Investors should prioritize Meta (META) as it disrupts the healthcare sector with its Muse Spark 1.1 model, which outperforms competitors on medical benchmarks at a 7x lower operating cost. Look for entry points into the "Edge AI" theme through private leaders like Liquid AI, which is securing major automotive partnerships by moving AI processing from the cloud to on-device hardware. The "Sovereign AI" trend is accelerating, making companies that offer on-premise model customization, such as Mira Murati’s new startup Inkling, high-conviction targets for enterprise adoption. In the biotech space, monitor the "Damage Repair" sector following Revel Pharmaceuticals' breakthrough in reversing protein aging, a key milestone for the longevity investment thesis. Shift focus away from massive, general-purpose models toward Small Language Models (SLMs) that offer specialized, offline functionality for industries like finance and robotics.

Based on the transcript from Peter Diamandis’ Moonshots podcast featuring Ramin Hasani (CEO of Liquid AI), here are the investment insights and asset mentions extracted for a general audience.

Liquid AI (Private)

• Liquid AI is a foundation model lab originating from MIT’s CSAIL, focusing on "Liquid Neural Networks."
• Technology: Unlike traditional "Transformer" models (like GPT), Liquid AI uses architectures inspired by the nervous system of the C. elegans worm.
• Small Language Models (SLMs): They specialize in efficient, general-purpose AI that can run "on-device" (CPUs, NPUs) rather than requiring massive cloud data centers.
• Key Partnership: Announced a major deal with Mercedes-Benz to power in-car AI that functions offline, ensuring privacy and safety-critical reliability.

Takeaways

• Edge AI Opportunity: Liquid AI represents a shift toward "Edge AI"—bringing intelligence to physical devices (cars, robots, laptops) where connectivity is unreliable and privacy is paramount.
• Efficiency over Scale: While giants like OpenAI focus on massive models, Liquid AI is betting on "computational efficiency," making them a key player for enterprise and hardware integrations.
• B2B Strategy: Their business model focuses on "Model Plus Platform," allowing enterprises to customize and fine-tune small models for specific vertical industries (Biotech, Automotive, Finance).

Thinking Machine Labs / Inkling (Private)

• Founded by Mira Murati (former CTO of OpenAI), the startup recently released its first model, Inkling.
• Specifications: An open-weight model with 975 billion parameters (using a Mixture of Experts architecture where only 41B fire at once).
• Strategic Pivot: Murati is betting on customization over "leaderboard dominance." The model is designed to be downloaded and run on-premise by companies.

Takeaways

• Sovereign AI Trend: There is a growing investment theme around "Sovereign AI"—companies wanting to own their own model weights rather than sending proprietary data to a closed API (like OpenAI or Anthropic).
• Fine-Tuning as a Service: The primary investment insight here is the move toward "Fine-Tuning." Investors should look for companies that enable enterprises to adapt foundation models to their own private data.

Meta (META)

• The podcast highlighted Meta’s release of Muse Spark 1.1, a medical AI benchmark.
• Performance: It reportedly beat GPT-5.6 (OpenAI) on medical benchmarks while being 7x cheaper to operate.
• Distribution: Meta is integrating these high-level medical diagnostics for free across WhatsApp and Facebook, reaching 3.5 billion users.

Takeaways

• Healthcare Democratization: Meta is positioned as a leader in "Healthcare Abundance," demonetizing medical diagnostics. This adds massive value to their ecosystem and user retention.
• Open Source Advantage: By releasing powerful open-weight models (Llama, Muse), Meta is effectively setting the "ceiling" for the industry and challenging the paid subscription models of competitors.

Weco AI (Private)

• A London-based startup that recently published evidence of Recursive Self-Improvement (RSI).
• The Breakthrough: They claim their AI agent (AID²) can rewrite its own code and research strategies. In tests, 8 days of machine self-improvement outperformed 2 years of human effort.

Takeaways

• The "Hard Takeoff" Scenario: RSI is the "Holy Grail" of AI. If AI can effectively design the next generation of AI, the pace of innovation moves from linear to exponential.
• Investment Theme: Look for "Meta-Learning" startups—companies building AI that builds better AI. This is the engine behind the theoretical "Singularity."

Revel Pharmaceuticals (Private)

• A longevity startup (partnered with Alphabet’s Calico) that published a breakthrough in Nature Communications.
• The Discovery: They developed an enzyme (CML-ACE) that acts as a "molecular lawnmower" to remove Advanced Glycation End-products (AGEs)—the "scars" on proteins that cause aging, stiff arteries, and wrinkles.

Takeaways

• Longevity Escape Velocity: This is a significant milestone in the "Longevity" sector. It suggests that damage previously thought to be permanent (chemical cross-linking in the body) is now reversible.
• Sector Growth: Investors should watch the "Damage Repair" (Senolytic/Glycation) space within Biotech, as these technologies move from theoretical to human-tissue validated.

Investment Themes & Risks

1. Regulatory Capture

• Context: CEOs from OpenAI, DeepMind, and Tesla are calling for a "FINRA-like" regulatory body for AI.
• Insight: Analysts in the podcast warn this may be "Regulatory Capture"—incumbents trying to build a "moat" or "cartel" to prevent smaller startups from competing by setting impossibly high safety standards.

2. The "Small Model" Revolution

• Context: The discussion emphasized that the "Scaling Laws" (just making models bigger) are hitting economic and physical limits.
• Insight: The next wave of investment value is in SLMs (Small Language Models) that are specialized, cheaper, and can run without an internet connection.

3. Patent System Risks

• Context: Palmer Luckey (Anduril) argued that the current patent system is a national security risk because it forces inventors to publish "instruction manuals" that adversaries (like China) can rip off.
• Insight: There may be a shift toward "Trade Secrets" and "National Security Patents" (Invention Secrecy Act), which could change how tech companies protect their intellectual property.