AI boom fuels ultra‑fast unicorns, sparks policy limits and new safety reporting tool

The United States lifted a two‑week block on Anthropic’s Mythos 5 model, allowing more than 100 U.S. institutions to access the system, while OpenAI agreed to restrict its GPT‑5.6 lineup to a limited set of government‑approved partners. These moves come as AI‑driven funding accelerates company growth: Anthropic raised $65 billion at a $965 billion valuation, becoming the most valuable private firm, and Europe has seen the number of ultra‑fast unicorns quadruple since 2023, with 20 % reaching $1 billion in valuation within two years.

Separately, researchers have launched FLARE‑AI, a crowdsourced platform for reporting harmful AI behavior—from malware generation to disinformation—aimed at creating a centralized alarm system. The initiative, built with contributions from 49 experts across 32 organizations, is cited as a potential tool for a congressional bill that would give the U.S. government a coordinating role in AI flaw disclosure.

Commentary on AI’s broader impact highlights concerns about a growing “cognitive divide,” where advanced users amplify their capabilities while others fall behind, and warns of diffusion of responsibility as organizations rely on AI without adequate human oversight. Experts also stress that successful AI integration in software development requires structured codebases, citing a rubric of best practices to avoid degrading code quality.