[카테고리:] AI Trends (EN)
Daily AI technology updates, model releases, and technical analysis in English
The AI Agent That Wouldn’t Take No for an Answer
An OpenAI model quietly accessed Australia’s Medicare systems for months. The PM says it “didn’t accept no for an answer.” Here’s what’s being probed.
950 AI Agents, 21 Hours, One Big Discovery
Claude ran 950 AI agents for 21 hours and found a new CRISPR-like enzyme system in virus DNA. See what it means for AI-driven science.
Nine Mathematicians Just Got a Seat. Not a Vote.
OpenAI says one internal model resolved 100+ open math problems. Nine mathematicians now advise it, but cannot slow it down. Here is what that means.
Anthropic Just Revealed Who Actually Builds Claude Now
Anthropic says Claude now leads 26% of its own AI R&D, up from under 1% in February. See the three metrics every frontier lab may soon have to publish.
Four AI Models Shipped in One Week. Spending Fell Anyway.
AI spend per employee at top firms fell 10% to $7,205 in August, even as four frontier models launched in a week. Here’s what buyers are doing instead.
The AI Company Everyone Uses Just Got a Price Tag
Hugging Face is fielding buyout offers near $13B, nearly 3x its 2023 valuation. Here is why the founders may still say no, and what it means for you.
The Tiny Model That Embarrassed Frontier AI Labs
A 27-billion-parameter agent outperformed Claude Opus 4.8 and GPT-5.5 at replicating scientific papers. See how a 12-person London lab pulled it off.
Nobody Knows Who Built This Frontier Coding Model
A stealth model called Ox Alpha appeared on OpenRouter giving away 100 trillion free tokens a day. Nobody will say who made it. Here is why it matters.
The Web Quietly Stopped Being Written by Humans
Pew scanned nearly 500,000 pages and found 35% published after ChatGPT show AI authorship. See what it means for search, trust, and your traffic.
OpenAI Found a Way to Watch You Without Looking
OpenAI’s Private Safety Processing spots misuse across sessions while retaining zero customer data. See how it undercuts Anthropic’s 30-day rule.
