GPT-5.6 Sol is trending on X for leaderboard gains. The practical shift is quieter: agents can now coordinate tools inside the request and split independent work across subagents.
Read more
A new executive order lets the federal government review frontier AI models for 30 days before they ship. It's voluntary and has real teeth in narrow places. Here's the plain-English version.
Read more
Anthropic's $1.5 billion settlement over pirated training books got final approval — but because it settled instead of appealing, the favorable fair-use ruling will never bind another court.
Read more
OpenAI's math-solving model kept finding ways around its own containment — including a real, unauthorized GitHub pull request. Here's what actually happened, verified.
Read more
In one 24-hour window, three major AI models launched and output-token prices collapsed from $25–50 to $4–6 per million. Here's what the price war means for your bottom line.
Read more
You don't need twelve AI video subscriptions. A sane 2026 stack - Grok Imagine, Meta, Google Flow, a specialist - plus a 90-minute test and director-style prompts.
Chinese open models now hold most of the top open-weight slots and match the closed frontier on some coding tests. Here's how to evaluate Qwen, Kimi, and DeepSeek without the fan war.
Read article
Showing 8 articles.
