Meta tells staff token counts and AI dashboards no longer determine impact scores, even as internal testing of its Hatch agent expands.
In Brief
DOJ backs OpenAI in New York Times copyright case, calls AI training fair use
Pangram CEO warns internet is 'dangerously close' to dead internet theory
Pangram's Max Spero on why AI detection is harder than 'real or fake'
HiddenLayer raises $100M as AI security spending races toward $2.83B
Reliance Jio opens JioPC cloud service to turn 8-year-old PCs into AI machines
Adobe acquires Indian marketing AI startup Rilo in team-and-tech deal
Top stories
The Russian startup fused a 753B-parameter GLM-5.2 with a 4B Qwen-3.5 at one-twentieth the cost of the larger model.
Protect Democracy sues four federal agencies for the unclassified framework governing which frontier AI models get released.
Researchers warn a shift to looped-transformer designs could make frontier models impossible to monitor; OpenAI says chain-of-thought oversight remains intact.
The AI assistant checks messages against Amazon's own send records to flag impersonation scams before customers act on them.
A 20-page DOJ filing argues fair use covers LLM training, siding with OpenAI against The New York Times in the Southern District of New York.
Some 360,000 customers a year ask Amazon whether a message is real; the AI checks it against every message Amazon has ever sent.
Featured
PangramThe 24-person Brooklyn startup has torpedoed book deals and prize wins with percentage scores — and now Substack scans every post.
Jaeden Schafer5 min read
Yotta IPOThe Hiranandani-backed operator is preparing a January–March 2027 listing to fund capacity expansion for AI workloads.
Jaeden Schafer4 min read
Deep Dive
A podcaster's retelling of the July incident triggered a public dispute over anthropomorphism — and who bears responsibility for OpenAI's containment failure.
Frontier Labs
Foundation models, benchmarks, and the research race at OpenAI, Anthropic, Google, Meta, and beyond.
The agency is piloting remote sensing and machine learning to sharpen yield forecasts that farmers say have moved markets against them.
The twinned 5.1 release cuts token costs, reduces false-positive refusals, and finally brings Zero Data Retention to Fable.
The AI upscaler carries a 50–60% performance hit and needs 6x frame generation to hit 1080p on midrange cards.
The AI economy
Funding rounds, earnings, strategy, and the competitive moves defining the AI market.
Google is offering Disney, Warner Bros. Discovery, and Universal hundreds of millions to train Gemini on their libraries — with a per-character price tag.
Revenue jumped 34% to $3.41B as AI-driven attacks fuel demand; Console is the third acquisition in Nikesh Arora's dealmaking spree.
Two former Sequoia IT leaders spun out an autonomous observability tool that flags risky system changes before they cascade into downtime.
This week
Models
Jaeden Schafer5 min read
Security
Jaeden Schafer5 min read
Security
Jaeden Schafer5 min read
Security
Jaeden Schafer4 min read
Security
Jaeden Schafer4 min read
Opinion
AstraThe unreleased frontier model is the first to cross OpenAI's 'critical cybersecurity threshold,' with restricted access planned at launch.
Latest
All news →Tools
Jaeden SchaferSep 1
Tools
Jaeden SchaferSep 1
Business
Jaeden SchaferSep 1
Tools
Jaeden SchaferSep 1
Security
Jaeden SchaferSep 1
The beats
Nine beats covering every corner of AI — from frontier models to the workflows people actually use. Pick one to jump into the latest reporting.
Daily coverage of what just happened in AI.
Latest in News
USDA to test satellites and AI for crop estimates after farmer backlash
Funding, revenue, strategy, and the AI economy.
Latest in Business
Meta drops AI-usage metrics from performance reviews as Hatch agent rolls out
Reviews, tutorials, and picks for the AI tool stack.
Latest in Tools
Google launches Pics, a Workspace AI image editor aimed at Canva
Roles, salaries, and how AI is reshaping work.
Latest in Careers
Insurance claims adjusters are the workforce's biggest AI haters, Glassdoor finds
Threats, breaches, and the security side of AI.
Latest in Security
Lawsuit seeks to force Trump administration to reveal AI safety review rules
AI, religion, and the spiritual questions AI raises.
Latest in Faith
Pope Leo XIV's AI encyclical flagged as 46% AI-written by Pangram detector
Foundation models, benchmarks, and the research race.
Latest in Models
Mostik builds a mathematical bridge letting AI models share weights directly
Commentary and long-form takes on where AI is headed.
Latest in Analysis
Pangram CEO warns internet is 'dangerously close' to dead internet theory
The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.
The briefing read inside teams at
