This 24-Year-Old Raised $64M to Build an AI Smarter Than the World's Best Mathematicians

Imagine an AI that doesn’t just output answers — it remembers, adapts, and reasons over time like a living system. In this episode of The Neuron, Corey Noles and Grant Harvey sit down with Zuzanna Stamirowska, CEO & Cofounder of Pathway, to break down the world’s first post-Transformer frontier model: BDH — the Dragon Hatchling architecture.Zuzanna explains why current language models are stuck in a “Groundhog Day” loop — waking up with no memory — and how Pathway’s architecture introduces true temporal reasoning and continual learning. We explore:• Why Transformers lack real memory and time awareness • How BDH uses brain-like neurons, synapses, and emergent structure • How models can “get bored,” adapt, and strengthen connections • Why Pathway sees reasoning — not language — as the core of intelligence • How BDH enables infinite context, live learning, and interpretability • Why gluing two trained models together actually works in BDH • The path to AGI through generalization, not scaling • Real-world early adopters (Formula 1, NATO, French Postal Service) • Safety, reversibility, checkpointing, and building predictable behavior • Why this architecture could power the next era of scientific innovationFrom brain-inspired message passing to emergent neural structures that literally appear during training, this is one of the most ambitious rethinks of AI architecture since Transformers themselves.If you want a window into what comes after LLMs, this interview is essential.Subscribe to The Neuron newsletter for more interviews with the leaders shaping the future of work and AI: https://theneuron.ai

Carina Hong dropped out of Stanford's PhD program to build "mathematical superintelligence" — and just raised $64M to do it. In this episode, we explore what that actually means: an AI that doesn't just solve math problems but discovers new theorems, proves them formally, and gets smarter with each iteration. Carina explains how her team solved a 130-year-old problem about Lyapunov functions, disproved a 30-year-old graph theory conjecture, and why math is the secret "bedrock" for everything from chip design to quant trading to coding agents. We also discuss the fascinating connections between neuroscience, AI, and mathematics.Lean more about Axiom: https://axiommath.ai/ Subscribe to The Neuron newsletter: https://theneuron.ai

Nick Talken started a 3D printing materials company in a trailer lab in his co-founder's backyard, sold it to a 145-year-old German chemical giant, then spun out an AI platform that's now transforming R&D for Fortune 100 companies. Albert Invent's foundational AI model—trained on 15 million molecular structures—is helping scientists at companies like Kenvue (maker of Tylenol, Neutrogena, and Listerine) compress projects from 3 months to 2 days. We dig into how enterprises train bespoke AI models on proprietary data, why you can't just use ChatGPT for chemistry, and what becomes possible when AI can "think like a chemist."Subscribe to The Neuron newsletter: https://theneuron.aiAlbert Invent website: https://www.albertinvent.comKenvue partnership announcement: https://www.businesswire.com/news/home/20251014240355/en/

Most enterprise knowledge is trapped in meetings—and then lost forever. Otter.ai CEO Sam Liang explains how his company turned meeting transcription into a $100M+ revenue business by solving a problem most companies don't even realize they have.In this episode, we cover:- Why meetings are your company's most expensive activity (and how to measure ROI on them)- Building a "meeting-centric knowledge base" that captures voice data other systems miss- How Otter organizes enterprise knowledge like Slack—but for spoken conversations- Real-time sales coaching that feeds reps answers during customer calls- AI avatars that attend meetings on your behalf (and ask questions for you)- The technical challenges of understanding dialects, tone, and context in voice AI- How one financial company used Otter to onboard new clients instantly with full conversation history- Privacy vs. utility: designing permission systems for meeting data- The future of active AI agents that contribute to meetings, not just transcribe themSam previously worked on the blue dot location platform for Google Maps and now runs a company that's transcribed over 1 billion meetings. If you're thinking about how AI can actually improve enterprise workflows (not just automate busywork), this conversation is packed with specific, tactical insights.A special thank you to this episode's sponsor, SAS: https://www.sas.com/en/whitepapers/how-aiot-is-reshaping-industrial-efficiency-security-and-decision-making.html?utm_source=other&utm_medium=cpm&utm_campaign=-globalResources mentioned:• Otter.ai $100M ARR announcement: https://otter.ai/blog/otter-ai-breaks-100m-arr-barrier-and-transforms-business-meetings-launching-industry-first-ai-meeting-agent-suite• HIPAA compliance: https://otter.ai/blog/otter-ai-achieves-hipaa-compliance• Otter.ai: https://otter.aiSubscribe to The Neuron newsletter: https://theneuron.ai➤ CHAPTERS0:00 - Introduction & Sam's Background1:16 - From Meeting Notes to Enterprise Knowledge04:48 - Building a Meeting-Centric Knowledge Base06:14 - Why Meetings Are Your Most Expensive Activity05:40 - Solving Information Silos with AI07:56 - A Message from our Sponsor SAS9:11 - Meeting Transcriptions Alone Aren't the Answer17:34 - Leader Dashboards & AI Workflows18:49 - AI Avatars: Send Your Digital Self to Meetings21:45 - Active AI Agents That Talk Back23:13 - Privacy, Permissions & Corporate Culture26:08 - Technical Challenges: Understanding Context & Tone34:37 - Privacy vs. Utility Trade-offs37:25 - The Future of Meetings in 202739:27 - Competing with Microsoft & Google43:02 - How Otter Generated Over $1 Billion in Customer ROI46:05 - What Excites & Concerns Sam About AI49:09 - Security Risks of AI Avatars49:50 - Final Thoughts on the Future of AI at WorkHosted by: Corey Noles and Grant HarveyGuest: Sam Liang, Co-founder & CEO, Otter.aiPublished by: Manique SantosEdited by: Kush Felisilda

In this episode, we sit down with Pavan Davuluri, Corporate Vice President of Microsoft's Windows + Devices business, to explore how Windows is evolving into an AI-native platform. Pavan leads the team responsible for strategy, design, and delivery of Windows products across the full stack - from silicon and devices to platform, OS, apps, experiences, security, and cloud. With 23 years at Microsoft, he's driven the creation of the Surface line and now oversees how hardware and software fuse together with AI at the center. We explore how Copilot is being deeply integrated into Windows, the engineering shifts required to make Windows a more proactive and intelligent platform, and how Microsoft balances powerful automation with user control. From Surface design standards influencing the broader ecosystem to supporting OEM partners in the AI PC era, Pavan reveals the principles guiding Windows' transformation and what the computing experience will look like in the next five years.Subscribe to The Neuron newsletter: https://theneuron.aiMicrosoft Surface: https://www.microsoft.com/surfaceWindows AI features: https://www.microsoft.com/windows/ai-features

While everyone obsesses over which AI model is smartest, a quiet revolution is happening in the infrastructure layer underneath. Modular just raised $250M at a $1.6B valuation to solve a problem most people don't know exists: AI is locked into expensive, vendor-specific hardware ecosystems. Tim Davis, Co-Founder & President of Modular, joins us to explain why his company is building the "hypervisor for AI"—making it possible to write code once and run it on any GPU, from NVIDIA to AMD to Apple Silicon. We dive into why this matters for businesses, what the Android analogy really means, how companies are seeing 70-80% cost reductions, and whether we're even on the right path to superintelligence.Subscribe to The Neuron newsletter: https://theneuron.aiTry Modular: https://modular.comGetting Started Guide: https://modular.com/get-started

In this episode, we sit down with Scott Guthrie, EVP of Microsoft's Cloud + AI Group, to explore the architecture behind Azure's AI Superfactory. Scott oversees Microsoft's hyperscale cloud computing solutions including Azure, generative AI platforms, and next-generation infrastructure. We dive into Microsoft's strategic approach to AI datacenter buildout, the innovative Fairwater architecture with its 120,000+ fiber miles of AI WAN backbone, and how Microsoft is balancing performance, sustainability, and cost at planet-scale. From dense GPU clusters drawing 140kW per rack to closed-loop liquid cooling systems, Scott reveals the engineering trade-offs behind infrastructure that powers frontier AI models with trillions of parameters. Whether you're an enterprise leader planning AI adoption or a developer curious about cloud architecture, you'll leave understanding how Microsoft is executing on next-gen infrastructure that transforms global challenges into opportunities.Subscribe to The Neuron newsletter: https://theneuron.ai

Retool CEO David Hsu reveals that 48% of non-engineers are now shipping software. We explore how AI is democratizing software development, why engineers might stop coding internal apps within 18-24 months, and what this means for the future of work. David shares insights from Retool's survey of 10,000+ companies, Retool’s new AppGen program, and how "tomorrow's developers" are using AI to build real production applications on enterprise data.Subscribe to The Neuron newsletter: https://theneuron.aiLearn more about Retool: https://retool.com

Computers can see and hear, but they've never been able to smell—until now. In this episode, we sit down with Alex Wiltschko, Founder & CEO of Osmo, to explore how his company is using AI to digitize scent. Alex walks us through how they "teleported" the smell of a fresh plum across their lab, created the world's first AI-designed fragrance molecules, and built Osmo Studio—a platform that lets anyone design custom fragrances in one week instead of two years. We discuss the read/map/write framework for digitizing smell, why scent is tied directly to memory and emotion, and how this technology could eventually detect diseases like cancer and Parkinson's earlier than any current diagnostic. Plus: what does the Museum of Pop Culture smell like, and can AI really create a fragrance from a Bon Iver song?Links: Osmo: https://www.osmo.aiScent Teleportation Update: https://www.osmo.ai/blog/update-scent-teleportation-we-did-itOsmo Studio: https://osmostudios.ai/ Subscribe to The Neuron newsletter: https://theneuron.aiCheck out the sponsor of this video, Flora: https://dub.florafauna.ai/neuronSubscribe to The Neuron newsletter: https://theneuron.aiHosted by: Corey Noles and Grant HarveyGuest: Alex Wiltschko (Founder & CEO, Osmo)Published by: Manique SantosEdited by: Kush Felisilda

Behind every AI response, there's an invisible army of humans who trained it. In this episode, we talk with Casper Elliott from Invisible Technologies - the company that's trained 80% of the world's top AI models. We explore how models actually learn, why data quality matters more than quantity, what enterprises get wrong about AI deployment, and whether AI will really automate everyone's jobs. Casper shares insights from working with frontier labs, reveals the surprising skills that make great AI trainers (hint: League of Legends helps), and explains why the future needs more humans, not fewer.Subscribe to The Neuron newsletter: https://theneuron.aiLearn more about Invisible Technologies: https://invisibletech.ai