Dr. @RichardSocher, CEO of Recursive, tells @JonKrohnLearns Krohn how AI could deliver a century of scientific breakthroughs in a decade, why recursive self-improvement may arrive within a year without an overnight takeoff, and which software and jobs will endure. Watch the episode here: superdatascience.com/1037

Carly Taylor founder of Rebel Data Science, tells @JonKrohnLearns why trying to predict where #AI is headed leads to paralysis, how leaders can buy their teams room to explore, and why she deliberately chases the situations that scare her. Watch the episode here: superdatascience.com/1035

Prof. Jeff Hancock and Dr. Kate Niederhoffer, who coined the term workslop, tell @JonKrohnLearns why AI-generated work costs the recipient more than it saves the sender, what actually drives people to produce it and what relation slipping does to a team. Watch the episode here: superdatascience.com/1033

Ish Shah and Tyler Cox of Dell Technologies tell @JonKrohnLearns why agentic workloads devour so many more tokens than chatbots did, how to pick a model that is merely good enough and what moving the work under your own desk does to the bill. Watch the episode here: superdatascience.com/1031

Dr. Caitlin Malone, host of Linear Digressions, tells @JonKrohnLearns how AI let her restart a podcast abandoned for five and a half years, why managing agents is the same skill as managing people, and what Bainbridge's ironies of automation warn us about. Watch the episode here: superdatascience.com/1029

Dr. Dilani Kahawala, co-founder and CEO of Anna, tells @JonKrohnLearns what it takes to build a consumer agent with no app UI, why long-running agents are a different problem from Claude Code and how a $20 subscription covers always-on tokens. Watch the episode here: superdatascience.com/1027

Dr. Luis Serrano, founder of Serrano Academy, tells @JonKrohnLearns how transformer embeddings bend space the way the sun bends starlight, why RAG is a workflow and not an agent, why agent evaluation is harder still, and how GRPO powers reasoning models. Watch the episode here: superdatascience.com/1025

Aishwarya Srinivasan, co-founder of The Gen Academy, tells @JonKrohnLearns where defensibility comes from when code is nearly free, how to evaluate non-deterministic agentic systems end to end, why reinforcement learning is roaring back, and which career skill matters most now. Watch the episode here: superdatascience.com/1023

Tristan Handy, founder of dbt Labs, tells @JonKrohnLearns why he chose SQL over Spark, how the semantic layer stops agents getting metrics wrong, why he turned down a decade of acquisitions, and how skill files compress year-long migrations into six weeks. Watch the episode here: superdatascience.com/1021

Priyanka Vergadia, founder of The Cloud Girl, tells @JonKrohnLearns why AI raised the technical floor and made taste the ceiling, how to structure Claude skills so your output stops being slop, and why 70% of an AI budget belongs in upskilling. Watch the episode here: superdatascience.com/1019

Pete Johnson, Field CTO of AI at MongoDB, tells @JonKrohnLearns which AI deployments are actually returning ROI, why the embedding model you pick can make or break a RAG pipeline and how tokenmaxxing became a cautionary tale of AI vanity metrics. Watch the episode here: superdatascience.com/1017

Cathy O'Neil Harvard math PhD, ex-quant, author of Weapons of Math Destruction tells @JonKrohnLearns why simple algorithms terrify, how AI degrades rather than replaces workers, why auditing an algorithm means auditing a bureaucracy, and how to demand accountability. Watch the episode here: superdatascience.com/1013

Dr. Catherine Williams, one of the world’s first data scientists and now Chief Data Officer at the nonprofit Candid, tells @JonKrohnLearns why going deep on the math behind machine learning matters more than ever, how she was hired at AppNexus for her mathematical fluency despite having almost no data skills, why fine-tuning has largely gone the way of the dinosaur and why the most valuable data professionals are the ones who can move up and down the whole stack of complexity. Watch the episode here: superdatascience.com/1011

Investor and entrepreneur Steve Mock joins @JonKrohnLearns to share the real, human stories behind @aisavedme.org, from a founder finally understood by her family, to a dog’s cancer treatment tracked down with AI, plus how he built the entire site himself as a self-described non-programmer by letting “the AI manage the AI.” Watch the episode here: superdatascience.com/1009

80,000 Hours co-founder @ben_j_todd returns to tell @JonKrohnLearns why there’s no permanent “solid ground” in an AI career, only a fast-moving frontier to chase, why extreme power concentration may be the world’s most neglected problem, and why a team of five can now do what once took thirty, making how you spend the next few years matter more, not less. Watch the episode here: superdatascience.com/1007

Bestselling author and former pro poker player Gilbert Eijkelenboom tells @JonKrohnLearns Krohn why the people side of data science matters more than the math, from the “And, But, Therefore” storytelling framework, to why only about 15% of us are truly self-aware, to why doing uncomfortable things for 30 days actually rewires your beliefs. Watch the episode here: superdatascience.com/1005

Frank Basso, VP of Infrastructure at Lightning AI, tells @JonKrohnLearns how he builds AI data centers end to end, from co-locating tens of thousands of GPUs and 20-megawatt liquid-cooled superpods, to why “screaming banshee” server halls hit 110 decibels, and why the data center isn’t actually why your power bill went up. Watch the episode here: superdatascience.com/1003

In a role reversal for this episode #1001 special, @JonKrohnLearns becomes the guest as podcast founder Kirill Eremenko interviews him about how AI erased the technical career moat he spent 20 years building, why that paradoxically makes skilled people more valuable than ever, whether we’re in an AI bubble, and the number-one reason AI projects fail. Watch the episode here: superdatascience.com/1001

TrueFoundry’s Nikunj Bajaj speaks to @JonKrohnLearns about how enterprises like Nvidia and Siemens are realizing over $100 million in returns from single AI agent deployments, and why the conventional advice to “start small” is exactly the wrong way to roll out agents in a large organization. Watch the episode here: superdatascience.com/996

Jazmia Henry tells @JonKrohnLearns how she builds foundation models for the oil and gas industry, where 90% of documents are unstructured and an AI model's accuracy can be the difference between a routine shift and a worker losing a limb. Watch the episode here: superdatascience.com/995