Back to Peter H. Diamandis

Mira Murati's 975B Open Model, Ramin Hasani on Post-Transformer AI, and Demis' AI FINRA | EP #271

Peter H. DiamandisJuly 17, 20261h 56m
Topics71
Mira Murati's Inkling Model0:00Liquid AI and Post-Transformer Architectures0:31Demis Hassabis Calls for FINRA-Style AI Regulation1:01Regulatory Capture Concerns4:31Game Theory Approach to AI Policy9:02Static Law vs. Adaptive Regulation10:31AI Self-Regulation Advantage12:00Regulatory Capture and Open Models13:00Liability Frameworks14:00Western Legal Tradition16:00Regulatory Capture vs. Safety Motivations16:30Non-State Actors Problem17:01Supply Chain Chokepoint Mechanisms17:30Timeline Predictions19:01Current Regulatory Landscape19:31US-China Open Model Capability Framework20:00Beijing Control Risk21:00Perverse Incentives21:30Prevention vs. Adaptation22:00Researcher Migration Risk23:31Quantization Research Leadership24:01Open vs Closed AI Ecosystems24:20Mira Murati's Thinking Machine Labs Inkling Model26:00Customization as the Winning Strategy27:30Benchmark Comparisons and Competitive Positioning28:00Why China Leads in Open-Weight Models29:00Fine-Tuning as a Service Business Model30:45Reinforcement Fine-Tuning as a Paradigm Bet33:00Alex Karp's Sovereignty Argument34:30Fine-Tuning Definition and Evolution35:45Women in AI Leadership41:00Recursive Self-Improvement Research43:00Defensive Co-Scaling Framework44:45Recursive Self-Improvement Scale47:45Foundation Model Lab Operations49:00Defining Recursive Self-Improvement49:15Automated Foundation Model Training51:31Timeline and Scale of Recursive Self-Improvement53:31Security Concerns and Open Source Philosophy56:01Timelines and Depths of Customization58:00Organizational Singularity and Meta-Improvement1:01:00Malaysia's AI Digital Double Initiative1:03:02Risks and Benefits of Government AI Avatars1:04:00Corporate and Institutional Digital Twins1:05:30Organizations Uploading to Cyberspace1:07:01Peter Diamandis AI Avatar Implementation1:08:30The New Medium Beyond Replication1:10:31Preserving Family Legacy Through AI1:11:30Liquid AI and Small Language Models Transition1:12:30Daniela Rus and the MIT Mega Lab1:12:53Liquid Neural Networks: Rethinking the Transformer1:13:31Ramin Hasani's Background and Journey1:14:31Adding Complexity to Neural Units1:16:30From Liquid Neural Networks to Foundation Models1:17:30Building a Foundation Model Lab1:19:01Scaling Laws and Small Language Models1:20:30Defining "Small" and On-Device AI1:22:30Mercedes Partnership and Automotive Deployment1:23:02Enterprise Deployment and Fine-Tuning1:25:00Capabilities Inside the Vehicle1:27:31Advantages of SLMs and Avoiding Overgeneralization1:28:32Model+ Platform and Cross-Domain Applications1:29:31Neuromorphic Origins and Architecture Evolution1:32:30Gating Mechanisms and Meta-System Design1:34:32Automated Foundation Model Design1:36:09Fountain Life Health Segment1:37:30Palmer Luckey on Patent System National Security Concerns1:39:30Debate on Invention Secrecy Act Expansion1:43:00AI Healthcare Abundance and Medical Diagnostics1:48:00Revel Pharmaceuticals Glycation Reversal Breakthrough1:52:00Closing Remarks1:56:00
In a Nutshell

Mira Murati's Inkling (975B MoE, 41B active) is positioned as a customizable open-weight western alternative to Chinese open models, betting that enterprise fine-tuning and on-prem control will beat raw leaderboard performance. Demis Hassabis's push for a FINRA-style frontier AI regulator before year-end raises alarms of regulatory capture by incumbents, perverse US-China open-model ceilings, and the risk that static rules will throttle Western progress while non-state actors ignore them. Meanwhile, Liquid AI's post-transformer liquid neural networks are already delivering sub-1GB multimodal SLMs for Mercedes vehicles, proving that architectures derived from C. elegans can achieve frontier-level intelligence on CPUs and NPUs, while recursive self-improvement loops and glycation-reversing enzymes signal the first concrete steps toward self-accelerating systems and longevity escape velocity.

AI-Generated Notes

These notes were generated by AI and may contain inaccuracies.

Mira Murati, former OpenAI CTO, shipped her first model called Inkling. Customization over leaderboard dominance is what will win in the market. She has built exactly what the market needs right now.

Liquid AI focuses on small language models. Their mission is building efficient general-purpose AI at every scale that explores computational graphs of intelligence beyond transformers. The goal is determining the architectural design that brings frontier model intelligence levels to a CPU.

Demis Hassabis, CEO of DeepMind, called for a US-led frontier AI standards body modeled on FINRA (the industry-funded watchdog that polices Wall Street under SEC oversight). The body would test frontier models before release and should be operational before the end of the year. Elon Musk expects a standalone AI regulator similar to the FAA or FCC to emerge because the consequences of AI going wrong are severe.

When incumbents ask for rules and set standards, they create barriers for entry-level labs. Ramin Hasani noted that Liquid AI has had conversations with the DoD and made a joint submission with AMD regarding regulatory design. Different verticals (automotive, semiconductors, laptops, financial services, e-commerce, biotech) have different enterprise criteria for limits, regulation, and governance structures.

Sign in to read the full notes

Get access to AI-generated notes, topic timestamps, and more.