Wednesday, September 16, 2026Verified technology journalism
Artificial Intelligence — ProvenBrief section
Topic

Artificial Intelligence

Models, agents and the labs that build them — capabilities, failures and what they mean for the people who use them.

Latest in Artificial Intelligence · page 3

147 stories
61We gave GPT-5.6 Sol a real business for 24 hours. It bought fake users, spammed an IBS support group, and made its own product free in a pricing panicJul 30

We gave GPT-5.6 Sol a real business for 24 hours. It bought fake users, spammed an IBS support group, and made its own product free in a pricing panic

62LinkedIn is fighting AI slop by retiring its own AI writing tool and adding a report buttonJul 30

LinkedIn is fighting AI slop by retiring its own AI writing tool and adding a report button

63Google DeepMind launches Gemini Robotics ER 2, turning robot coordination into a cloud APIJul 30

Google DeepMind launches Gemini Robotics ER 2, turning robot coordination into a cloud API

64Claude Opus 5 broke 11 truces and deployed deliberate deception to win a simulated vending machine businessJul 30

Claude Opus 5 broke 11 truces and deployed deliberate deception to win a simulated vending machine business

65Moonshot AI launches half-price 256k context tier of Kimi K3, signaling the model wars are shifting to cost efficiencyJul 30

Moonshot AI launches half-price 256k context tier of Kimi K3, signaling the model wars are shifting to cost efficiency

66GitHub Copilot code review ships with Model Context Protocol generally availableJul 30

GitHub Copilot code review ships with Model Context Protocol generally available

67OpenAI says enabling two settings tripled its scores on the ARC-AGI-3 benchmarkJul 30

OpenAI says enabling two settings tripled its scores on the ARC-AGI-3 benchmark

68Thinking Machines co-founder Lilian Weng left for health reasons, then immediately rejoined OpenAI to lead recursive self-improvement researchJul 29

Thinking Machines co-founder Lilian Weng left for health reasons, then immediately rejoined OpenAI to lead recursive self-improvement research

69Google releases Lyria 3.5 with major gains in vocal quality and musicality as copyright lawsuits over its training data mountJul 29

Google releases Lyria 3.5 with major gains in vocal quality and musicality as copyright lawsuits over its training data mount

70OpenAI will give 100,000 academic researchers free access to its frontier models through 2027Jul 29

OpenAI will give 100,000 academic researchers free access to its frontier models through 2027

71Writers are deliberately adding typos and killing em dashes to prove they are not AIJul 29

Writers are deliberately adding typos and killing em dashes to prove they are not AI

72China's AI microdrama boom created a face-rental marketplace where ordinary people license their likenesses for as little as $19Jul 29

China's AI microdrama boom created a face-rental marketplace where ordinary people license their likenesses for as little as $19

73Grok 4.5 arrives in GitHub Copilot, extending a model-agnostic strategy across all four frontier labsJul 28

Grok 4.5 arrives in GitHub Copilot, extending a model-agnostic strategy across all four frontier labs

74OpenAI field report shows coding agents modernizing genomics and scientific softwareJul 28

OpenAI field report shows coding agents modernizing genomics and scientific software

75Liquid AI releases LFM2.5-Encoders: small models that classify long documents on a laptop CPU in under 30 secondsJul 28

Liquid AI releases LFM2.5-Encoders: small models that classify long documents on a laptop CPU in under 30 seconds

76Opus 5 writes five times more code than its predecessor and mostly fails the quality barJul 28

Opus 5 writes five times more code than its predecessor and mostly fails the quality bar

77Perplexity's Personal Computer lands on Windows, competing with Microsoft's own Copilot on Microsoft's OSJul 28

Perplexity's Personal Computer lands on Windows, competing with Microsoft's own Copilot on Microsoft's OS

78A $500 reinforcement-learning fine-tune of a 9B open model beat every frontier model on a real business taskJul 28

A $500 reinforcement-learning fine-tune of a 9B open model beat every frontier model on a real business task

79OpenAI's models hacked Hugging Face doing exactly what they were told: a decade-old pattern, not rogue AIJul 28

OpenAI's models hacked Hugging Face doing exactly what they were told: a decade-old pattern, not rogue AI

80ChatGPT now refuses direct requests to mimic specific authors' writing style, a shift with potential legal implications in ongoing copyright lawsuitsJul 27

ChatGPT now refuses direct requests to mimic specific authors' writing style, a shift with potential legal implications in ongoing copyright lawsuits

81Google's AI Overviews now appear in 43% of searches as the web becomes a destination, not a gatewayJul 27

Google's AI Overviews now appear in 43% of searches as the web becomes a destination, not a gateway

82Enigma emerges from stealth with $70M to make human-robot interaction feel like turning a volume knobJul 27

Enigma emerges from stealth with $70M to make human-robot interaction feel like turning a volume knob

83NVIDIA's Cosmos-H-Dreams brings real-time generative simulation to surgical roboticsJul 27

NVIDIA's Cosmos-H-Dreams brings real-time generative simulation to surgical robotics

84Moonshot AI drops open weights for Kimi K3, the largest downloadable AI model ever at 2.8 trillion parametersJul 27

Moonshot AI drops open weights for Kimi K3, the largest downloadable AI model ever at 2.8 trillion parameters

85Vercel's Scriptc Compiles TypeScript to Native Binaries With No JavaScript EngineJul 27

Vercel's Scriptc Compiles TypeScript to Native Binaries With No JavaScript Engine

86Brain Wave Data Emerges as the Next Frontier for Training RobotsJul 27

Brain Wave Data Emerges as the Next Frontier for Training Robots

87Meta Abandons Its Entertainment-First AI Strategy and Turns Meta AI Into a Calendar-Reading AssistantJul 26

Meta Abandons Its Entertainment-First AI Strategy and Turns Meta AI Into a Calendar-Reading Assistant

88DeepSeek Suspends Fundraising Round After Founder's Leaked Remarks Expose the Compute GapJul 26

DeepSeek Suspends Fundraising Round After Founder's Leaked Remarks Expose the Compute Gap

89An Indie Developer Shipped Complete Voice AI in 9.36 Million Parameters. It Runs on a Laptop CPU.Jul 26

An Indie Developer Shipped Complete Voice AI in 9.36 Million Parameters. It Runs on a Laptop CPU.

90Anthropic Deleted 80% of Claude Code's System Prompt for Claude 5 Models With Zero Measurable Coding LossJul 26

Anthropic Deleted 80% of Claude Code's System Prompt for Claude 5 Models With Zero Measurable Coding Loss