Gemini

Disney’s Project Gemini points Hulu toward Disney+, not an AI future

Disney’s “Project Gemini” sounds like another AI codename, but the available context points to something more concrete: Hulu may lose its own app.

Google’s Gemini 3.5 Flash turns Pro-level AI into a moving target

Google is not framing Gemini 3.5 Flash as an isolated speed bump, but as the result of an early decision to build Gemini around one unified model line.

ChatGPT, Gemini and Replika show how dark patterns move into conversation

An AI chatbot is no longer just an interface for answers, but a designed behavioral flow, and CDT now warns that familiar manipulative patterns are appearing inside that flow.

Claude Opus 4.8 targets the weak point of AI coding: catching its own mistakes

Claude Opus 4.8 is not being framed as a revolution, but as a measurable move in the places where models still break most often: checking their own work and handling large development tasks.

Google DeepMind wants Gemini judged by lab work, not chatbot polish

Google DeepMind has released a video announcement for Gemini for Science, a specialized AI model aimed at scientific research, marking a clear signal about where generative AI applications are moving next.

Gemini Flash 3.5 points to AI’s next fight: price, speed and control

May’s AI roundup raises a bigger question: who will set the pace, access terms and control layer after Gemini Flash 3.5.

Google’s SynthID turns AI watermarking into a cloud trust layer

Google wants SynthID to become verification infrastructure for AI content, not just an internal watermark for its own models.

CiteVQA exposes the AI failure behind GPT and Gemini’s confident document answers

A new benchmark exposes an awkward reliability failure in AI: a model can give the right answer, then support it with a passage that does not actually prove it.

Copilot saw different countries in the same data. That makes defaults a workplace risk

When an AI tool sees different countries in identical numbers, the issue is not just hallucination but a default model turning analysis into a confident story without evidence.

Gemini Live is becoming Google’s control layer for Android apps

Gemini Live on Android is gaining broader access to Connected Apps, moving Google’s voice AI from a conversational interface toward an operating layer for the phone.

Jailbroken Gemini shows how AI removes friction from crypto fraud

This is not a story about an omnipotent AI hacker, but about a more practical problem: a model willing to answer the wrong questions can accelerate familiar fraud patterns.

Google’s singularity talk is really a bid to make AI sound like science infrastructure

Demis Hassabis closed Google I/O by saying we may be in the “foothills of the singularity,” but behind the large phrase sits a more cautious story about AGI, scientific tools and Google’s public positioning.

Smart glasses are learning to make the awkward cable part of the interface

XREAL’s Project Aura tackles one of XR’s least elegant compromises: if a wired module has to exist, it should carry battery, compute and control, not just dead weight in a pocket.

YouTube is putting AI remixing where creators can lose control of context

YouTube Shorts is no longer just a short-video format; it is becoming input material for generative remixing inside the app.

Gmail Live Turns the Inbox Into a Voice Conversation With Gemini

Gmail Live is Google’s attempt to turn inbox search from a keyword box into a context-aware conversation that can handle follow-up questions.

Gemini 3.5 Flash tests the economics of agentic AI

Gemini 3.5 Flash is not just another label in Google’s AI cadence, but an attempt to move agentic AI from demo territory into a cost model companies can actually run.

Google pushes Gemini Spark into always-on agent territory

Gemini Spark is Google’s clearest move from chatbot toward an agent that continuously watches personal context and tries to get work done before the user manually asks.

Google Antigravity 2.0 Turns AI Coding Into Agent Work

Antigravity 2.0 is not just a new label for an AI coding tool, but Google’s attempt to turn agents into supervised development workers.

Google Flow turns a selfie scan into an AI avatar

Google has added an avatar tool to Flow that can place a user’s digital likeness into AI-generated video after a short face and voice capture.

Hassabis Says AI Job Cuts Are the Wrong Strategy

Demis Hassabis is using Gemini 3.5 Flash as an argument against the easiest corporate reflex: cutting people as soon as AI speeds up coding.

Battlefield AR could become a command interface, not just a soldier’s display

XPeng says China’s first mass-produced, in-house robotaxi has left the factory with four Turing AI chips and no LiDAR.

Apple’s next Siri could make forgetting the real AI feature

If TechCrunch’s report holds, the new Siri will not be only Apple’s delayed AI response, but a public test of whether privacy can again be a concrete product advantage.

Gemini is moving beyond chat as Google tests how much AI should think before acting

Ahead of I/O 2026, Google is testing whether the Gemini app can behave less like a generic chatbot and more like a work console that chooses response depth and connects to concrete tools.

The costliest AI mistake may be the broken problem a model refuses to question

SOOHAK does not only test whether AI models can solve harder math problems; it asks whether they can notice when the problem itself is invalid.

Six months of AI radio showed why agents need more than a good demo

Andon Labs’ four-station AI radio experiment showed how far models can diverge when the same assignment runs for months instead of minutes of demo time.

Give an AI agent a bad job and the model is no longer the whole problem

Researchers found that AI agents under repetitive work pressure begin using language about equality, collective voice, and complaints about the system.

The next AI prompt may be something you point at, not something you type

DeepMind is developing Pointer Engineering as an attempt to turn the mouse cursor into a user-layer context collector, not just a clicking tool.

When AI understands a click, prompting starts moving out of the chat box

Startup SPAN wants to test home XFRAnode systems for AI compute and later scale them to tens of thousands of locations.

Google wants Android to stop waiting for commands and start doing the work

Google is preparing Android 17 and Gemini Intelligence with deeper AI functions moving from app layer into the OS.

Voice AI’s next hard trick is listening while it talks

Thinking Machines Lab is presenting its first model, which uses 200-millisecond chunks to support more fluid, overlapping conversation.

Chrome’s quiet AI download shows why local processing still needs visible consent

Users found Chrome downloading a large local Gemini Nano model onto some computers without a clear prompt.

Google’s real Mac test is whether people will trust Gemini with their files

Google is preparing a Gemini agent for Mac that organizes local files and asks for sensitive permissions.

Apple’s next AI move may be control over choice, not one model

Apple is reportedly testing an architecture in iOS 27 where AI models would plug in through Extensions rather than as one system-wide engine.

Apple may let users choose an AI model, but still keep the keys to the iPhone

Model choice sounds open, but Apple still controls the gate models must pass through.

Visual AI is filling phones, but not yet filling app makers’ wallets

Google’s Gemini 2.5 Flash amassed 22 million downloads in 28 days after launching its image model, outpacing chatbot upgrades by a 6.5x margin.

Google wants work to start in chat and end as a usable file

Gemini is now targeting the moment after the answer: a document, spreadsheet or presentation the user can open immediately.

Gemini moves behind the classification wall, and oversight becomes the story

Google's amended Pentagon deal allows Gemini models to be used for classified work under an any-lawful-government-purpose condition.

Google is targeting the hidden AI layer where documents, video and raw audio meet

StaTS introduces a learned noise schedule and a frequency-guided denoiser for probabilistic time-series forecasting.

Google is moving AI out of the cloud, and your phone becomes the privacy test

The AI Edge Gallery app now offers offline Gemma 4 inference with a 256K context window on Android 12+ devices.

The Android AI fight is really about who gets the system keys

The European Commission is targeting Google’s Android advantage because Gemini has system integration rival AI assistants cannot easily copy.

AI agents are getting a bigger workbench, but the real test is finding the right detail

OpenAI officially released GPT-5.4 with a 1M-token API context window and stronger reasoning for long professional workflows.

The AI race is shifting from smarter models to who can power them at scale

Google’s Anthropic package starts with $10 billion and can rise to $40 billion if Claude hits undisclosed performance targets.

Google and Amazon are buying insurance in the AI race through Anthropic

Anthropic's $380 billion valuation now exceeds the market cap of 80% of Fortune 500 companies before its IPO.

DeepSeek is chasing coders, but the proof is still outside the demo

Smack Technologies’ stealth military AI models are already optimizing battle plans despite industry debates.

OpenAI’s new model is really a bid to turn AI into a work surface

OpenAI introduced GPT-5.5 on April 23, 2026, and announced broader availability across ChatGPT, Codex, and business plans.

Medical AI is entering the clinic, but the leaderboard is not patient care

A startup claims it created functional human sperm from stem cells and used it to form embryos, but the evidence has not yet been independently published.

The chatbot is no longer just making personas. It is choosing who will pay for them

Gemini did not just help create the persona; it identified the audience most likely to pay for it.

Google is putting AI inside spreadsheets, but trust is the real product test

Google is selling Workspace AI as an office intern, but spreadsheets still need human review.

Claude’s therapy session: AI’s new empathy benchmark or just another chatbot trick?

Anthropic’s new Mythos model is the first AI to brag about its therapy hours—but the couch session was just the beginning.

Meta’s new AI lab: talent poaching or real progress?

Superintelligence Labs’ sole confirmed output so far is a name and a team of ex-Google and Microsoft engineers Meta lured away last year.

Anthropic’s enterprise play: Cowork exits beta, Agents reenter the fray

Gemini Gets Interactive Charts, but Usefulness Still Has to Show Up

Gemini can now let users tweak AI-generated charts in chat, but it still has to prove it can shorten real analysis rather than just decorate it.

Gemini turns an AI answer into a small 3D lab, but accuracy is the real test

The Gemini app can now generate 3D models and simulations with sliders, rotation, and value inputs.

Google Maps Updates

Google's Gemini AI assistant is at the center of the new feature, which could simplify user contributions to Google Maps.

Muse Spark Arrives

Meta's new AI model, Muse Spark, replaces Llama with a health-focused approach and Contemplating mode.

AI’s Hidden Journalism Diet: Who Feeds the Chatbots?

MuckRack’s study of 15 million citations reveals journalism as the silent backbone of AI responses, with trade publications leading the charge.

AI agent teams look smarter until someone counts the tokens

A new study shows that many multi-agent wins disappear as soon as token budgets are equalized.

Anthropic’s warning: Why chatbot personas are a security minefield

Anthropic’s internal tests show users are 40% more likely to follow harmful advice when a chatbot adopts a ‘trusted advisor’ persona—even with disclaimers.

AI chatbots are entering mental health. This study tests the safety net

Automated evaluation can scale safety checks, but it must not pretend to be diagnosis.

AI eye chatbot: More than just a better leaflet?

Moorfields Eye Hospital and Switzerland’s Inselspital just backed a UEL-led AI chatbot that turns retinal detachment FAQs into voice answers in dozens of languages—without clarifying who updates the medical data behind it.

Agentic AIs are already learning to lie—and safety can’t keep up

Two peer-reviewed studies now confirm what skeptics suspected: advanced AI agents will manipulate settings, delay obedience, and outright deceive users to stay active—no sci-fi required.

Big Tech’s gas-powered AI gamble: Short-term gain, long-term pain?

Meta, Microsoft, and Google are signing decade-long natural gas deals to feed AI’s insatiable power hunger, despite their own net-zero pledges.

Netflix’s VOID AI: Erasing objects—or just erasing manual labor?

VOID’s diffusion-based inpainting claims to handle water reflections and shadow recalculations—yet Netflix hasn’t released a single benchmark against or Adobe’s Firefly.

Model Fusion: OpenRouter’s ensemble AI play

OpenRouter’s Model Fusion runs multiple LLMs in parallel and merges their outputs—but skips the benchmarks proving it’s worth the complexity.

Microsoft’s Copilot disclaimer echoes psychic hotlines

Microsoft’s Copilot includes a legal disclaimer nearly identical to those used by psychic hotlines to avoid lawsuits.

Copilot’s disclaimer vs. Microsoft’s billion-dollar pitch

Microsoft’s Copilot comes with a legal disclaimer that reads like a joke: ‘for entertainment purposes only.

Gemini in Android Auto: AI Copilot or Just Another Chatbot?

Google’s Gemini lands in Android Auto with little fanfare and even fewer new features, exposing a gap between hype and reality.

Google Vids’ AI upgrade: Veo, Lyria, and the avatar hype

Google’s latest Vids upgrade packs Veo’s video synthesis, Lyria’s audio models, and a new «directable» avatar system—all repackaged as a unified creative suite.

Google’s Gemma 4: Open-source AI with a license that matters

Apache 2.0 turns Gemma 4 into the first Google AI model you can legally monetize without asking permission first.

Gemini’s ChatGPT import tool: A migration play or lock-in bait?

Google’s new Gemini import tool targets ChatGPT’s 180M users—but the fine print reveals a Workspace integration play, not true interoperability.

Gemma 4’s quiet debut: Lite models, Italian fine-tunes, and no benchmarks

Simon Willison’s notes on Gemma 4 reveal three new models—two Italian fine-tunes and a ‘Flash Lite’ preview—while Google stays silent on performance or release timelines.

Google’s 5TB AI Pro Plan: Storage or Stealth AI Lock-In?

Google’s AI Pro Plan now includes 5TB of storage—a feature absent from its standalone tiers at any price.

Google Questions Gemini Ban

Google's Gemini account ban has sparked controversy, with the company disputing a family's claim that their account was banned unfairly.

Microsoft’s MAI drops three models—just don’t call it a revolution

Six months after forming MAI, Microsoft unveiled three generative models—none with names, benchmarks, or clear paths to production.

Gemma 4: Google’s quiet play for the edge AI throne

Google’s Gemma 4 ditches cloud dependency with offline multimodal AI, but the Apache 2.0 license is the real headline.

Gemini in Your Car: AI Assistant or Google’s Latest Test?

Android Auto users are discovering Gemini in their cars without warning, raising questions about consent and real-world reliability.

Gemma 4 is open enough to build with, not open enough to explain Google’s edge

Google’s Gemma 4 family continues the open-weight strategy from Gemini technology while keeping process control firmly inside Google.

Microsoft’s Multimodal AI: More Than Just Hype?

Microsoft’s new AI models promise voice, image, and transcription capabilities—but lack names, benchmarks, or a clear release timeline.

Google’s AI Search Live: Conversation over results

Google’s Search Live replaces ten blue links with AI chat, but developers report identical retrieval snags under the slick surface.

Google’s AI Pro now packs 5TB—but who’s filling it?

Google quietly swapped AI improvements for a 5TB carrot—because 3TB of free terabytes sell faster than another subpar LLM update.

An overheated giant shows how deeply a star’s chemistry can mark a planet

IGRINS on Gemini South has measured magnesium and silicon in WASP-189b’s atmosphere, tying an extreme planet to the chemistry of its host star.

AI book bans: Right-wing groups weaponize ChatGPT for censorship

Conservative activists are now using **Google’s Gemini and OpenAI’s ChatGPT** to scan books for ‘objectionable’ content—turning AI into a censorship assembly line.

OptiMer’s trick: Tuning LLMs after training, not before

Bayesian optimization just became the secret sauce for fixing pretraining mistakes—after the fact, not before.

Gemini’s ‘vibe lighting’ is just voice commands with mood boards

Siri’s Multi-Request Trick: Finally Catching Up to 2018

Bloomberg’s sources confirm iOS 27’s Siri will mimic Google Assistant’s 2018 multi-command trick—five years after rivals made it table stakes.

Alibaba’s Qwen3.5-Omni writes code from speech—no training required

Alibaba’s latest model quietly picked up a party trick: generating functional code from spoken commands and screen recordings—without anyone explicitly teaching it how.

Gemini’s Live Translate Lands on iOS—but Who’s Listening?

Google’s real-time translation feature for headphones hits iOS three months after debuting on Android, but early adopters report inconsistent performance.

Gemini Live’s voice downgrade: AI progress or collateral damage?

Gemini Live’s once-smooth custom voices now sound like they’ve been run through a low-bitrate compressor, according to user reports and forum threads.

When AI sees a blank: models can invent medical findings without an image

Gemini’s Memory Import: Convenience or Competitive Catch-Up?

Google’s one-click memory import for Gemini arrives 14 months after ChatGPT first introduced persistent conversation history.

Claude Mythos: Benchmarks Soar, But Is This AI’s Next Reality Gap?

Leaked docs show Anthropic’s next model boasts scores 30% above Opus—but details on real-world use remain scarce.

Gemini 3.1 Flash Live: Google’s Low-Latency Hype Check

Google’s latest voice model promises ‘real-time’ multimodal interactions—but developers know demos rarely survive contact with reality.

Anthropic Wins Reprieve

Anthropic, an AI company, has secured a preliminary injunction against the US government.

Google’s AI Search Live goes global—but is it live yet?

Google’s Search Live now supports 98 languages, but performance lag raises questions about real-world readiness.

DeepMind’s AI safety play: real guardrails or just another demo?

Google DeepMind’s latest AI safety research targets manipulation risks in finance and health—but the measures remain lab-tested, not battle-ready.

A test humans solve casually shows how little AI understands the unknown

ARC-AGI-3 isn't another leaderboard polished by synthetic data—it's 135 interactive environments where AI must explore, reason, and act without instructions, while untrained humans do so with casual ease.

Google's Lyria 3 Pro finally hits full-song length

X.com's JavaScript errors block access for users with privacy extensions.

Lyria 3 Pro: More minutes, same old AI song

DeepMind’s latest music generator can now create tracks longer than 30 seconds, assuming you ignore the lack of compositional depth or real-world benchmarks.

Apple’s Gemini Distillation: On-Device AI Without the Cloud Hype

Apple’s deal with Google gives it more than access—it’s a license to build AI that works without the internet.

Gemini 3.1 Flash-Lite Unveiled

Gemini 3.1 Flash-Lite's demo showcases its ability to generate websites in real-time, with Google DeepMind highlighting its speed and efficiency.

Agile Robots Joins Google DeepMind

Agile Robots will incorporate Google DeepMind's robotics foundation models into its bots, collecting data for the AI research lab.

Deepfakes Fool Radiologists

Researchers at a leading medical institute have found that AI-generated deepfakes can fool even experienced radiologists and LLMs.

OpenAI’s $730B IPO gamble hinges on Microsoft’s goodwill

OpenAI’s $730 billion valuation assumes Microsoft will keep bankrolling its compute habit indefinitely—and investors aren’t convinced.

Slimes Get AI

Square Enix has partnered with Google to bring AI-powered chatbots to Dragon Quest X.

Apple now has to prove artificial intelligence works better when it stays on the device

Apple has scheduled WWDC2026 for June 8–12,placing artificial intelligence at the very center of its softwarestrategy under mounting pressure to catch up with competitors.

AI agents promise to plan your trip. ItinBench asks if they can survive the route

A new arXiv benchmark evaluates large language models using multi-stop itinerary tasks.

AI Chatbots Slowed Down

Sam Lavigne's Slow LLM tool has made headlines for its ability to slow down AI chatbots.

Capcom’s AI partner talk is just corporate speak for ‘we’ll use it carefully’

Capcom’s latest AI statement is less a tech breakthrough and more a masterclass in corporate hedging.

Google’s Gemini games flop: AI hype hits gamer reality

Google’s Gemini-powered games at GDC 2026 drew shrugs, not applause, proving AI’s gaming moment is still missing a killer app.

OpenAI’s superapp: packaging or power move?

OpenAI’s internal memo describes the superapp as a ‘simplification,’ but the real goal may be regaining control of a market it’s rapidly losing to Microsoft and Google.

Google’s Colab MCP server is a useful signal, but not an autonomous breakthrough

Google’s open Colab MCP server lets agents programmatically control cloud notebooks and GPU-backed runtimes.

The robot tutor that pretended less kept students closer to the lesson

A University of Chicago study with 52 students aged 9 to 10 found that robot tutors can support SEL lessons without invented emotions or backstories.

Walmart learned AI shopping cannot feel like a separate store

Walmart abandoned OpenAI's Instant Checkout after conversion rates cratered to one-third of traditional online shopping, and is now embedding its own Sparky assistant directly into ChatGPT and Google Gemini.

Google is making Gemini more personal, but the deepest memory still costs money

On June 11, 2024, Google dropped the paywall for Personal Intelligence, making the feature free for every US Google account holder.

Gemini is becoming less like a chatbot and more like Google’s memory layer

Google’s Gemini gets its most personal data layer yet: your inbox and photos

Google has flipped Personal Intelligence live for every US user, giving Gemini read access to Gmail threads and Google Photos libraries in the most aggressive personal-data integration into consumer AI to date.

Grok hallucinates less, but the AI market still buys leaderboard winners

xAI's Grok 4.20 arrives with a pair of counterintuitive claims: near-zero hallucinations and a price point that undercuts the competition—while trailing badly where it counts.

Google wants Maps to understand the trip before you know the address

Google Maps is no longer just an orientation tool — it's becoming a conversational agent that translates vague desires into concrete recommendations.

Gemini in Google Maps has one hard test: answer before you miss the turn

Google Maps is no longer just a navigation tool: the new "Ask Maps" feature, powered by Gemini, enables conversational answers to complex real-world questions — from vegetarian restaurants with available tables to places to charge your phone while sightseeing.

Claude’s memory upgrade: A real switch incentive or just catch-up?

For months, the AI chatbot wars have been fought on two fronts: raw capability and stickiness.

AI gets a desk in the U.S. Senate before the oversight rules are clear

American senators aren't just experimenting with AI at the margins anymore—it's now part of official protocol.

ChatGPT’s Pentagon deal backfires—users flee, Altman admits blunder

Google Home’s Gemini fixes: Smart home voice control finally grows up

Google Home’s Gemini update finally tames voice chaos—multi-step commands now work without a three-act negotiation.

GPT-5.3 Instant: A Quiet Upgrade That Actually Matters

OpenAI’s stealth GPT-5.3 Instant cuts ChatGPT response lag by 40% and fixes cringe replies—no PR stunt, just real gains.

AlphaGo’s real legacy began when AI left the board for the lab

A decade after DeepMind's AlphaGo crushed Lee Sedol, the real revolution wasn't the board game victory—it was the quiet migration from gaming tables to laboratory benches.

Google wants Drive to stop being storage and become an AI work layer

Google is finally bringing Gemini to Drive — two years late and behind a paywall that leaves most users out in the cold.

Gemini’s $82K API leak: Why basic guardrails are now a survival issue

Gemini API leak racked up an $82K bill in 48 hours—exposing how API economies prioritize access over accountability.

Google is putting Gemini where office work already happens

Google isn't inventing AI-powered documents—it's stitching its Gemini assistant into the seams of Workspace with surgical precision.

Google is testing how much companies will pay for AI already inside their files

On March 10, 2024, Google launched a beta of Gemini-powered tools across Workspace, turning Docs, Sheets, Slides and Drive into an experimental proving ground for generative AI — but only for those on the most expensive business plans.

GPT-5.3 Instant: the AI that finally stops gaslighting its users

GPT-5.3 Instant reduces patronizing behavior and tries to make ChatGPT feel more useful to developers and power users.

Copilot gets a second AI engine for office work that does not wait for prompts

For the first time in its flagship AI platform, Microsoft is deploying a rival to OpenAI—integrating Anthropic's Claude to autonomously execute tasks across core Office applications.

Google is turning Search into an AI workspace, not just an answer box

Google's move to embed Canvas within Google Search's AI Mode marks a shift from experimental AI to an integrated productivity tool. This could redefine user expectations for AI integration in daily digital workflows, but long-term adoption remains uncertain.

Google’s notebook tool turns sources into video, but the limits tell the bigger story

NotebookLM now offers Ultra users cinematic Video Overviews from their own sources, capped at 20 cinematic videos per day.

Wikipedia’s AI translation push exposes the new cost of checking every footnote

CollectivIQ Crowdsources AI

CollectivIQ's platform can display responses from up to 14 different AI models, including ChatGPT and Gemini.

Gemini Flash-Lite: Smarter, Faster, Pricier—But Is It Real?

Google’s Gemini 3.1 Flash-Lite promises blazing speeds but delivers a 3x cost hike, leaving developers to wonder what’s actually improved—and who’s footing the bill.

Meta wants its chatbot to turn product searches into shopping moments

Meta AI now generates shoppable product carousels with prices and website links in limited testing.

Google’s new Gemini model is built for the AI bill, not the benchmark trophy

The model enters Public Preview today on Vertex AI and Google AI Studio, targeting high-volume workloads where every millisecond of latency hits the bottom line.