MTS

Khosla Ventures founding partner Samir Kaul joins MTS to discuss why compute is no longer the real bottleneck in AI biology, how the semiconductor supply shortage will break within five years, and why energy constraints will surpass chip limits in the next 12 months. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love. MTS is a live news and interview show cove...

Cyber ethnographer Ruby Justice Thelot explains why social media platforms are aging out young users, making a push into AI vertical dramas and 90-minute videos, and shedding personal communication. He tracks how the decline of traditional home hosting is spurring the rise of commercial offline spaces like run clubs and supper clubs.-- https://x.com/being_on_line --Timestamps:(00:00) Introduction...

Former OpenAI governance and DeepMind safety researcher Richard Ngo analyzes why rapid recursive self-improvement timelines are overhyped and breaks down how frontier labs justify race dynamics. He explains why AI capabilities will remain jagged for years and argues that the alignment field is bottlenecked by conceptual confusion and political groupthink. Check out our sponsor Lovable: https://lov...

Tutor Intelligence co-founder and CEO Josh Gruenstein discusses raising a $75M Series B to build and deploy factory robots across America. He explains why foundation models like GPT Astra unlock real-world physical automation, why 90% of US warehouses still lack a single robot, and how low-cost robotic labor could radically deflate physical goods. Check out our sponsor Lovable: https://lovable.dev...

Luke Alvoeiro from Factory breaks down Droid 1.0, introducing a programmatic tool calling engine that generates multi-step scripts to cut context bloat and boost agent reliability. He reveals how Droid secured an 8% lead over Claude Code on Terminal Bench 4 using identical frontier models, and shares his 5-year outlook for fully autonomous software factories. Check out our sponsor Lovable: https:/...

Larridin co-founder and CEO Russ Fradin breaks down data from their enterprise AI research and Anthropic's S-1 filing, showing how AI spend currently drives massive productivity gains for top engineers while lowering output for bottom performers. He outlines why enterprise agent adoption remains tiny, how companies must audit workflows to justify hundreds of billions in model spend, and when real...

Representative George Whitesides discusses the reality of recursive self-improvement inside frontier labs, his push to monitor automated AI R&D, and his call for a mandatory one-month pause on frontier model development modeled on aerospace flight halts. He also details congressional battles over funding the Center for AI Safety and Innovation, federal oversight of classroom AI use, and the ge...

Foundation for American Innovation research fellow Daniel King breaks down the impending race to industrialize outer space through rapid advancements in robotics and AI. He unpacks the real-world economics of orbital data centers, the strategic risks of chokepointed terrestrial launch infrastructure, and the rising push for commercial space titling over speculative galaxy distribution schemes. Che...

Base director Antonio García Martínez joins MTS to discuss how on-chain infrastructure is unlocking global access to US equities and fueling autonomous AI agent transactions. He explores the rise of tokenized stocks, why agent-to-agent e-commerce will outpace human retail, and how programmable finance is becoming the native monetary layer for AI. Check out our sponsor Lovable: https://lovable.dev....

MNX founders Stephen Grugett and Alex Masmej explain how they built a new financial exchange to trade AI valuations, compute futures, and fast-takeoff risk. They break down why OpenAI and Anthropic pull ahead in short timeline scenarios, the market mechanics behind compute commodities, and how decentralized infrastructure responds to cyber threats. Check out our sponsor Lovable: https://lovable.de...

General Compute co-founder and CTO Jason Goodison breaks down their $100M Series A to build a dedicated inference cloud on Cerebras wafers. He explains why fast inference requires moving away from standard GPU racks, how the industry is shifting from a dial-up era to broadband interactivity, and why open-source models will drive the next generation of real-time AI agents. Check out our sponsor Lov...

Foundry Robotics founder and CEO Adarsh Kulkarni discusses raising a $100M Series A to build software-defined, AI-orchestrated factories from scratch. He explains why retrofitting legacy sites fails, outlines the four-nines reliability barrier in manufacturing, and predicts 1 million robots will deliver 5 billion work hours annually within 36 months. Check out our sponsor Lovable: https://lovable....

Edison Scientific founder and CEO Sam Rodriques breaks down their new partnership with Pfizer to deploy superintelligence across pharma R&D. He details how their AI scientists Robin and Cosmos are already discovering novel disease treatments, and maps out why drug repurposing will deliver rapid medical breakthroughs before entirely new AI-designed compounds arrive. Check out our sponsor Lovabl...

AMD's Ian Weinberg demos the Ryzen AI Halo developer box live in studio with Sophia Dew, running massive AI workloads locally. He details how 192GB of unified memory unlocks local 300-billion-parameter models and multi-agent workflows, and explains how semantic routing diverts 70% of enterprise reasoning tasks away from frontier models to on-prem silicon. Check out our sponsor Lovable: https:/...

Stanford professor and Hoover Institution senior fellow Allison Okamura challenges the current race toward humanoid robots, arguing that single-task and soft vine architectures will transform industries like medicine and eldercare far faster. She warns against applying language-model agentic AI directly to physical robots without safety guarantees, outlining what robotics adoption will actually lo...

Alexander Barry, head of capabilities tracking at Epoch AI, breaks down a new study showing how open-weight AI models can reverse engineer software patches into zero-day browser exploits. He discusses testing GLM 5.3 against Firefox, the shrinking patch gap, and whether upcoming models like Kimi K3.1 Pro threaten frontier cyber security. Check out our sponsor Lovable: https://lovable.dev. Turn ide...

Harmonic co-founder and CEO Tudor Achim discusses expanding their Aristotle formal verification system from math proofs to C code and chip design. He explains why AI-powered cyber attacks make code verification an urgent national security priority for critical infrastructure, and how formal verification eliminates entire classes of exploits. Check out our sponsor Lovable: https://lovable.dev. Turn...

Cohere Chief AI Officer Joelle Pineau discusses the company's massive Series E fundraise, expanding enterprise deployments, and building frontier AI models outside the US and China. She also breaks down Cohere's acquisition of Prague-based cybersecurity harness startup Aisle and the imperative for organizations to use AI for defense. Check out our sponsor Lovable: https://lovable.dev. Turn...

Mistral CRO Marjorie Janiewicz joins MTS to break down the open-weight release of Mistral Large 4. She discusses how the frontier model matches closed-source coding and cyber defense capabilities at a fraction of competitors' compute investment, while laying out why enterprise sovereignty and model customization will render generic chatbots obsolete within a year. Check out our sponsor Lovable...

Sarah Constantin of Principles of Intelligence explores the blind spots in AI agent evaluation and why fast doom scenarios misjudge physical science. She also unpacks why current US tort law fails to hold companies liable for unintentional AI cyberattacks, alongside reflections on rationalist culture. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people love. MTS is...

Databricks VP and Neon co-founder Nikita Shamgunov explains why cloud hyperscalers offer the wrong primitives for AI agents and how Lakebase solves database state at scale. He breaks down separating compute from storage, the shift to AI software factories, and why competitors like Microsoft are copying Neon's branching architecture. -- https://x.com/nikitabase -- MTS is a live news and intervi...

Mistral research scientist Sagar Vaze breaks down the open-weight release of Mistral Large 4, a 1-trillion parameter multimodal model trained on 1E25 FLOPS. He explains how on-policy distillation delivers extreme token efficiency, why enterprise fine-tuning beats massive closed models, and how open weights protect cyber defenders from provider lockouts. Check out our sponsor Lovable: https://lovab...

Secure AI Project co-founder Nick Beckstead breaks down the push for frontier AI safety laws, arguing why government limits on training run compute are necessary to prevent catastrophic loss of control. He evaluates regulatory gaps exposed by the Hugging Face incident, California's executive order on third-party audits, and the coming impact of automated AI research on timeline speeds. Check o...

FCC Senior National Security Counsel Adam Chan explains the commission's crackdown on hardware containing Chinese surveillance chips and its broader bans on foreign drones, routers, and power inverters. He discusses how frontier AI cyber defense models protect critical telecommunications infrastructure against advanced nation-state threats like Salt Typhoon, and why the US must maintain its le...

Formation Bio co-founder and CEO Benjamin Liu breaks down the company's new $600M fund to license stalled drug candidates and run them through clinical trials with AI. He argues that AI drug discovery is creating massive abundance, while clinical development remains the true bottleneck for saving human lives. Check out our sponsor Lovable: https://lovable.dev. Turn ideas into software people l...