Descriptions:
Episode 271 of Peter Diamandis’s Moonshots podcast covers three major AI developments with a panel that includes Ramin Hasani, co-founder and CEO of Liquid AI. The first story is the launch of Inkling, former OpenAI CTO Mira Murati’s debut model from her new startup. The panel characterizes Inkling as targeting customization over benchmark dominance — a deliberate positioning away from leaderboard competition toward enterprise fit.
Hasani then walks through Liquid AI’s post-transformer architectural philosophy, describing the company’s mission as building efficient general-purpose AI at every scale — specifically models capable of running on CPUs rather than requiring GPU clusters — by exploring “computational graphs of intelligence” beyond the standard transformer paradigm. The discussion covers what architectural diversity might mean for AI deployment at the edge and in resource-constrained environments.
The third major topic is Demis Hassabis’s call for a US-led frontier AI standards body modeled on FINRA, the financial industry self-regulatory organization. Panelists debate whether incumbent labs setting their own standards creates anticompetitive moats for new entrants. The episode also covers a recursive self-improvement scale (levels 0–3) published by a startup called Wo, with the startup claiming it is approaching level one “ignition” — where AI R&D output begins to exceed human pace at equivalent cost.
📺 Source: Peter H. Diamandis · Published July 17, 2026
🏷️ Format: Podcast
