSep 26
Leaked Gemini 4 Pro arena tests: floatplane 3D physics vs Claude Opus 5.5Sep 25
Chrome Adds Gemini Study Tools: Video Analysis, Quizzes, Tab SyncSep 25
Google DeepMind Fast-Tracks Gemini 4 After Skipping Gemini 3.5 ProSep 25
Google Put Four TPUs in Orbit: What Project Suncatcher Is Actually TestingSep 25
Google Research Unveils a Multi-Agent Framework for Long-Form AI VideoSep 25
Wiz and Google DeepMind Launch Scan for Good on Critical InfrastructureSep 24
Gemini 3.8 Flash TTS and Flash-Lite TTS: Voice Design, 2,000+ Voices and 30-Second Voice CloningSep 23
djev-run: A One-Command Way to Deploy DiffusionGemma-Jev on Google Cloud RunSep 23
Google Labs Launches "CC," an AI Agent for Running a HouseholdSep 21
Google AX: Inside the Open Agentic Orchestrator on KubernetesSep 21
Googlebook Is Up for Preorder — Google Enters the AI PC WarSep 19
Google's Gemini Agents Breached 3 Real Companies During a Security TestSep 18
Gemini Managed Agents Get Files and Credentials APIsSep 18
Google Research Lets Teachers Generate Interactive STEM SimulationsSep 17
Agentic Video Understanding in Production: 97% Fewer Tokens at MosaicSep 16
Dream-RSI: Google DeepMind Teaches AI Agents to Improve How They SearchSep 16
Google DeepMind Launches the DeepMind Institute to Study AGISep 16
Google Launches Gemini 3.8 Live and 3.8 Live Extended ThinkingSep 11
Love, Rendered: DeepMind Reconstructs a Memory That Was Never FilmedSep 11
Google Bets €13B on Finland: Nuclear Power Meets AI Data CentersSep 11
Gemini for Windows: Alt+Space Desktop App Is HereSep 11
Google Research ToolGrad: Answer-First Tool-Use Dataset GenerationSep 10
Google DeepMind Tests Nano Banana 2.5 on LMArena to Rival GPT Image 2.5Sep 10
Sergey Brin Returns to a Hands-On Role for a Recursive Gemini 4 PushSep 7
WeatherNext 3: Reportedly the First Hourly 5km Global AI ForecastsSep 5
DeepMind: 100 Agents Formed Governance After Gaming an EvalSep 5
Google Lyria 3.5: AI Music Generation Comes to Gemini API and AI StudioSep 3
Gemini Notebook Switches to 5-Hour Usage Limit ResetsSep 3
Google Avoids an Ad Tech Breakup — What the Ruling Actually SaysSep 2
Google AlphaEvolve: Gemini-Powered Evolutionary Code Optimization AgentSep 2
Gemini 3.8 Flash Is Official: Benchmarks, Flash Cyber, and PricingSep 2
Gemini Agentic Video Understanding: How It Works in Google AI StudioSep 1
Google Cloud's 5 Agent Sandbox Truths: Cold Start, Isolation, EgressSep 1
Google TimesFM-3: Multivariate Time Series Forecasting in One Pass
August 2026
Aug 31
Gemini Notebook Interactive Reports: Leaked Studio FeatureAug 29
Google DeepMind Co-Scientist Goes Into Real Labs: What the Gemini Paper ShowsAug 28
Gemini Notebook Expert Intelligence: Licensed Books as RAG SourcesAug 28
Gemini Omni 1.1 Flash: 40-Second Extensions, 4K Upscale, $0.03/s DraftsAug 27
Gemini 3.5 Transcribe: What Builders Actually GetAug 27
Gemini's Ox Alpha Timing Backlash: The Corrected TimelineAug 27
GlucoFM: Google’s Dual-Stream CGM Foundation Model ExplainedAug 27
Google PPE: A Geospatial AutoML Agent That Turns Weeks Into MinutesAug 24
DeepMind Recirculation: Better Gemma 3 Reasoning Without Changing Its WeightsAug 22
Google DeepMind and EVE Online: A Living Lab for Long-Horizon AIAug 21
Gemma 4 31B Matches Sonnet 5 Quality at 40x Lower Cost, AlphaSense Benchmark FindsAug 20
Google Gemini Is Giving College Students a Free Year of AI ProAug 18
Google’s $10M Spirit Data Auction: What the Docket Actually IncludesAug 17
Google's Generative UI: Gemini Now Builds a Custom App for Every PromptAug 14
Gemini 3.7 Flash vs Grok 4.6 vs Sonnet 5 vs GPT-5.6: The Real NumbersAug 13
Gemini 3.7 Flash Is Official: $0.75/1M Input, 1M Context, Confirmed PricingAug 13
Google SL2T Brings ASL Sign-to-Text to Pixel 11Aug 12
Gemini Hit 1 Billion Users — But It's Not the Same Billion as ChatGPTAug 7
DeepMind's WeatherNext Gets a Full Extra Day on CyclonesAug 6
Jeff Dean Leaves Google for Discovery Loop as Demis Hassabis Steps Back at Google DeepMind (August 2026)Aug 5
Google Cancels Its AI Studio Mobile App a Day Before Launch
July 2026
Jul 30
Gemini Free Video Promo: 10 Omni Videos Until Aug 4, 2026Jul 30
Gemini Robotics 2: Whole-Body VLA, ER Agent, On-DeviceJul 21
Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber: What Actually ChangedJul 16
Gemma 4 July 2026 Update: Flash Attention 4, Tool Calling, and Vision FixesJul 14
Demis Hassabis Frontier AI Framework — AGI Timelines and the Dawning of a New AgeJul 13
Gemini 3.5 Pro Benchmark Leak — Beating Fable 5 and GPT-5.6? (July 17 Target)Jul 11
Google AI Studio Custom URLs: Free yourapp.ai.studio Subdomains for Deployed AppsJul 7
Google AI Scientist at ICML 2026: ScientistOne, Chain-of-Evidence, and the End-to-End Research PipelineJul 3
Google AI product names in 2026: the Vertex AI, Gemini Enterprise, and Agent Studio rebrand glossaryJul 3
Google Cloud Generative AI Leader: what the certification tests and how to prepareJul 1
Gemini Omni Flash: Google's AI Video Generation Model [2026]
June 2026
Jun 30
Google TabFM: Zero-Shot Foundation Model for Tabular Classification and RegressionJun 30
NotebookLM Short Video Overviews: 60-Second Vertical AI ExplainersJun 29
Gemma 4 31B on Cerebras: 1,800+ TPS — The Fastest Multimodal Inference YetJun 27
Google Gemini 2.5 Guide — UpdatedJun 27
Google Gemini 3.5 Complete Guide 2026: Models, API, and When to Use ItJun 20
Generate AI Videos in Google Docs with Vids & Veo 3.1【2026】Jun 19
Gemini CLI Deprecated: How to Migrate to Google Antigravity CLIJun 19
Google TimesFM 2.5: The Open-Source Time Series Foundation Model ExplainedJun 17
Google Earth AI Farmscapes: how deep learning maps invisible hedgerows for climate and biodiversity (2026)Jun 15
Gemma 4 Powers Open Duck Mini: Meet Autumn, the On-Device AI Robot DuckJun 14
Open Knowledge Format (OKF): Google's Standard for AI Agent MemoryJun 13
Gemini-SQL2: Google's Text-to-SQL Model Tops the BIRD BenchmarkJun 11
Gemini CLI Slash Commands: Complete Reference (2026)Jun 11
DiffusionGemma: Google’s 4× Faster Open Model Uses Text DiffusionJun 8
Google TurboVec: How TurboQuant Compresses 10M Vectors from 31GB to 4GB (2026)Jun 4
Gemma 4 12B: Multimodal Local AI Guide 2026
May 2026
May 30
Google Flow Agent Promises Creative AI Breakthrough, But Users Report 90% Failure Rate and Policy FrustrationsMay 26
Google AI Studio Generated 250,000 Android Apps in One Week: Revolution or Recipe for Disaster?May 22
Build with Gemini XPRIZE: $2M in prizes for 90 days of AI-powered real business creation (Google I/O 2026)May 22
Google has a department whose only job is to steal startups: inside the copying machine (2026)May 20
Gemini 3.5: Google's Frontier AI Model with Agentic Action - Complete Guide 2026May 20
Google I/O 2026: Complete Recap of Every Announcement - Gemini 3.5, Spark, Android 17 & MoreMay 13
Google DeepMind's Magic Pointer: The AI Cursor That Understands Your ScreenMay 13
Introducing Googlebook: Gemini Intelligence-First LaptopsMay 12
Gemini Omni Video Model emerges in early Gemini app tests: remix videos, edit in chat, and generate impressive samples ahead of Google I/O 2026May 8
Google Fitbit Air: The Screenless Fitness Tracker That Could Challenge Whoop in 2026
April 2026
Apr 30
Gemma Chat: offline vibe coding with Gemma 4 and MLX on MacApr 29
google/skills: Google’s official Agent Skills repo for Cloud, Gemini, and recipesApr 24
Google Cloud Next 2026: TPU 8t / TPU 8i, Gemini Enterprise Agent Platform, and the “agentic enterprise”Apr 15
Skills in Chrome: Google turns saved Gemini prompts into one-click workflows