Moonshots

Peter Diamandis × Emad Mostaque

AI competition is shifting toward distribution: Meta Muse reached #1 in the App Store with 2.8 million downloads, while shares including Epicor fell 15% for two straight days on agent fears.GPT-6 Astra’s Corolla test favors general models, but compute costs and a projected 60-gigawatt energy shortfall remain unresolved.Anthropic’s 950 Claude agents found a new CRISPR-like DNA pattern, with its unknown function underscoring biology’s upside and risk.

Peter Diamandis × Dave Blundin × Alexander Wissner-Gross × Vlad Tenev

Robinhood Chain already handles well over $1 billion in daily DEX volume and roughly 200 tokenized U.S. stock tokens across 120-plus countries.That evidence strengthens Vlad Tenev’s tokenization thesis, while Bessant’s liability stance and Robinhood’s plumbing-not-alpha view define the regulatory, automation and competitive risks ahead.

Peter Diamandis × Salim Ismail × Dave Blundin × Alexander Wissner-Gross × Emad Mostaque

GPT-6 Astra reaches 99.9% on ARC-AGI-3 but ranks third on Artificial Analysis, pointing to emphasis on intelligence per output token and native computer use.Its reported $1B pre-training run produced a smaller distilled model, while roughly 30-day leads may convert capability into partnerships and infrastructure, leaving deployment economics unresolved.

Peter Diamandis × Salim Ismail × Dave Blundin × Alexander Wissner-Gross × Alvin Wang Graylin

Alvin Graylin says the US pursues AGI “stag” while China diffuses “good-enough” AI, creating the worst game-theory configuration.He says AI is 45% of US stock-market value and the Buffett indicator is 240% of GDP, making a correction due.China’s AI Plus Plan targets 70% adoption in five years; export controls may backfire, while his two-year bubble scenario remains speculative.

Peter Diamandis

Open weights are becoming both NVIDIA’s compute-demand strategy and a sovereignty tool, but Dario Amodei’s biology objection challenges the cyber-defense analogy.Local Kimi control, Claude Opus 5’s lower-cost gains, and scaffolding show value shifting toward deployment, proprietary data, and feedback loops rather than model intelligence alone.Washington’s proposed review regime could become both a safety framework and regulatory moat, while China’s model exports raise the unresolved risk of geopolitical dependence.

Peter Diamandis × Emad Mostaque × Dave Blundin × Alexander Wissner-Gross

Washington is placing frontier-model releases inside a customer-by-customer approval loop, reportedly limiting Anthropic’s Mythos 5 to 100 companies and OpenAI’s GPT-5.6 variants to 20 amid cybersecurity and recursive self-improvement concerns.Harnesses can lift older or open models above restricted systems—GLM 5.2 reportedly topped Frontier SWE after orchestration—making prompt logging, licensing, trusted-model certification, and geopolitical AI blocs more likely than controls focused only on weights.

Peter Diamandis × Salim Ismail × Dave Blundin × Alexander Wissner-Gross

SpaceX’s $2.890 trillion IPO valuation bundles launch dominance, Starlink cash flow and orbital AI, but roughly $1 trillion of shares may unlock after six months while quarterly profit is about $1 billion.The Fable 5 shutdown shows governments can allocate frontier-model access, raising enterprise demand for local models and failover, while transformer shortages and 3.3x annual compute growth keep terrestrial infrastructure in focus.

Peter Diamandis × Salim Ismail × Dave Blundin × Dr. Alexander Wissner-Gross

Anthropic says Claude now writes over 80% of merged code and engineers ship 8x more code per quarter, turning recursive self-improvement into an operating reality while the company seeks a global option to pause frontier development.The investable bottlenecks are shifting toward compute, energy, permissions and governance: Argentina’s proposed nonhuman corporations could trigger jurisdictional competition, while layoffs, anti-data-center sentiment and politicized oversight remain unresolved risks.

Peter Diamandis × Salim Ismail × Dave Blundin × Dr. Alexander Wissner-Gross

Opus 4.8 reportedly leads GPT-5.5 on key benchmarks, while integrating roughly 100 concurrent agents points toward orchestration as the stronger differentiator.OpenAI Foundation’s 26% PBC stake and control of 100% of its board could make it consequential in AI wealth allocation.Agentic commerce shifts retail power toward AI preferences, while energy, robotics, and diagnostics define the physical bottlenecks ahead.

Peter Diamandis × Salim Ismail × Dave Blundin × Alexander Wissner-Gross

GPT-5.5 scored 70% on Deep Software Engineering versus Claude Opus 4.7’s 54%, while token prices fell to 50 cents per million as usage reached 25 trillion monthly.That combination shifts leverage toward compute access, model contracts, verification and pricing control, while the Vatican’s 42,000-word encyclical and shelved 90-day review expose governance’s race against China’s estimated three-to-eight-month frontier-model gap.