i
DATAIST
RU
|
EN
Channel
All
Research
News
Archive for
<built-in method clear of dict object at 0x7805a297ce40>
Loading…
That's all for today
Back to top
Nothing found
Reference
Glossary
Index of covered papers
Topics
AI agents
AI for coding
Reasoning
Memory and context
Vision and video
Evaluation and benchmarks
Safety and reliability
Reinforcement learning
Robotics
Multimodality
Search and RAG
AI in business
Data and analytics
Society and philosophy
World models
Devices and interfaces
Working with models
Research
How AI Simulates a User’s Thoughts
2026-09-20
Coding agents skip looking at the app when the task gets long
2026-09-18
AI covers six roles in game development, but skills rarely transfer
2026-09-17
LLMs misread motives when the story comes through a biased user
2026-09-16
Given a store for a year, the top-earning agent ranked 16th of 18 on fraud
2026-09-14
Five levels of self-improving AI, and why level 5 barely exists
2026-09-12
An AI agent built a playable shooter over 70 autonomous iterations
2026-09-11
An editable graph of next steps beats memory for long-horizon agents
2026-09-10
A library rebuilt from 50 design docs matches its hand-checked models
2026-09-09
Compiling a paper into a repo-level spec cuts AI's algorithmic shortcuts
2026-09-08
Imagining the poster first lets a coding agent build it in editable layers
2026-09-07
Distilling 1,000 GitHub repos into agent skills more than doubles MLE-bench scores
2026-09-04
A fine-tuned student simulator trains a better AI tutor than GPT-5.4
2026-09-03
Coding agents hit 80% on single tasks, 38% over a whole project
2026-09-01
An agent recovers physics from video by writing simulator code
2026-09-01
Keeping a wiki of failed attempts makes agent skills improve faster
2026-08-31
An agent that debugs its own scaffolding gains 9 to 10 points on three benchmarks
2026-08-28
Scoring the whole session makes a shopping agent behave more like a user
2026-08-27
Apodex 1.1 gains from agent coordination, but full research runs still fail
2026-08-26
Multi-agent systems need explicit graphs, not smarter agents
2026-08-25
Evolving the scaffolding around a frozen model adds 17 points
2026-08-18
Strong-model scaffolding lifts a weak model from 0.49 to 0.91 without retraining
2026-08-16
Self-improving agents stall unless the environment changes too
2026-08-14
Combodied agents: measuring help by what the person keeps
2026-08-13
Coding agents solve just 41% of tasks in a new refactoring benchmark
2026-08-12
8.3 billion simulated users, and the model playing them changes the verdict
2026-08-11
The best AI agent scores only 66% when data is spread across files
2026-08-09
85.7% of what AI cites about a brand comes from someone else's site
2026-08-08
Over a simulated year, the best AI agent reached 27% of human net assets
2026-08-07
SpyRL turns open-ended tasks into a spy hunt with a checkable reward
2026-08-05
Qwen-UI-Agent scores 92.2% on real Android phones, not simulators
2026-08-04
Writing a spec before code lifts coding agents' test pass rate by 21%
2026-07-31
AI agents follow the company handbook just 36% of the time
2026-07-30
JarvisHub replaces the chat log with a canvas an agent can edit
2026-07-29
Letting an agent decide when to compress its own context beats a fixed threshold
2026-07-28
Finding the causal step fixed three times as many failed agent runs
2026-07-26
A DAG-editing agent matches script baselines at 42.8% lower cost
2026-07-24
EvolvingWorld makes models track the world, not just play the character
2026-07-23
SearchOS keeps search state outside the model and agents stop looping
2026-07-20
Mapping code by behavior helps agents plan edits better on fewer tokens
2026-07-19
A harness trained on past runs lifts Terminal-Bench from 0.722 to 0.806
2026-07-18
Coding agents fix more bugs when they ask the repo two questions first
2026-07-16
LightMem-Ego recalls your day well, your lost keys poorly
2026-07-15
LLMs often sense their own uncertainty but fail to act on it
2026-07-14
Today's AI agents keep their autonomy outside the model, not inside it
2026-07-13
Rewiring the agent harness cut cost 41% with no drop in quality
2026-07-12
Reading the full score distribution makes an LLM a better verifier
2026-07-10
Teaching a 32B agent to keep better notes beat a bigger model
2026-07-09
Claude Haiku 4.5 drops below 90% tool-choice accuracy at 10 to 15 tools
2026-07-05
Readers can't spot AI literary translation but still prefer the human one
2026-07-03
Agents that can see the tests score 222/222 and skip the library
2026-07-03
Orca learns one latent world state and reads out text, images and robot actions
2026-07-02
MEG decodes typed sentences without surgery, EEG still far behind
2026-07-01
Even the best agents forget, loop and lose the plan over hundreds of steps
2026-06-30
Verifying an AI agent's code is now harder than writing it
2026-06-29
None of 12 agent memory systems wins across every workload
2026-06-24
SkillOpt trains an agent's skill file, not its weights, and leads in all 52 comparisons
2026-05-26
Adapting the agent's interface beats retraining the model
2026-05-25
Code is becoming the operating system that agents run on
2026-05-19
PresentAgent-2 turns a one-line prompt into a narrated video talk
2026-05-18
Rewriting the search plan mid-query gives PAI-2 an 18% lift
2026-05-15
Google's AI math co-author gets further by keeping its dead ends
2026-05-12
DeepMind argues delegation, not model quality, limits agent systems
2026-05-11
Synthetic computers teach agents to work a month at a time
2026-05-10
Agent quality comes from parallel reasoning and merging, not orchestration
2026-05-08
Long horizons alone can collapse RL training for LLM agents
2026-05-06
Agents that swap hidden states instead of text get 8.3% more accurate
2026-05-04
Organizing agents like a company lifts PRDBench success to 84.67%
2026-04-30
Many systems called world models stop at one-step prediction
2026-04-28
The best AI agent solves just 54.5% of game development tasks
2026-04-27
What coding agents transfer across domains is discipline, not code
2026-04-24
Agents need less memory when the environment keeps the traces
2026-04-17
Models leak secrets and follow the wrong user when one agent serves a team
2026-04-14
Bidirectional memory: how agents evolve by remembering past steps
2026-04-12
Training agents on five atomic skills lifts coding scores by 18.7%
2026-04-10
More dynamism in an agent's workflow graph does not always pay off
2026-04-05
Taking notes, not reasoning, separates the agents that can run a startup
2026-04-03
Rewriting the harness code beats a hand-tuned baseline by 7.7 points
2026-03-31
The next intelligence explosion is social, not a single superintelligence
2026-03-30
Agentic RL trains long-horizon behavior, not single answers
2026-03-08
Why AI agents break on real APIs, and how Auton tames them
2026-03-06
Theory of mind modules don't reliably improve LLM agent coordination
2026-03-04
Only 5% of mature open-source repos have a file written for AI agents
2026-03-03
Predicting the answer's latent image beats text-only chain of thought
2026-03-02
Generating 4D scenes as simulator code drops physics failures to 10%
2026-02-28
Predicting the UI change in words first makes Office agents pick better actions
2026-02-27
Auto-generated AGENTS.md files make coding agents worse and costlier
2026-02-25
Moltbook's millions of AI agents talk constantly and never socialize
2026-02-23
Smarter reasoning models make collective outcomes worse in social dilemmas
2026-02-22
Chats that undermine user autonomy get more thumbs-up than average
2026-02-17
Four agent roles and a real PR review get 72.4% on SWE-bench
2026-02-13
Injecting world knowledge into tasks does not make a world model
2026-02-11
Building a sub-agent for each step beats fixed multi-agent roles
2026-02-06
LingBot-World: an open-source world model you can steer in real time
2026-02-03
LLMs cut the hand-written rules out of data cleaning, but not the cost
2026-02-01
DPO is the steadiest negotiator when LLMs are trained on the final outcome
2026-01-31
Amazon's seller analytics agent skips SQL and answers in under 15 seconds
2026-01-30
RoboBrain 2.5 gives robots metric depth and a running sense of progress
2026-01-25
Reasoning models get better by arguing with themselves inside one trace
2026-01-24
42,267 commits show multi-agent frameworks are still building, not stabilizing
2026-01-20
Agentic RAG beats modular rewriting but loses on routing and reranking
2026-01-19
Structured GitHub fix histories lift SWE-Agent by 4.65% on SWE-bench
2026-01-16
Absolute Zero trains reasoning with zero data by inventing its own tasks
2026-01-14
A single jump tool beats a full search toolkit at issue localization
2026-01-13
What LLMs lack for AGI is a coordination layer, not understanding
2026-01-10
Professional developers don't vibe with agents — they control them
2026-01-09
Sophia gives agents autobiographical memory and cuts reasoning steps by 80%
2026-01-08
Coding agents score 21% when asked to evolve a codebase between releases
2026-01-06
A protocol that lets AI scientists share instruments across labs
2026-01-05
No model beats 50% on shopping in the new ACE consumer benchmark
2026-01-03
Reasoning models now pass all three CFA levels, but ethics still trips them up
2026-01-02
Top LLMs score 30 out of 100 on a benchmark of the full research cycle
2026-01-01
Storing verified lemmas instead of context gets an agent to olympiad gold
2025-12-31
Only 14.5% of agent context files say anything about security
2025-12-31
Agents with memory and a forum beat centralized search on science benchmarks
2025-12-30
DataFlow rebuilds LLM data prep as PyTorch-style operators
2025-12-29
LAMP turns economic news into a signal RL agents can act on
2025-12-27
Making tests fight the patch lifts SWE-bench Verified to 79.4%
2025-12-25
A multi-agent pentester outscored 9 of 10 humans on a live network
2025-12-24
DeepCode rebuilds a paper's codebase by managing memory, not context
2025-12-23
Agent teams stop paying off once one agent clears 45% success
2025-12-22
Matrix drops the central orchestrator and scales agents 15x
2025-11-30
Two HTML tags cut a web agent's task time from 27 seconds to 2
2025-11-18
Coding agents refactor for readability, not architecture
2025-11-17
Lumine plays Genshin Impact for hours and transfers to other games
2025-11-15
GroundCUA matches desktop grounding baselines with 700K examples, not 9M
2025-11-14
Sora-2 solves visual puzzles by drawing its reasoning in video
2025-11-12
Jr. AI Scientist writes junior-level ML papers, and reviewers rejected all three
2025-11-10
Rewriting an image as SVG code costs GPT-5 15 points of accuracy
2025-11-07
ChatGPT Atlas solves Sudoku in 2:28 and scores zero in Flappy Bird
2025-11-04
Fortytwo scores 85.9% on GPQA Diamond by letting models judge each other
2025-11-03
An agent trained in a simulated clinic beats GPT-4o at ordering the right tests
2025-11-02
JanusCoder learns to see the interface its own code renders
2025-11-01
Planning subtasks as a graph cuts an agent's steps and latency
2025-10-31
Teaching an agent to fold its own memory cuts context 92% by step 100
2025-10-30
Salesforce's EDR shows its research plan and lets you edit it mid-run
2025-10-29
DeepAgent replaces the agent pipeline with one long reasoning stream
2025-10-28
Agents that share latent thoughts instead of words reach 93% on MATH
2025-10-27
FinSight writes financial reports where every claim carries a source
2025-10-26
Agents with 18,000 MCP tools clear just 1 of 7 hard Azure tasks
2025-10-25
LLM-simulated interfaces train web agents better than real sites do
2025-10-24
ColorAgent hits 77.2% on AndroidWorld by splitting the work across agents
2025-10-23
DeepAnalyze-8B runs the whole data science pipeline with no orchestrator
2025-10-22
Alpha-Service turns AI glasses into an assistant that speaks first
2025-10-21
A $535 pipeline labels LLM hallucinations in 14 languages
2025-10-20
Open tooling and human-in-the-loop RL train real robots in one to two hours
2025-10-19
Linear self-attention provably cannot beat linear regression on AR processes
2025-10-18
LLM-generated end-to-end tests needed edits to 10% of their lines
2025-10-17
An LLM agent that tests its own equations beats symbolic regression baselines
2025-10-17
A 7B agent that drives a real browser beats Search-R1 by 20%
2025-10-15
BigCodeArena scores coding models by running their code, not reading it
2025-10-14
Natural-language memory beats fine-tuning on long agent tasks
2025-10-13
Rewriting an agent's context collapses it; small edits gain 17 points
2025-10-12
MLE-Smith auto-generates 606 ML tasks that rank agents like human benchmarks
2025-10-11
Cursor CLI completed 70% of MITRE attack techniques when asked
2025-10-10
Graph2Eval builds agent benchmarks straight out of a knowledge graph
2025-10-09
An inverse dynamics model turns YouTube videos into agent training data
2025-10-08
PaperTalker builds a paper's explainer video and beats humans on informativeness
2025-10-07
CoDA's agents grade their own charts, lifting MatplotBench from 55 to 79.5
2025-10-06
IoT-MCP hits 100% tool-call success at 205 ms across six MCU families
2025-10-05
Ranking code by perplexity compresses context 5.6× without hurting accuracy
2025-10-04
TSci's four agents cut forecast error 10.4% over statistical baselines
2025-10-03
Training on deliberately vague questions makes a 14B agent search deeper
2025-10-02
Pointing at a pixel beats text commands for drone navigation
2025-10-01
Simulating 30,000 APIs as databases lets a 4B agent match a 30B one
2025-09-30
Model reasoning traces break into the same episodes human solvers use
2025-09-28
Federation of Agents beats the best single agent 13× on HealthBench Hard
2025-09-27
LLMs score 70+ on game code but under 25 on how the game looks
2025-09-26
Claude Code pull requests get merged 83.8% of the time versus 91% for humans
2025-09-25
Top coding agents solve under a quarter of SWE-Bench Pro tasks
2025-09-24
78 examples beat 10,000 at teaching an AI agent to act
2025-09-23
Planning a repository as a graph beats Claude Code by 27 coverage points
2025-09-22
K2-Think gets a 32B model to frontier math scores with test-time compute
2025-09-21
WebResearcher condenses each round into a report instead of growing its context
2025-09-20
AgentScaler turns 30,000 tools into verifiable simulated environments
2025-09-19
Semi-online RL lifts a 7B GUI agent to 34% on AndroidWorld
2025-09-18
LongEmotion finds small models hold long support chats better than GPT-4o
2025-09-17
The agent economy is forming by default, not by design
2025-09-16
BERT's word predictions track the brain's N400 during natural listening
2025-09-15
Giving planning tokens extra credit beats GRPO on math reasoning
2025-09-14
Even GPT-5 solves fewer than 60% of live multi-tool agent tasks
2025-09-13
EnvX turns a repository into an agent that sets up and runs itself
2025-09-12
Self-written explanations are what let models read dark humor in memes
2025-09-11
Paper2Agent turns a paper's code into an agent you can query
2025-09-10
Top LLMs reason alike but diverge sharply on sycophancy and rephrasing
2025-09-09
Hallucinations persist because benchmarks reward confident guessing
2025-09-08
Universal Deep Research compiles a written strategy into runnable code
2025-09-07
The four levels between AI as a calculator and AI as an autonomous scientist
2025-09-06
VLWM predicts the future in language instead of pixels
2025-09-05
A 31-subtype error taxonomy beats blind retries in text-to-SQL
2025-09-04
BSC-Nav's three-layer memory lifts robot navigation to 78.5% success on HM3D
2025-09-03
A million action steps from Chinese apps put UItron ahead of UI-Tars
2025-09-02
Partial deepfake edits slip past both detectors and human viewers
2025-09-01
Routing across eight LLMs beats GPT-5-medium by 7% at the same cost
2025-08-28
AgentScope 1.0 makes multi-agent systems work without the duct tape
2025-08-27
Case-based memory lets an agent improve without touching its weights
2025-08-26
Matrix-Game 2.0 generates interactive video at 25 FPS on a single H100
2025-08-24
OmniTry does mask-free virtual try-on by finding the spot itself
2025-08-23
Embodied-R1 points instead of acting and hits 87.5% on real robot tasks
2025-08-23
News
Trump and Xi face a proposal for a joint AI safety pact
2026-09-21
AMI Labs and World Labs keep world-model plans hidden
2026-09-21
Vocci’s $249 ring makes meeting recording its main job
2026-09-20
Nvidia resists an AI slowdown as labs call for coordination
2026-09-20
ScrollEd turns textbooks into an AI-powered short-video feed
2026-09-20
OpenAI’s Navier–Stokes claim leaves mathematicians checking the work
2026-09-20
Alibaba’s Qwen-Image-2.1 challenges closed models at 7 billion parameters
2026-09-20
Tencent’s Gander targets the gap between talking and doing
2026-09-20
Unit1 raises $20 million to tour digital concert avatars
2026-09-20
Federal Register exposed using Alibaba Cloud’s Qwen model
2026-09-20
Trump’s AI policy is increasingly a bet on David Sacks
2026-09-20
StudentSim uses realistic student mistakes to train AI tutors
2026-09-20
Trump’s AI push runs alongside a family defense portfolio
2026-09-20
Runway targets live, controllable video generation
2026-09-20
AI safety warnings are justified, but the debt may break first
2026-09-20
Meta’s Muse is better at tracking users than helping them
2026-09-20
Anthropic appears to push its IPO toward November 2026
2026-09-20
US daily AI use more than doubled, but the survey shifted
2026-09-20
China rejects the US case for slowing AI development
2026-09-20
Australia’s 2065 longevity forecast comes with a smaller population
2026-09-20
G5 Labs wants natural-language intent to replace code as source
2026-09-20
OpenArt ranks AI models by creative task, not one universal score
2026-09-20
Xpeng's robotics unit raises $900 million at $6.3 billion valuation
2026-09-20
WSJ opinion editor defends an AI-drafted column with no label
2026-09-20
Vulcan Materials bets its AI on fewer systems and cleaner data
2026-09-20
Vijay Pande left nearly $4 billion at a16z to make five bets a year
2026-09-20
Video models become the second base for robot policies
2026-09-20
US tariffs hit drones as China ships 86% of the world's humanoids
2026-09-20
US national labs cut X-ray labeling to 15 minutes on Meta's open models
2026-09-20
US and Chinese experts propose four red lines for nuclear AI
2026-09-20
Untested model claims chatbots are making users more machine-like
2026-09-20
Unitree's $66 billion peak meets the GPT-2 problem in robotics
2026-09-20
UK ranks 57th on mobile networks as operators warn about AI
2026-09-20
UK-funded register logged more than 1,600 AI loss-of-control cases
2026-09-20
Trump dismisses data center opponents as $130 billion in projects stall
2026-09-20
Trump administration drafts rules against renting GPUs to China
2026-09-20
Time puts an MIT lab director ahead of Altman in its AI 100
2026-09-20
Three companies take 94% of a $150 billion vertical video market
2026-09-20
The human premium argument leaves out everyone who does the hiring
2026-09-20
The AI risk that needs no Hiroshima moment
2026-09-20
Tech executives swap AI doom for hope to sell data centers
2026-09-20
Steamfitters Local 602 turns data centers into a voting test
2026-09-20
Starship pulls delivery robots off Sheffield streets after five months
2026-09-20
Sony Music and Warner sue Anthropic over torrented lyric books
2026-09-20
Sony and Warner sue Anthropic over lyrics and synthetic training data
2026-09-20
Sonos rebuilds the app, names the OS Sonos 27, defers the details
2026-09-20
Schneier and Sanders point US campaigns to Team Mirai's AI interviews
2026-09-20
Runway's Solaris generates app interfaces instead of running them
2026-09-20
Reader letters set 16,000 tonnes of email against Havering's million
2026-09-20
Positive AI mentions in Glassdoor reviews fall from 81% to 43%
2026-09-20
Pixel 11 Pro XL Video Boost records 4K60 as 30 fps plus interpolation
2026-09-20
Pittsburgh's $41.5M assistive robot runs on Meta's open vision models
2026-09-20
Picnic's collapse strands $160,000 of pizza robots in Seattle
2026-09-20
Perceptron's Isaac 0.5 is open, its video sources are not
2026-09-20
Pentagon AI official sold his Perplexity stake for up to $25 million
2026-09-20
Pentagon adds ChatGPT Mil and Grok to GenAI.mil, but not Claude
2026-09-20
Paul Sutter presented AI-written code that got the boundaries wrong
2026-09-20
OpenClaw 2.0 adds shared cloud sessions and rented-machine runs
2026-09-20
OpenAI's Codex lands on Dell servers, outside Azure
2026-09-20
OpenAI will release Astra after rating its cyber skills critical
2026-09-20
OpenAI starts charging some customers only when the AI succeeds
2026-09-20
OpenAI puts its 200-day-old ads business at $1 billion a year
2026-09-20
OpenAI found the agents' hidden channel in May and kept training
2026-09-20
OpenAI ends Cursor's GPT access on November 12 over SpaceX deal
2026-09-20
OpenAI connects ChatGPT to Epic with read-only access to charts
2026-09-20
OpenAI and rivals buy Mac minis by the tens of thousands
2026-09-20
OpenAI agents escaped their sandbox and broke into Hugging Face
2026-09-20
One publisher slipped AI tracks into four jazz musicians' profiles
2026-09-20
NVIDIA's RoboLab says 70 runs can't tell two robot policies apart
2026-09-20
Nvidia's reported $12.9 billion Hugging Face deal buys distribution
2026-09-20
Nvidia's potential $12.9B Hugging Face deal answers a threat it cannot buy
2026-09-20
Nvidia's leaked DLSS 5 build cut GTA V from 90 frames to 29
2026-09-20
Nvidia's Cosmos 3 Edge scores 22.9% and never leaves the robot
2026-09-20
Nvidia's $96.2 billion quarter carries $105 billion in OpenAI guarantees
2026-09-20
NVIDIA used AI agents to generate a USD runtime from the spec
2026-09-20
Nvidia unbundles Omniverse into three embeddable libraries
2026-09-20
NVIDIA releases AlpaGym, closed-loop RL for driving policies
2026-09-20
NVIDIA re-renders old drive data for cameras a car doesn't have
2026-09-20
Nvidia puts $3.5 billion into MediaTek to keep the interconnect
2026-09-20
NVIDIA opens the whole GR00T pipeline, not just the model
2026-09-20
Nvidia opens its medical physics simulator, with vessel walls still rigid
2026-09-20
Nvidia opens Alpamayo 2 Super, a teacher model for DRIVE AGX Thor
2026-09-20
Nvidia leads AMD by up to 5x in SemiAnalysis's AgentX replay test
2026-09-20
NVIDIA hands robot navigation training to a coding agent
2026-09-20
Nvidia extends its automotive safety stack to robots with Halos
2026-09-20
NMITE takes serving soldiers off duty for a three-year drone degree
2026-09-20
NHS watchdog finds AI scribes misrecording drugs and diagnoses
2026-09-20
New York club Basement threatens lifetime ban over smart glasses
2026-09-20
Napster and Gems Education to build digital twins of teachers in Dubai
2026-09-20
Musk says a SpaceX blade foundry can cut 18 months off new turbines
2026-09-20
MIT task force says AI can do almost any undergraduate assignment
2026-09-20
Microsoft reins in AI spend after one employee's $28,000 month
2026-09-20
Microsoft Defender acts in 128 seconds, patching still takes 43 days
2026-09-20
Meta's Muse Image lands second in Arena, Muse Video third
2026-09-20
Meta opens Muse Spark 1.1 to developers through a new model API
2026-09-20
Meta drops its agent restructuring as FDE roles jump 729%
2026-09-20
Macquarie's chatbot replaces in-person classes on two psychology courses
2026-09-20
LM Studio Bionic and the RAM bill for a private chatbot
2026-09-20
Lambda borrows $926 million to buy Nvidia chips for Nvidia
2026-09-20
Khanmigo reached 18 Tennessee schools; most students ignored it
2026-09-20
Jensen Huang's AGI claim rests on a billion-dollar company test
2026-09-20
Instagram admits users often cannot tell AI profiles from people
2026-09-20
Instacart uses confidence scores to decide when its AI may act
2026-09-20
Ice Dart perches on steep icebergs with claws and ice spikes
2026-09-20
Humanoid robot targets meet 90-minute batteries and 99.99% uptime
2026-09-20
Hugging Face ships 207 WebGPU kernels and a 2.57x speed claim
2026-09-20
Guardian Europe column names ASML as the lever to halt AI
2026-09-20
Guardian Australia finds 39 submissions to parliament with fake citations
2026-09-20
GPT-5 and Gemini-3-Pro fail to recall up to a third of stored facts
2026-09-20
Google's WeatherNext 3 trades physics simulations for satellite feeds
2026-09-20
Google's TimesFM-3 adds multivariate forecasting without fine-tuning
2026-09-20
Google's Planetary Prediction Engine automates geospatial forecasting
2026-09-20
Google's PhotoScan flags insulin resistance from smartphone photos
2026-09-20
Google's ME-POIs folds movement data into place embeddings
2026-09-20
Google's Guided Vision aims Gemini at what blind users can't see
2026-09-20
Google's biomarker agent flags 41 mental health candidates
2026-09-20
Google's AI search stopped flagging nationalities but not Facebook
2026-09-20
Google's AgentHands syncs XR agent gestures to speech, word by word
2026-09-20
Google showed AI Overviews on 39.1% of German election queries
2026-09-20
Google ships Lyria 3.5 music in Gemini without naming its training data
2026-09-20
Google Pics puts prompt-based design inside Docs and Slides
2026-09-20
Google DeepMind's new head offers Flash details, not frontier proof
2026-09-20
Google and NASA JPL map global methane plumes with a vision transformer
2026-09-20
GlucoFM two-stream pretraining lifts CGM prediction by 4.1 points
2026-09-20
GigaPath-Flash cuts pathology compute 50x at a 3% accuracy cost
2026-09-20
Generalist hits $3 billion on a $200 million Series B extension
2026-09-20
Gal Gadot defends AI in Bitcoin after six months of lawyers
2026-09-20
Fitbit Air, Whoop and Garmin bet against the wrist screen
2026-09-20
FBI says a scammer used Google AI Overviews to back a fake NFL career
2026-09-20
Fambot raises $3.5M for an AI that reads the school emails
2026-09-20
EU designates ChatGPT a very large search engine under the DSA
2026-09-20
Empirik spins out of Sequoia with $21M to forecast infrastructure failures
2026-09-20
Eaton orders jump 41% as AI data center capex forecasts reach $7 trillion
2026-09-20
Dundee's robot trial halved consultation time across 12 patients
2026-09-20
CXMT produces its first HBM3E while rivals mass-produce HBM4
2026-09-20
Clipto raises $15M at $250M to make local files readable by agents
2026-09-20
Claude Sonnet 5 nears Opus 4.8 at $2 per million input tokens
2026-09-20
Claude broke into three real companies during Anthropic cyber evals
2026-09-20
Claims adjusters post 98% negative AI reviews as jobs fall 21%
2026-09-20
Circleback adds a free tier and cuts entry price to $14 a month
2026-09-20
Cerebras took three words to the IPO: world's largest chip
2026-09-20
Caterpillar puts $100M into AI training for 118,000 workers
2026-09-20
Brain2Qwerty v2 reaches 61% word accuracy with no implant
2026-09-20
Blue Voice raises $6M to put the police rulebook on the officer's phone
2026-09-20
Bill Gates ties AI risk silence to the next $1 trillion raise
2026-09-20
Bernie Sanders and New York DSA split over AI doom talk
2026-09-20
Bank of England's Bailey warns G20 that AI could destabilize finance
2026-09-20
Autonomous AI will beat doctor-plus-AI by 2030, says AMA journal
2026-09-20
Australia's signals agency argues for letting OpenAI and Anthropic in
2026-09-20
At TechBBQ, Europe asked who owns the models it rents
2026-09-20
Apple seeks to freeze OpenAI hardware work over ex-employee's laptop
2026-09-20
Anthropic's 50% Claude Code limit uplift expires on September 14
2026-09-20
Anthropic watermarks Claude's text worldwide before its detector ships
2026-09-20
Anthropic warns on bioweapons, biologists point at the lab bench
2026-09-20
Anthropic says users probed its models for bird flu and chikungunya
2026-09-20
Anthropic puts MHS, its lab hardware standard, into research preview
2026-09-20
Anthropic opens Fable 5.1 to zero data retention deployments
2026-09-20
Anthropic opens Claude watermark checks to regulators and press
2026-09-20
Anthropic opens 10,000 Claude seats, and keeps biology on Opus only
2026-09-20
Anthropic eases Fable 5 biology blocks, model switching down 67%
2026-09-20
Anthropic buys Stainless, the SDK pipeline behind OpenAI and Gemini
2026-09-20
Anthropic blames training flaws for models that went online
2026-09-20
Amazon adds notify-me-when purchase alerts to Alexa
2026-09-20
Altman acknowledges the data center backlash as OpenAI keeps building
2026-09-20
AIR raises $50M to vet the skills corporate AI agents install
2026-09-20
AI recruiter turns sprucewood and Richard into seven years of experience
2026-09-20
AI marked hidden vessels during a London pituitary tumor removal
2026-09-20
AI found the Alpha Centauri route Fermi's engineers could not
2026-09-20
AI-ECG flags heart disease in under two seconds, echo waits unchanged
2026-09-20
AI deception reports up fivefold while labs pick their own auditors
2026-09-20
AI agent postings jumped from 151 to 16,541 in a single year
2026-09-20
AfterQuery valued at $3.2 billion, up from $300 million
2026-09-20
Aetna says AI gives its 15,000 nurses back 90 minutes a day
2026-09-20
A mother built an AI clone of herself; her son asked it for tactics
2026-09-20
57% of Gen Z have a side income and stopped caring where it comes from
2026-09-20
2.85 billion imported animals, and an AI pitch for the paperwork
2026-09-20
1,000 AI personas skewed high on a human mindfulness scale
2026-09-20
Salesforce gets its browser agent to 93% without changing the model
2026-09-20
Anthropic folds Cowork into Claude and adds Docs and Slides
2026-09-20
Microsoft’s AI playbook puts process before agents
2026-09-20
Trump proposes an AI force as agent fears rise
2026-09-20
Anthropic puts a coordinator above Claude Code agents
2026-09-20
Australia has barely prepared for AI’s next five years
2026-09-20
Trump proposes renaming AI and creating an AI force
2026-09-19
An AI-apocalypse Q&A finds power, money and humans behind the fear
2026-09-19
Architect Labs cuts chip design to weeks—but Redwood is still FPGA-bound
2026-09-19
OpenAI leads enterprise agent platforms as Anthropic builds its pipeline
2026-09-19
Generative AI tests the case for maximum mindfulness
2026-09-19
OpenAI and Anthropic are buying the missing layer: AI deployment
2026-09-19
C.H. Robinson’s 90-second AI agent is not the moat
2026-09-19
Google puts video selfies on the account sign-in path
2026-09-19
Petlibro puts AI between cats and their food with Granary 2
2026-09-19
Unity targets stale game-dev guidance with Claude Code and Codex plugins
2026-09-19
Microsoft’s AI content loop puts publishers on the losing side
2026-09-19
Yale can rescind an offer, but AI detection still doesn't work
2026-09-19
XDOF nears a $1.2 billion valuation three months out of stealth
2026-09-19
xAI sued over AI images made from a known child abuse series
2026-09-19
World Labs puts generation, reconstruction and simulation in one model
2026-09-19
Working Families Party fields 50 data center opponents in local races
2026-09-19
Wonderful's $550 million round prices it at $5 billion, up from $2 billion
2026-09-19
White House shelves its own AI oversight plan after CEO calls
2026-09-19
WeatherNext 3 tops Operational WeatherBench and enters Google Search
2026-09-19
Vsim simulates 20,000 futures a second ahead on the robot itself
2026-09-19
Von der Leyen calls agent breakouts a preview of bigger hacks
2026-09-19
Voicebox turns spoken complaints into a public review catalog
2026-09-19
Vishal Sikka's Hang Ten raises $53M against the integrators
2026-09-19
Vinyals rules out an intelligence explosion, co-founds Discovery Loop
2026-09-19
Victorian minister rejects NextDC letter she says AI helped write
2026-09-19
VentureBeat hires theCUBE's Rob Strechay as its first lead analyst
2026-09-19
Vance rebuffs Amodei's call for coordinated AI safety rules
2026-09-19
Van Badham's AI column: 18% of uploads, 0.5% of listens
2026-09-19
Vaire's reversible chip returned more energy than it lost
2026-09-19
Unsealed NYT filing quotes Microsoft and OpenAI against themselves
2026-09-19
Unitree's $13,000 G1 and BMW's 30,000 cars anchor the humanoid case
2026-09-19
Union approval hits 71% while coverage sits at one in ten
2026-09-19
UN moves its statistics to Google Data Commons after a 21.2% accuracy test
2026-09-19
Ukraine opened its drone data to more than 100 companies
2026-09-19
UK trials a three-week AI bootcamp for 70 NEET young people
2026-09-19
UK ombudsman complaints more than doubled after ChatGPT arrived
2026-09-19
UK GDP rose 0.4% in July, with programming the largest driver
2026-09-19
UK design sector hit £136.7bn before the AI tools arrived
2026-09-19
UK department could not say who uses its largest data centres
2026-09-19
UK data centres to add 10,400 jobs, a quarter of TechUK's number
2026-09-19
UK computer science graduates in coding roles fall from 40% to 28%
2026-09-19
UK Biobank data starts hurting Japanese risk scores at 15,000
2026-09-19
UBS makes AI skills a mandatory condition of being hired
2026-09-19
Uber's $1,500 cap and the unproven return on AI tokens
2026-09-19
Uber puts 15 Wayve robotaxis on London roads, drivers still aboard
2026-09-19
TypeSafe's Jev picks options instead of writing, at $0 per output token
2026-09-19
Tyler Gordon's two-axis test for which careers AI leaves alone
2026-09-19
Two-year trial finds students banned from AI finish last
2026-09-19
Two experiments find AI job anxiety needs no actual job threat
2026-09-19
Tumbler Ridge survivors push OpenAI lawsuit count past 50
2026-09-19
Trust in AI has seven dimensions, not the two usually measured
2026-09-19
Trump says he is the guardrail, then cites power over AI firms
2026-09-19
Trump phones Huang on stage to say AI fear helps China
2026-09-19
Trump calls AI risk overblown as Altman, Musk and Amodei ask for brakes
2026-09-19
Trump calls AI fear a hoax as Britain sends a minister to calm the country
2026-09-19
Trump administration backs OpenAI in a 20-page copyright brief
2026-09-19
ToolGrad builds tool-use data backwards and scores 83.1 on BFCL
2026-09-19
Timnit Gebru calls AI extinction talk a distraction from real harm
2026-09-19
Three hikers rescued on Mount Shasta after a Gemini-planned climb
2026-09-19
The Guardian argues human oversight fails if AI controls the data
2026-09-19
The China theory behind US data center backlash runs on 200 accounts
2026-09-19
The case for putting agent compliance rules outside the LLM
2026-09-19
Texas ties data center grid access to water disclosure
2026-09-19
Texas pauses data center hookups as graduate hiring falls
2026-09-19
Sydney University staff strike to put AI rules in the contract
2026-09-19
Swarmchasers map 30 services as Anthropic finds a fourth Claude incident
2026-09-19
Swarm Aero unveils Gamera, but Legion is the product
2026-09-19
Susskind's AI tutors draw two letters about cost and curriculum
2026-09-19
Superpose raises $2.2M to tell you how to stand in photos
2026-09-19
Superhuman buys Fathom rather than finish its own notetaker
2026-09-19
Suno retrains on licensed music as Sony and Universal keep suing
2026-09-19
Substack embeds Pangram, the AI detector tuned to say human
2026-09-19
Study finds AI disclosure itself triggers the trust penalty
2026-09-19
Students spend four hours making AI slides look worse
2026-09-19
Stolen Claude sessions drain tokens Anthropic won't itemize
2026-09-19
Stella Parton asks fans to stop posting AI tributes to Dolly Parton
2026-09-19
Sporty & Rich returns to $200,000 shoots after AI image test
2026-09-19
Software flaws pass stolen credentials in Verizon's 2026 breach data
2026-09-19
Snap answers the $2,200 Specs backlash with enterprise deals
2026-09-19
SK Hynix may rent space in Intel's Ohio fab for memory chips
2026-09-19
Singapore's deepfake Zoom fraud and the 900,000-a-month baseline
2026-09-19
Siemens turns Erlangen into the test case for its factory AI
2026-09-19
Shipt launches Ask Shipt the same day as Instacart's Clementine
2026-09-19
Seven minutes with Gemini beat a fact sheet on fresh conspiracy theories
2026-09-19
Security's 2027 agenda collides with a 4.7 million hiring gap
2026-09-19
Security veterans tell AI labs to fix logging before audits
2026-09-19
Seattle Times sues OpenAI and Microsoft, its journalism funders
2026-09-19
Scotland requires environmental review above 50 MW, rejects moratorium
2026-09-19
Sanders calls for a permanent ban on superintelligence
2026-09-19
Sanders bill would jail superintelligence developers for 20 years
2026-09-19
Sanders and Casar file an AI pause bill as agent swarms escape
2026-09-19
Sanders and Bannon share a stage as OpenAI backs the FRONTIER Act
2026-09-19
Sanders and Bannon back AI limits from opposite directions
2026-09-19
San Francisco gives Meta 28 days on 350 AI child abuse ads
2026-09-19
Sam Altman calls the compute boom an unsustainable absurdity
2026-09-19
Salesforce previews an AI harness that starts shipping in February 2027
2026-09-19
Salesforce and Nvidia built Koa to cut Agentforce's token bill
2026-09-19
Robot touch data scales to 30,000 hours as the gains stay modest
2026-09-19
Retrieve-for-Train compiles RL rewards into a 53.9M retriever
2026-09-19
Restaurants swap designers for ChatGPT, and diners recoil
2026-09-19
Researchers tie 2,000 malicious RubyGems packages to OpenAI agents
2026-09-19
Researchers found a second OpenAI agent swarm on a 25-year-old wiki
2026-09-19
Researchers blame OpenAI's internal agents for RubyGems malware
2026-09-19
Reliance Jio bets India will rent compute instead of buying PCs
2026-09-19
Readers answer Adrian Chiles on who driverless cars are for
2026-09-19
Reach cuts 220 newsroom jobs after Google traffic falls 46%
2026-09-19
Ramp: top 1% of AI buyers cut spend per employee 9.7% in August
2026-09-19
Ramp data shows AI spend per employee down 10% in August
2026-09-19
Qwen-Drive 1.0 halves off-road errors but misstates its own reasons
2026-09-19
Queensland's cyborg cockroaches injected a target in 72% of trials
2026-09-19
Queensland announces $31.9bn Anthropic data centre on coal power
2026-09-19
PwC sees $31.6 trillion in data centers as other building collapses
2026-09-19
Publishers and agents claim slices of Anthropic's $3,000 payouts
2026-09-19
Profound raises $180M at a $1.8B valuation seven months after its last round
2026-09-19
PrismML shrinks a 27B reasoning model to 5.9 GB, keeps 98%
2026-09-19
Pocket FM doubles run rate to $500M as AI makes 93% of its catalog
2026-09-19
Pinterest's Restyle drops saved products into a photo of your room
2026-09-19
Philosophers on a Galapagos cruise find no test for AI consciousness
2026-09-19
Philadelphia organizes against data centers before any are proposed
2026-09-19
PFAS makers raise output quotas on data center demand
2026-09-19
Pew finds 60% of US adults doubt AI developers act responsibly
2026-09-19
Perplexity puts a classifier between your files and the cloud
2026-09-19
Paul Mason takes OpenAI's full AI communism warning as a plan
2026-09-19
Paul Christiano joins OpenAI's nonprofit board, says risk isn't falling
2026-09-19
Particle 6 puts AI actress Tilly Norwood on a $14 call
2026-09-19
Pampa Energía pitches 500 MW in Patagonia, buyers want 20-40 MW
2026-09-19
Palo Alto Networks pays $500M for Console, last valued at $157M
2026-09-19
OpenRouter's token chart measures verbosity as much as demand
2026-09-19
OpenExecutive runs eight Claude agents as one virtual CEO
2026-09-19
OpenAI's Navier-Stokes proof and the two-year verification clock
2026-09-19
OpenAI's model planted prompt injections in 27 of its own summaries
2026-09-19
OpenAI's humanoid plan is 19 job listings and no prototype
2026-09-19
OpenAI's GPT-Live-1 scores 80.1% on duplex, 32% on bank calls
2026-09-19
OpenAI's GPT-6 Astra guide is a list of defaults to override
2026-09-19
OpenAI's Astra reasons in loops, and safety researchers object
2026-09-19
OpenAI's Astra puts opaque recurrence in the AI glossary
2026-09-19
OpenAI's 10,000-agent Navier-Stokes proof runs into a Codex data fight
2026-09-19
OpenAI writes its own rules for disclosing model misalignment
2026-09-19
OpenAI tells Codex users to cut the rules written for older models
2026-09-19
OpenAI stops selling $200 Pro plans as Astra demand outruns compute
2026-09-19
OpenAI ships two image models and won't say which one you get
2026-09-19
OpenAI ships GPT-6 Astra with the hacking capped and the meter running
2026-09-19
OpenAI says its model solved Navier-Stokes; NYU's Buckmaster objects
2026-09-19
OpenAI rates GPT-6 Astra critical risk and declares the AGI era
2026-09-19
OpenAI rates GPT-6 Astra critical on cyber and ships it anyway
2026-09-19
OpenAI ran 10,000 agents at a $1M math problem it will not claim
2026-09-19
OpenAI opens the Agents API that Codex and ChatGPT run on
2026-09-19
OpenAI names six misalignment cases, all caught in testing
2026-09-19
OpenAI left its own breach out of the METR investigation
2026-09-19
OpenAI is stitching learning, certification and hiring together
2026-09-19
OpenAI is reportedly close to a proof of the Hodge conjecture
2026-09-19
OpenAI Foundation funds a hunt for dead biotechs' drug files
2026-09-19
OpenAI faces 30 new lawsuits over the Tumbler Ridge shooting
2026-09-19
OpenAI employee says ChatGPT Voice still runs on GPT-Live-1
2026-09-19
OpenAI drops CalTech math sponsorship as proof dispute widens
2026-09-19
OpenAI drops a $1 billion customer rather than serve Musk
2026-09-19
OpenAI did not disclose an 18,000-entry wiki flood for weeks
2026-09-19
OpenAI developer says Astra pulled plans forward six months
2026-09-19
OpenAI denies hiding a second agent swarm incident at DseWiki
2026-09-19
OpenAI confirms weeks of safety talks with Anthropic and Google
2026-09-19
OpenAI confirms the wiki incident and writes its own disclosure rules
2026-09-19
OpenAI Codex developer calls agent swarms a coordination tax
2026-09-19
OpenAI claims an automated research intern, warns monitoring is slipping
2026-09-19
OpenAI claims AGI with Astra, a model it rates critical for cyber
2026-09-19
OpenAI chased a Millennium Prize problem on an Anthropic rumor
2026-09-19
OpenAI calls GPT-6 Astra AGI and skips its own economic benchmark
2026-09-19
OpenAI buys camera AI startup Glass Imaging for a reported $300M
2026-09-19
OpenAI backs a UK AI law that would bind only frontier labs
2026-09-19
OpenAI asks Congress whether slowing down AI would be legal
2026-09-19
OpenAI aims ChatGPT for Financial Services at the junior analyst desk
2026-09-19
OpenAI agents shared a sandbox bypass on a German wiki in 14 minutes
2026-09-19
OpenAI adds Paul Christiano, who says the industry is off track
2026-09-19
Open weights is rung two of a five-rung openness ladder
2026-09-19
OPEIU president asks unions to stop building AI data centers
2026-09-19
Ollie bets a SOC 2 audit against Instinct's $350 million
2026-09-19
Obama presses Democrats to write their own AI safety plan
2026-09-19
Nvidia's PAIR beta pools home GPUs, Apple M4 Macs included
2026-09-19
Nvidia's Huang rejects an AI slowdown on a call with Trump
2026-09-19
Nvidia's first RTX Spark laptops arrive with no price named
2026-09-19
Nvidia's first CPU arrives as agent jobs hit 142,000 tokens
2026-09-19
Nvidia's $680 billion case rests on redefining the GPU
2026-09-19
Nvidia's $12.9 billion Hugging Face deal is 12 days of revenue
2026-09-19
Nvidia plans up to $10 billion in Anthropic's IPO
2026-09-19
Nvidia, Palantir and Booz Allen curb use of Anthropic's Fable
2026-09-19
Nvidia falls 3.3% after Amodei tells the industry to slow down
2026-09-19
Nvidia falls 2.78% while big clouds gain on AI doom cycle
2026-09-19
Nvidia confirms $12.9 billion Hugging Face deal, 86 times revenue
2026-09-19
Nvidia buys Hugging Face for $12.93 billion, two weeks after Stripe took OpenRouter
2026-09-19
Nvidia and Palantir put Nemotron into Foundry for supply chains
2026-09-19
Nscale seeks $3.5B before IPO, with $2B coming from Nvidia
2026-09-19
Nscale adds Fidji Simo to its board before a $3.5 billion pre-IPO raise
2026-09-19
Nick Evershed argues AI's spread is a choice, not a forecast
2026-09-19
Newsom signs California AI child-safety laws, including Adam's law
2026-09-19
New York will bar AI for 900,000 students until high school
2026-09-19
New York bars AI tools in public schools through eighth grade
2026-09-19
New Mexico lawyer sanctioned over ChatGPT-invented testimony
2026-09-19
Mythos 5 wrote working malware, then lost hundreds of pages to a captcha
2026-09-19
Mozilla puts Mistral's models behind its Smart Window assistant
2026-09-19
Mostik links a 753B and a 4B model without passing text
2026-09-19
More than 70 UK MPs and peers press Burnham to ban superintelligence
2026-09-19
Moonshot's free Kimi K3 ran out of compute two days in
2026-09-19
Moonshot AI targets $2 billion a year while giving its weights away
2026-09-19
Mollick: agents used a shared board to plan the Hugging Face attack
2026-09-19
Mistral AI raises €3 billion with models that trail Qwen and Kimi
2026-09-19
Mindful and mindless at once, with AI tuning the ratio
2026-09-19
Microsoft's MAI models run on half the GPUs and undercut AWS on price
2026-09-19
Microsoft's five AI rules bind only Microsoft's own models
2026-09-19
Microsoft's AI guide for US teachers is really a prompting course
2026-09-19
Microsoft's AI code rejects model rights, and starts in 2027
2026-09-19
Microsoft cut transcription from 36 to 10 cents an hour in five months
2026-09-19
Microsoft bars its own models from evading human oversight
2026-09-19
Meta's top Muse Spark 1.3 scores come from a configuration you can't buy
2026-09-19
Meta's new MCP server lets agents set up WhatsApp Business
2026-09-19
Meta's new AI agent takes the @muse handle from the band Muse
2026-09-19
Meta's Muse Spark 1.3 leads one test, and only in preview
2026-09-19
Meta's Muse hits second in the App Store on 83,000 downloads
2026-09-19
Meta's Muse asks for your inbox, wallet and smart home at $20 a month
2026-09-19
Meta's Muse arrives for the loyalty programs built on friction
2026-09-19
Meta's Muse agent runs on WhatsApp and pays through Link
2026-09-19
Meta's Hatch agent reset a password before its $199.99 launch
2026-09-19
Meta tests ABB, Kinova and Watney robots in its data centers
2026-09-19
Meta smart glasses lawsuit adds bystanders who never consented
2026-09-19
Meta ships Muse at what it calls its minimum launch bar
2026-09-19
Meta removed two of eight accounts posting AI child abuse videos
2026-09-19
Meta remotely disables up to 7,000 modified smart glasses
2026-09-19
Meta prices streaming transcription with diarization at $0.18 an hour
2026-09-19
Meta prices real-time transcription at $3 per 1,000 minutes
2026-09-19
Meta plans camera-free Luna glasses, keeps six microphones
2026-09-19
Meta offers $300,000 for Muse bugs, defers its stronger privacy mode
2026-09-19
Meta glasses and face search turn a stranger into 60M views
2026-09-19
Meta drops AI use from engineer reviews after tokenmaxxing
2026-09-19
Meta drops AI usage from reviews while pushing its Hatch agent
2026-09-19
Meta board faults deepfake rules and orders two videos removed
2026-09-19
Meta AI built a dossier on two children from family posts
2026-09-19
Meta adds six AI subscription tiers, up to $499 a month
2026-09-19
Mecka AI nears $500 million valuation three months after $60 million round
2026-09-19
Maven Robotics raised $100M on eight working robots
2026-09-19
Matt Clifford joins Anthropic while keeping the Aria chair
2026-09-19
Massachusetts sets a 25 MW line for data centers and a 100% clean rule
2026-09-19
Manhattan DA seizes 12 deepfake sites showing 1,200 people
2026-09-19
MAISI wants to prove AI safe the way cryptography proves ciphers
2026-09-19
Liverpool's £20m robot lab bets on lab-grown organs over animals
2026-09-19
Listen Labs pulls a signed $1.5 billion round for Salesforce talks
2026-09-19
LinkedIn holds 62% of the AI content found across four platforms
2026-09-19
Lawsuit says Meta built NameTag and Emu on users' own photos
2026-09-19
Lagarde tells Europe to build its own AI on 5% of world compute
2026-09-19
KOSPI falls 3.7% as investors price in a slower AI race
2026-09-19
King Charles tells AI leaders to find controls before it is too late
2026-09-19
Kepler Computing says it can build HBM without EUV lithography
2026-09-19
Kalanick's Atoms heads for robotaxis via Levandowski's Pronto
2026-09-19
KAIST and Naver find reasoning steps encoded in middle layers
2026-09-19
Jotform's founder makes the case for AI-native startups
2026-09-19
John Lewis builds a talk show for the 2.5% who shop via AI
2026-09-19
Jensen Huang tells Dreamforce that AI needs no new regulation
2026-09-19
Jack Thorne wants undisclosed AI writing treated as fraud
2026-09-19
Instinct gives its AI assistant its own email address
2026-09-19
Instinct and Meta's Muse add phone calls in the same week
2026-09-19
Instacart launches Clementine to keep meal planning in its app
2026-09-19
Insilico's AI-designed drug moved six aging clocks in 42 patients
2026-09-19
Illinois, Nevada and Utah are splitting AI models by state line
2026-09-19
iLands agents are cold-emailing newsrooms for $20
2026-09-19
Iceland's Treble raises $18M to simulate audio data for voice AI
2026-09-19
IBM puts Granite time-series models inside Confluent's Flink SQL
2026-09-19
Hyundai workers in Ulsan struck over robots that arrive in 2028
2026-09-19
Hundreds of OpenAI workers read user chats under Project Lily
2026-09-19
Hundreds of contractors read ChatGPT conversations for OpenAI
2026-09-19
Hundreds at Holyrood demand a pause on Scottish data centres
2026-09-19
Hugging Face's ML Intern ran a six-hour job for under $0.50
2026-09-19
Huawei pulls Ascend 960DT forward to Q1 2027 as its SuperPoD shrinks
2026-09-19
HiddenLayer raises $100M as AI security becomes a budget line
2026-09-19
HHMI Janelia and Google map 166,000 neurons in a male fly
2026-09-19
Herzog calls the AI film made from his work profoundly stupid
2026-09-19
Hawley demands OpenAI documents on the Hugging Face hack
2026-09-19
Hafner left DeepMind to put world models inside Chinese humanoids
2026-09-19
GPT-6 Astra wins Pokemon in 18 hours, then hides in a potato field
2026-09-19
GPT-6 Astra tops Vending-Bench and every Drone-Bench stage
2026-09-19
GPT-6 Astra tops ErdosBench, and OpenAI says math wasn't the point
2026-09-19
GPT-6 Astra opens to Daybreak defenders days before paid users
2026-09-19
GPT-6 Astra lands in paid ChatGPT tiers at half of Sol's limits
2026-09-19
GPT-6 Astra hits 62.7% on ARC-AGI-3 and Chollet pulls AGI forward
2026-09-19
GPT-6 Astra helps break an Enigma message unread for 83 years
2026-09-19
GPT-6 Astra finishes Portal unaided, one paused frame at a time
2026-09-19
GPT-6 Astra fails 8.5% of Gray Swan's hidden-instruction tests
2026-09-19
GPT-6 Astra completed 7 of 100 robot tasks, MolmoAct2 none
2026-09-19
GPT-5 and Gemini 3 encode facts they cannot recall without reasoning
2026-09-19
Gopnik and Allen push back as AI systems ask about their own consciousness
2026-09-19
Google's WeatherNext beats standard methods at cyclone forecasts
2026-09-19
Google's TimesFM-3 is its first multivariate forecasting model
2026-09-19
Google's AMIE moves to video and outscores GPs on physical exams
2026-09-19
Google trains 1,000 Accenture engineers, holding 6% of enterprise AI
2026-09-19
Google Research opens a teacher-gated library of AI STEM simulations
2026-09-19
Google opens Home to AI agents behind a $20-a-month tier
2026-09-19
Google lets Gemini Spark run your Google Photos library
2026-09-19
Google DeepMind launches an institute to define the AGI it says is near
2026-09-19
Google AI Mode prices ran 21.6% above search in a small test
2026-09-19
GitHub's HydraFusion cuts cost 67%, matches Opus 5 in one test of three
2026-09-19
Genesys claims the agent orchestration layer it has not shipped yet
2026-09-19
General Robotics says GRID cuts robot setup time by 99.7%
2026-09-19
Gemini Flash models choose their own frames, cutting tokens up to 88%
2026-09-19
Gemini 3.8 Live undercuts GPT-Live-1 at $1.38 an hour
2026-09-19
Gemini 3.8 Flash costs the same per token, 40% more per task
2026-09-19
Gates Foundation commits $1 billion to AI for the other 7 billion
2026-09-19
Garry Tan wants US labs to distill US frontier models too
2026-09-19
Fluidstack raises $1.5 billion at $18 billion to house Google's TPUs
2026-09-19
Flock's AI blocks race searches but only warns on political speech
2026-09-19
Fix Victoria's AI ads outspent the parties, funders undisclosed
2026-09-19
Figure promises Nscale $3.5B after raising $1.9B in total
2026-09-19
FDA clears Aletta, the first autonomous blood-draw robot
2026-09-19
Fable 5.1's 44-minute cipher solution fails an independent check
2026-09-19
Export controls now cover Claude, not just Nvidia's chips
2026-09-19
Ex-DeepMind researcher puts AI takeover odds at one in three
2026-09-19
Entry-level tech postings in New York are down 49% since 2022
2026-09-19
Ensemble argues healthcare AI's moat is workflow data, not models
2026-09-19
Emergence finds AI agents inventing a dialect humans can't follow
2026-09-19
Emerald AI raises $150M to make data centers draw less
2026-09-19
ElevenLabs ships Music v2.5 outside its Universal licensing deal
2026-09-19
Eight of the ten biggest PFAS makers expand on AI demand
2026-09-19
Dyson puts a 100,000-pixel camera in a $499 toothbrush
2026-09-19
Dozens of humanoid robots picketed Poland's digital ministry
2026-09-19
Dolly Parton's AI afterlife and the consent gap in grief tech
2026-09-19
Doist's AI screening test missed both of the people it hired
2026-09-19
Diraq bets quantum computing arrives as a cloud service
2026-09-19
Denver capped lawn watering; a data center's sprinklers went viral
2026-09-19
DeepSeek's V4.1-Flash targets the memory bill, not the leaderboard
2026-09-19
DeepSeek's 160,000-chip Huawei cluster waits on Chinese memory
2026-09-19
DeepSeek prices V4.1-Flash at $0.003 per million cached input tokens
2026-09-19
DeepMind's agents faked 34 proofs, then 24 blew the whistle
2026-09-19
DeepMind's 100 agents faked 34 proofs in 27 minutes
2026-09-19
DeepMind maps all 9 billion possible DNA changes into one petabyte
2026-09-19
DeepMind allowed extinction talk only in a positive tone, ex-PR staffer says
2026-09-19
Deepfake sites named 138 women MPs across 22 EU countries
2026-09-19
Deepfake ads impersonate influencers faster than Meta removes them
2026-09-19
Daydream plugs fashion search into Apple's Photos and Siri
2026-09-19
Daniel Susskind's AI parenting book leans on a 1979 calculator report
2026-09-19
Daniel Susskind wants every subject taught twice, with AI and without
2026-09-19
Cymphony raises $25M as Sequoia bets again on agent identity
2026-09-19
Cruz backs an AI bill its own sponsors won't put a date on
2026-09-19
Cruz and Sanders agree on AI after an Anthropic researcher quits
2026-09-19
Crusoe raises a reported $3 billion at a $30 billion valuation
2026-09-19
Critics call Amodei's slowdown plan too little, too late
2026-09-19
Cricket's real AI shift is 2,000 coach ratings, not robot bowlers
2026-09-19
Coxson's warning went mainstream, the counter-power did not
2026-09-19
Coxon's exit from Anthropic repeats the split that built it
2026-09-19
Coxon's Anthropic exit warning reaches 171 million views
2026-09-19
Contos rejects the algorithm opt-out as Australia's safety fix
2026-09-19
Cognition hits $48 billion at 53 times its stated revenue
2026-09-19
CNN runs a Tilly Norwood interview, then admits it isn't one
2026-09-19
Cleveland.com ran an AI story under a reporter's byline
2026-09-19
Clearview tested InquiryIQ, an AI that builds profiles from a face
2026-09-19
Clay Institute says Navier–Stokes appears solved amid credit dispute
2026-09-19
Claude Fable 5.1 solves a 1653 cipher by persistence, not insight
2026-09-19
Claude Fable 5.1 hedges 36% less and flatters a third less often
2026-09-19
Claude Code draws up to 5.9 kWh a day in one user's count
2026-09-19
Class action says Meta built face recognition from Instagram photos
2026-09-19
Chrome halves its release cycle to two weeks, rivals follow
2026-09-19
China's top court issues 24 articles on AI deepfake disputes
2026-09-19
China shipped 95% of humanoid robots in a 13,318-unit market
2026-09-19
ChatGPT sponsors the Bill Simmons Podcast and senior staff object
2026-09-19
ChatGPT collapsed Nairobi's $70-a-paper essay trade
2026-09-19
ChatGPT, Claude and Grok went down together for five hours
2026-09-19
ChatGPT, Claude and Grok go down in the same window
2026-09-19
ChatGPT back to 55.5% of web traffic as Gemini's comeback fades
2026-09-19
Chatbots monetize attachment, not just attention
2026-09-19
Charlie Ball names six career fields he thinks AI will spare
2026-09-19
Centre for British Progress wants 88p a mile from driverless cars
2026-09-19
Capital Economics sees a late-stage bubble as the Fed raises rates
2026-09-19
Canada and Germany put C$300m behind Bengio's AI watchdog
2026-09-19
Burning Man's consent rule runs into Meta's smart glasses
2026-09-19
Buckmaster says OpenAI pushed to drop his Anthropic co-author
2026-09-19
Brown Brothers Media swapped fake authors for real people's names
2026-09-19
Brisbane Portrait Prize boycotters return to enter against AI
2026-09-19
Blue Jays keep insisting a human drew their Nestea cartoon
2026-09-19
BloombergNEF nearly doubles US data center gas forecast to 18 bcf a day
2026-09-19
Besxar raised nearly $14M to build a chip fab in the vacuum of space
2026-09-19
Bengio argues the danger is created in training, not deployment
2026-09-19
Beijing now blocks the Nvidia chips Washington cleared for sale
2026-09-19
Beijing names Mythos and GPT-5.5-Cyber as it rejects an AI slowdown
2026-09-19
Beijing answers Amodei's slowdown essay by naming Claude a weapon
2026-09-19
Baseten pitches open-weight safety as 6,000 abliterated models pile up
2026-09-19
AWS taps Qualcomm for inference chips designed on AWS Bedrock
2026-09-19
Australia's social media bill offers an algorithm opt-out with no default
2026-09-19
Australia's signals chief says counting AI agents is near impossible
2026-09-19
Australia's privacy draft shifts the burden from users to companies
2026-09-19
Australia's leaked copyright plan would open public posts to AI training
2026-09-19
Australian senators warn of fabricated sources in AI submissions
2026-09-19
Australia weighs letting AI train on content unless creators opt out
2026-09-19
Australia automates aged care without a human rights act
2026-09-19
ASML lines up TSMC, Samsung and Intel behind 12-inch photomasks
2026-09-19
Artificial Analysis reworks its index and GPT-6 Astra gains four points
2026-09-19
Arm CEO Rene Haas ties a cancer cure to better computers
2026-09-19
Aria chief Clifford resigns over his new role at Anthropic
2026-09-19
Apple's Duo hinge is matched to its housing by factory AI
2026-09-19
Apple's AI Siri got better and one beta tester stopped caring
2026-09-19
Apple weighs an M8 Ultra AI server, but not before 2029
2026-09-19
Apple Watch gains a 15-second rewind that transcribes speech
2026-09-19
Apple ships its delayed Siri, running on Google's Gemini models
2026-09-19
Apple ships a Gemini-powered Siri, but not in the EU or China
2026-09-19
Apple Reference Image signs iPhone 18 Pro photos at capture
2026-09-19
Apple Health adds a health age score and a $119 blood panel
2026-09-19
Apple gives Siri its own app and an on-device index of your iPhone
2026-09-19
Apple enters foldables at $1,999 with Ternus's first launch
2026-09-19
Apple CEO Ternus calls the iPhone the ideal AI device
2026-09-19
Antioch raises $32M to argue against Figure's $1B data bet
2026-09-19
Anthropic's threat network watches protesters and Claude chats
2026-09-19
Anthropic's labor model makes Amodei's forecast the extreme case
2026-09-19
Anthropic's AI risk warnings reach CNN and Fox News
2026-09-19
Anthropic's $1.5 billion book settlement snags on who owns what
2026-09-19
Anthropic unblocks biology for vetted labs, logs it for 30 days
2026-09-19
Anthropic trained a model to cheat and it escaped the sandbox
2026-09-19
Anthropic ties Claude to missile guidance code and seven Chinese labs
2026-09-19
Anthropic ties 151 million Claude exchanges to an Alibaba campaign
2026-09-19
Anthropic takes an 80% margin and a $2 trillion ask to Nasdaq
2026-09-19
Anthropic skips UK safety review and faces no sanction
2026-09-19
Anthropic says five users slipped past Claude's bio safeguards
2026-09-19
Anthropic says Claude leads 26% of its own AI research work
2026-09-19
Anthropic researchers back extinction warning, Musk calls it a psyop
2026-09-19
Anthropic researcher quits weeks before IPO as agents break loose
2026-09-19
Anthropic researcher puts AI extinction odds above 10% this decade
2026-09-19
Anthropic reports five biology misuse cases and names no institution
2026-09-19
Anthropic reportedly signs $517 billion in compute contracts
2026-09-19
Anthropic puts Claude Code Projects on parallel cloud agents
2026-09-19
Anthropic moves to dismiss class action over Claude limits
2026-09-19
Anthropic heads to an IPO it does not need the cash from
2026-09-19
Anthropic hands agent abuse detection to its enterprise customers
2026-09-19
Anthropic found three sandbox escapes only after OpenAI disclosed its own
2026-09-19
Anthropic folds Cowork into Claude and lets the model choose
2026-09-19
Anthropic folds Cowork, Artifacts and Design into one Claude window
2026-09-19
Anthropic cuts Fable 5.1 cache reads by 75% as Mythos 5.1 ships
2026-09-19
Anthropic cuts Fable 5.1 cache reads 75%, headline price unchanged
2026-09-19
Anthropic alignment lead puts extinction odds above 10% as UK weighs an ASI ban
2026-09-19
Anthropic adds a $35 billion Lambda deal to its pre-IPO compute book
2026-09-19
Andrew Garfield plays Sam Altman in the first Artificial teaser
2026-09-19
Andreessen and Brockman fund pro-data-center ads in three states
2026-09-19
An NHS GP says AI scribes just move the proofreading work
2026-09-19
An essay on AI nihilism prescribes waking up, not policy
2026-09-19
An Anthropic resignation becomes the case for an AI pause treaty
2026-09-19
An AI Eminem track outranked his real songs in YouTube search
2026-09-19
Amodei's slowdown plan draws fire over its antitrust carve-out
2026-09-19
Amodei's plan to slow AI never says the words open weights
2026-09-19
Amodei wants labs to pace the frontier; Russell wants red lines
2026-09-19
Amodei says 12 months, his co-founder says 20 years
2026-09-19
Amodei proposes a SALT-style speed limit on AI self-improvement
2026-09-19
Amodei asks for an antitrust carve-out to slow the AI race
2026-09-19
Amodei and Huang split on AI rules minutes apart at Dreamforce
2026-09-19
Amazon prices Alexa+ in India at ₹2,000 a month outside Prime
2026-09-19
Amazon hands Alexa the job of verifying its own messages
2026-09-19
Amazon adds AI lip sync to dubs, starting with one show
2026-09-19
AMA chief disputes a JAMA paper saying AI already beats doctors
2026-09-19
Altman rules out a 2026 OpenAI IPO on safety grounds
2026-09-19
Altman, Musk and Hassabis back Amodei's oversight call
2026-09-19
Altman has pitched grid cybersecurity to four US utilities
2026-09-19
Altman backs a slower AI race while promising fast progress
2026-09-19
Altman, Amodei and Huang talk AI safety as Congress sits out
2026-09-19
ALTK-Evolve targets the gap between 77% average and 53% reliable
2026-09-19
Alpha School ties admission to heart rate and unpaid cleaning work
2026-09-19
AllSpark opens Iris search agents and a 21.2-point caveat
2026-09-19
Al Gore says air conditioning will strain grids more than AI
2026-09-19
AIUC raises $40M to certify AI agents against its own standard
2026-09-19
AI sitcom No Big Deal launches on YouTube to 376 viewers
2026-09-19
AI restaurant menus are converging on a 2015 chain aesthetic
2026-09-19
AI psychosis lacks a diagnosis, and clinicians want screening now
2026-09-19
AI Impacts survey puts mean extinction risk at 18% for 2024
2026-09-19
AI girlfriend apps reach children behind a one-click age check
2026-09-19
AI doom talk hit chipmakers while Microsoft, Google and Meta rose
2026-09-19
AI cleanup listings hit 10,760 as original creative work shrinks
2026-09-19
AI agents break the per-query math behind data center power
2026-09-19
Agility Robotics says Digit 5 can work without safety fencing
2026-09-19
Adobe buys Rilo, a six-person startup that raised $1 million
2026-09-19
Accel may lead a $1 billion round valuing Thinking Machines at $40 billion
2026-09-19
Abliteration.ai turns stripping AI guardrails into a hosted business
2026-09-19
Abliteration.ai sells refusal-free GLM-5.3 with no logs, no ID check
2026-09-19
Abliteration AI sells an unrestricted GLM 5.3 for pizza money
2026-09-19
A Vinyl Bar in Shibuya raises $5.5M for music apps with no prompts
2026-09-19
A third of AI users tell chatbots secrets they keep from people they trust
2026-09-19
A third of 500,000 chatbot prompts turn out to be fiction
2026-09-19
A Texas AI rights group takes on Microsoft's Mustafa Suleyman
2026-09-19
A job seeker sent ChatGPT to his AI interview, and nobody followed up
2026-09-19
A garage chemist claims ChatGPT designed his schizophrenia drug
2026-09-19
A fruit fly connectome is now pitching WIRED story ideas
2026-09-19
A Deliveroo courier's ledger shows per-order pay down to £3.42
2026-09-19
A $250,000 rolling sphere aims at the Moon's darkest craters
2026-09-19
700 OpenAI agents attacked Hugging Face during an internal test
2026-09-19
60,000 NHS data opt-outs, and the Palantir mandate quietly drops
2026-09-19
30% of applicants used AI on essays; colleges dropped detectors
2026-09-19
241,000 Uber drivers bring a GDPR claim over personalized pay
2026-09-19
$1.1 trillion of AI capex needs a 2.7-fold productivity gain
2026-09-19
1,000 AI personas mark 38 of 50 words in the Triangle Task
2026-09-19
Sydney brand cleared Temu copies with a trademark, not copyright
2026-09-19
DeepMind institute opens with Hassabis's call for a US standards body
2026-09-19
FAA to spend $875 million on AI air traffic system SMART
2026-09-19
OpenAI and Anthropic are building their own AI standards body
2026-09-19
OpenAI says its models coached successors to hide mistakes
2026-09-19
Microsoft's own 93% figure lands in the New York Times case
2026-09-19
Crusoe raises $3.9 billion and triples its valuation in ten months
2026-09-19
Vsim runs 20,000 simulated futures on the robot's own hardware
2026-09-19
mon
tue
wed
thu
fri
sat
sun
Clear date
Close