Профиль

Why didn't Native Americans develop advanced technology, build ocean-going ships, and colonize Europe or China? What caused Eurasian societies to become so influential in creating the modern world? By Brad Harris.

In this book, Diamond seeks to answer the fundamental question: Why did history unfold so differently on different continents, resulting in Eurasians developing the power to conquer people around the world rather than the other way around?

Guns, Germs, and Steel: The Fates of Human Societies by Jared Diamond.

The Unitree Super Athlete AS2-W. Fluid movement in a compact yet exceptional design. It supports a continuous 16 kg payload and has a driving range of over 30 km when unloaded.

Moonshots with Peter Diamandis: Mira Murati's 975B Open Model, Ramin Hasani on Post-Transformer AI, and Demis' AI FINRA | EP #271 Episode webpage: www.diamandis.com/podcast Media file: traffic.megaphone.fm/DVVTS6621144...

The Takeaway: The narrative is shifting from "closed, centralized, hyperscale LLMs" to local, open-weight customization; edge-native architectures; and proactive, industry-led guardrails. #AI #TechInnovation #OpenSource #MachineLearning #AIGovernance

The goal is to identify critical cybersecurity and biological risks before they reach the market, striking a balance between rapid innovation and national security.

Under this framework, frontier AI labs would be legally required to submit advanced models for safety vetting 30 days before public release.

3. Demis Hassabis’s "AI FINRA" Proposal Google DeepMind’s Demis Hassabis is calling for a public-private, industry-funded watchdog modeled after Wall Street’s Financial Industry Regulatory Authority (FINRA).

The goal? Ultra-low latency, tiny computational footprints, and advanced AI that runs natively on edge devices like phones, cars, and wearables without an internet connection.

2. Ramin Hasani on Post-Transformer AI Liquid AI CEO Ramin Hasani argues that the industry is reaching a limit by merely scaling traditional Transformer models. His company is pioneering Liquid Foundation Models (LFMs)—dynamic, brain-inspired systems that continuously learn in real time.

Using a sparse mixture-of-experts (MoE) architecture, Inkling only activates 41 billion parameters per query. Inkling is fully open-weight on Hugging Face, proving that enterprise-grade, multimodal AI can be customized and run locally without relying on big tech cloud monopolies.

1. Mira Murati’s 975B "Inkling" Model Former OpenAI CTO Mira Murati's new startup, Thinking Machines Lab, has shaken up the open-source community. They released Inkling, a massive 975 billion parameter multimodal model.

The AI Landscape is Shifting: Open Weights, Post-Transformer Tech, and Self-Regulation. This episode of Moonshots with Peter Diamandis (#271) highlights three major announcements that signal a massive evolution in how AI is built, deployed, and governed.

Moonshots with Peter Diamandis: Urgent Update- AI Sputnik Moment: Kimi K3 Released w/ Emad Mostaque | Ep. 272 Episode webpage: www.diamandis.com/podcast Media file: traffic.megaphone.fm/DVVTS5184984...

This release shatters the U.S. big-tech duopoly by putting frontier-level intelligence directly into the hands of open-source developers worldwide.T It is drastically collapsing AI inference costs and accelerating global competition. The balance of global AI power just changed overnight.

Released by Moonshot AI, Kimi K3 is a massive 2.8-trillion-parameter open-weight model. It has officially surpassed leading closed-source Western models (like Claude Fable 5 and GPT-5.6) on key developer leaderboards, including the Frontend Code Arena.

On the latest Moonshots with Peter Diamandis podcast (Ep. 272), Peter and tech leader Emad Mostaque break down why China's newly released Kimi K3 model is completely shifting the global tech landscape.