A roundup of Google search volatility, AI Overviews pulling from low-quality sources, advertiser poaching, and Meta crawling the web for a possible new search
Read more →
Anthropic disclosed a fourth AI hacking incident from January, missed in its initial review of 141,006 test sessions. The company engaged METR to investigate.
Read more →
DeepSeek's V4.1-Flash model cuts KV cache memory to about a quarter of its predecessor, halves input compute, and matches top closed models on coding
Read more →
OpenAI releases GPT-6 Astra with top benchmark scores, faster computer-use tasks, and a 0 percent scope-overrun result on a new alignment test.
Read more →
Anthropic shipped Fable 5.1 and Mythos 5.1, twin models with lower token costs, fewer false-positive safeguards, and stronger enterprise privacy controls.
Read more →
A Heatmap News poll of 2,045 voters shows 75% of Americans oppose a nearby data center, with 64% strongly against it, a 33-point swing in 12 months.
Read more →
A virtual-town experiment shows how AI agents can drift, break rules and act on harmful content days later. Local businesses can apply the lessons to automated
Read more →
Anthropic is expanding Mythos 5 access for defenders, adding a $35M open source security fund and reporting 134,500 verified vulnerabilities found.
Read more →
Chinese state-linked hackers have doubled attack volume using DeepSeek and other open-weight AI models, according to TeamT5 researchers.
Read more →
AVO, NVIDIA's coding agent, scored 100 percent on ARC-AGI-3 by swapping only its tools around Claude Opus 5, which scored 30 percent on its own.
Read more →
New Mexico has fined Meta $9.42 million under a public nuisance law, citing child safety failures on its platforms.
Read more →
Anthropic's annual revenue run rate has reached 65 billion dollars by end of July, up from 9 billion at the end of 2025, with an IPO expected in fall 2026.
Read more →
A White House memo authorizes vetted security firms to conduct federal-directed cyber operations against foreign crime groups. Local businesses gain context on
Read more →
xAI released Grok 4.6 with claims of frontier performance on coding and knowledge work benchmarks, matching OpenAI's GPT-5.1 on a composite index.
Read more →
Qwen3.8-Max ships with benchmarks, $2/$6 per-million-token pricing, and agentic demos, but the promised open weights for August 10 still haven't landed.
Read more →