Épisodes

  • 998: In Case You Missed It in May 2026

    998: In Case You Missed It in May 2026
    Jun 5 2026
    In this month’s episode of ICYMI, Jon Krohn explores how AI agents are simultaneously creating new risks and unlocking powerful new ways of working with data. Hear from Anneka Gupta, Cal Al-Dhubaib, Trevor Manz, Jazmia Henry, Jeremy Mumford, and Jacob Miller, discussing why the old cybersecurity playbook breaks down in the age of Claude Mythos, how the notebook became an AI agent’s working memory, what it really takes to build a foundation model from scratch, and why failing slowly is the most expensive mistake an AI team can make. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/998⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (00:40) Why Claude Mythos Changes Everything About Cybersecurity (08:11) Why Your Notebook Should Be Your Agent’s Working Memory (13:19) What It Actually Takes to Build a Foundation Model From Scratch (20:46) Failing Slowly Is the Most Expensive AI Mistake
    Afficher plus Afficher moins
    28 min
  • 997: How This Text-to-Video-Game AI Startup Hit 20M Users

    997: How This Text-to-Video-Game AI Startup Hit 20M Users
    Jun 2 2026
    Dr. Andrey Kurenkov returns to the show to talk about Astrocade's astronomical growth from pre-alpha to over 20 million engaged users, what it actually takes to build a vibe-coding platform that scales, and how the broader AI landscape has shifted since his last appearance. Andrey shares behind-the-scenes lessons from building B2C user-generated content products, why the real moat is community rather than tech, and his current thinking on humanoid robotics, AGI, and the AI risks people actually overlook. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/997⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (02:11) The Astrocade elevator pitch and how it grew to 20M users (16:19) Why there's no secret sauce behind the platform (24:56) UGC as the real moat, not the AI (46:57) Why household humanoid robots are now 2–3 years away (58:33) What AGI actually means, and why Andrey is an ASI skeptic
    Afficher plus Afficher moins
    1 h et 10 min
  • 996: TrueFoundry’s Nikunj Bajaj on How to Get $100M Returns on AI Agent Deployments

    996: TrueFoundry’s Nikunj Bajaj on How to Get $100M Returns on AI Agent Deployments
    May 29 2026
    TrueFoundry co-founder and CEO Nikunj Bajaj speaks to Jon Krohn about how enterprises like Nvidia and Siemens are realizing returns of over $100 million from single agent deployments, the AI gateway architecture that makes it possible to connect, observe, and govern agents at scale, and why the familiar advice to “start small” is the wrong way to roll out AI agents inside a large organization. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/996 Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.⁠⁠⁠ In this episode you will learn: (01:21) What TrueFoundry does and why agents in production need a control plane (06:32) Breaking down the AI gateway: the model, MCP, and agent gateways (16:47) Taming tool sprawl with scoped, read-only MCP access (19:10) Why the agent gateway is the hard part and the kill switch most teams lack (22:24) The five-workflow framework behind $100M agent deployments
    Afficher plus Afficher moins
    30 min
  • 995: End-to-End Foundation Models for the Energy Industry, with Jazmia Henry

    995: End-to-End Foundation Models for the Energy Industry, with Jazmia Henry
    May 26 2026
    Jazmia Henry joins Jon Krohn to break down what it actually takes to build end-to-end foundation models for the energy industry. From wrangling decades of handwritten oil-and-gas documents into usable training data, to bespoke tokenizers, reinforcement learning, and inference at scale, Jazmia walks through every stage of the stack. Along the way she explains why reinforcement learning models are "bursty," what reward hacking is and how her Grounded Continuous Evaluation framework fixes it, and revisits the 2023 NeurIPS paper that argued, to widespread skepticism at the time, that scaling bad data degrades model performance. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/995⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (10:06) The User Agnosticism Tenet (20:02) The Zillow Offers parable (23:25) Why workflows should come before agents (29:57) Why data engineering is the bedrock of AI (52:41) Why velocity is the only durable moat
    Afficher plus Afficher moins
    1 h et 9 min
  • 994: AI’s Putting Recent Grads Out of Work; Here’s How to Get Hired Anyway!

    994: AI’s Putting Recent Grads Out of Work; Here’s How to Get Hired Anyway!
    May 22 2026
    Unemployment for recent computer-science graduates now rivals rates for fine-arts and anthropology majors, and undergraduate CS enrollment fell 11% in 2025. In this Five-Minute Friday, Jon Krohn walks through the data on both sides of the debate, from Stanford research showing a 13% employment drop for young workers in AI-exposed jobs, to Federal Reserve studies finding no statistically detectable link between AI adoption and reduced hiring. Jon shares his own view on where the truth lies and offers five concrete pieces of advice for graduates and senior professionals alike on how to get hired in 2026. Additional materials:⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/993⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.
    Afficher plus Afficher moins
    11 min
  • 993: How to Build AI-First Organizations, with Jacob Miller and Jeremy Mumford

    993: How to Build AI-First Organizations, with Jacob Miller and Jeremy Mumford
    May 19 2026
    For years, AI content has come in the form of “use this library, use this tool” tutorials that age out within months. Jacob Miller and Jeremy Mumford, co-authors of the brand new Wiley book Architected Intelligence, wanted to write something different, a guide to the higher-level principles of building AI products and AI-first organizations that will still be relevant in five or ten years. In this episode, the two Pattern engineers walk Jon Krohn through the core ideas of their book: why you should design products and processes so they can be executed by a human, an AI agent, or any hybrid combination; why most companies are still treating hallucinations as a model problem when they’re actually a data curation problem; why the natural progression of AI development goes skills, workflows, agents, not straight to agents; and why velocity, not models or data, is the only durable competitive advantage left. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/993⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (10:06) The User Agnosticism Tenet (20:02) The Zillow Offers parable (23:25) Why workflows should come before agents (29:57) Why data engineering is the bedrock of AI (52:41) Why velocity is the only durable moat
    Afficher plus Afficher moins
    1 h et 10 min
  • 992: Tokenmaxxing vs AI Hardware Bottlenecks

    992: Tokenmaxxing vs AI Hardware Bottlenecks
    May 15 2026
    While “tokenmaxxing”, the social media trend of maximizing AI token consumption as a vanity metric, takes off online, the physical infrastructure behind AI is slamming into serious bottlenecks. In this Five-Minute Friday, Jon Krohn maps out the four overlapping supply-chain constraints choking AI compute: GPUs (with NVIDIA Blackwell sold out through mid-2026), high-bandwidth memory (quintupled demand since 2023, only three manufacturers worldwide), CPUs (agentic AI requires 12x more CPUs per GPU than chatbots), and electricity (Gartner projects power shortages will restrict 40% of AI data centres by 2027). Find out why the five biggest hyperscalers are on track to spend $725 billion on AI infrastructure in 2026, where the reasons for optimism lie, and why Jon says you should definitely not tokenmaxx. Additional materials:⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/992⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information.
    Afficher plus Afficher moins
    15 min
  • 991: Pair Programming with AI in Your Python Notebook, with Dr. Trevor Manz

    991: Pair Programming with AI in Your Python Notebook, with Dr. Trevor Manz
    May 12 2026
    Dr. Trevor Manz of Marimo talks to Jon Krohn about Marimo Pair, an open-source agent skill that teaches coding agents like Claude Code how to drive a reactive Python notebook, reading cell state, running Python in the kernel, taking screenshots of cells, and iterating on data tasks the way agents iterate on traditional software. Trevor also unpacks recursive language models, his AnyWidget project that bridges Python and the web, and his journey from a Wisconsin small town and Harvard bioinformatics research to founding-engineer life at Marimo. Listen to the episode to hear why no matter where AI takes us, curiosity and going deep on a topic will always be valuable. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/991⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email natalie@superdatascience.com for sponsorship information. In this episode you will learn: (07:04) What Marimo Pair is and how it teaches agents to use notebooks as a tool (13:03) How agent skills work as folders of markdown files (24:15) Trevor's day-to-day workflow combining Claude Code and Marimo Pair (31:51) Recursive language models and why they could be the future of agentic reasoning (57:33) Career advice on curiosity, going deep, and becoming a domain expert
    Afficher plus Afficher moins
    1 h et 9 min