A Startup Beats Google, Power Users Break Away, and AI Gets Regulated

A six-person startup just beat Google on one of the hardest reasoning benchmarks — using Google’s own model. And inside companies, the top 5% of AI users are quietly gaining the equivalent of an extra workday every week.

In this episode, Jeff and Annie break down Poetiq’s ARC-AGI-2 win and why meta-systems — critique, refine, verify — may now matter more than picking the “best” model. They unpack OpenAI’s first State of Enterprise AI report, including the widening productivity gap between casual users and power users. Finally, they run through Quick Hits on chips, regulation, XR glasses, factuality benchmarks, the emerging AI licensing economy, and a major shift in how algorithmic bias could be judged.

🔍 In this episode:

• Poetiq’s orchestration layer beats Gemini Deep Think on ARC-AGI-2
• OpenAI data reveals the productivity chasm between median and power users
• RSL 1.0, agent standards, and regulation reshape the emerging “AI internet economy”
  
  

🎧 Watch the full episode of What the AI?! on YouTube, Spotify, or Apple Podcasts → https://www.whattheai.fm

Relevant links:

• Poetiq ARC-AGI-2 benchmark results and verification
• OpenAI State of Enterprise AI report
• Nvidia H200 export approval and U.S. revenue cut reporting
• Trump executive order push on national AI rules
• Google XR glasses pre-announcement
• Google DeepMind FACTS benchmark announcement
• OpenAI hires Slack CEO Denise Dresser as CRO
• OpenAI GPT-5.2 model update
• RSL 1.0 web licensing standard
• Cloudflare support for AI content licensing
• EU antitrust investigation into Google AI Overviews
• Google updates AI Mode with more publisher links
• OpenAI and Disney licensing deal coverage
• DOJ ends enforcement of disparate impact standard