LatestAI
Cats in AI Training Data: From Shelter Photos to Neural Networks
How cats became CAPTCHA subjects, benchmark categories and learned visual features—and what those examples do not prove about image generators.
Read the story →Artificial Intelligence — Featured
All posts →LatestArtificial Intelligence
Showing 9 of 59 posts
AIAI Found Hundreds of Security Bugs. Here Is What the Research Actually Shows.
Anthropic’s 500-bug report, Mozilla’s fixes and Calif’s MAD Bugs are distinct efforts. What their AI-assisted security research demonstrates.
AIMercor Breach: Hackers Claim an AI Data Haul, but the Scope Remains Unclear
Mercor confirms a LiteLLM-linked breach. What is known about the alleged data sale, Meta’s reported pause and the…
AIAnthropic, the Pentagon and the Court Fight Over AI Guardrails
Why Anthropic challenged the Pentagon, what the March injunction actually does, and how the government’s appeal differs from…
AIGoogle Gemma 4 Is Out Today and the Numbers Are Hard to Ignore
Google launches four Gemma 4 models under Apache 2.0. What the open-model rankings mean, how the sizes differ,…
AIOpenAI Just Raised $122 Billion. Yes, Billion. With a B.
OpenAI announces $122 billion in committed capital at an $852 billion valuation. How funding, revenue, investor access and…
AIClaude Cowork and Copilot Cowork: Related Technology, Different Products
Claude Cowork began in January; Microsoft’s Copilot Cowork reached Frontier in March. What each does, and how Researcher’s…
AISora Is Closing: What the Announcement Means for Creators
OpenAI has announced the end of the Sora app. A March 2026 look at what creators can prepare,…
AIWhen AI Agrees Too Much: What the Stanford Sycophancy Study Found
A Stanford study found that overly agreeable AI can affect judgments and willingness to repair conflicts. What the…
Showing 9 of 59 posts