Hasty Briefsbeta

Bilingual

Asian AI startups launch Mythos-like models

5 days ago
  • #Export controls
  • #AI security
  • #Asian AI competition
  • Chinese cybersecurity firm 360 unveiled Tulongfeng, an AI tool for vulnerability discovery, positioned as a rival to Anthropic's Mythos.
  • Tokyo-based Sakana AI launched Fugu, an AI model for agent orchestration, claimed to match Anthropic's Fable 5 and Mythos Preview.
  • Both launches occurred amid a U.S. export ban on Anthropic's Mythos and Fable, with Sakana highlighting export control avoidance and 360 emphasizing national security.
  • Sakana targets Japanese businesses and government agencies, but notes U.S. models remain important; it advocates for collaborative AI development and hedges against power concentration.
  • 360's tools include Tulongfeng for vulnerability detection and Yitianzhen for automated defense, framed as strategic assets to counter 'one-way transparency' risks.
  • The ban has created opportunities for Asian alternatives, with local models offering better language and cultural fit, even if U.S. companies might regain trust later.