1030: Garbage In, Gospel Out: Why Agents Need Better Data, with Salesforce's Gaurav Pathak

During their #sponsored discussion, Senior Vice President Product Management AI and Metadata at Salesforce, Gaurav Pathak talks to Jon Krohn about why AI agents need well-labeled, high-quality data to deliver reliable answers in the enterprise. Listen to the episode to hear Gaurav Pathak talk about the difference between a “data brawl” and “garbage in, gospel out”, who the “sin eaters” of enterprise AI are and the three skills that matter most for AI engineers today! Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/1030⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information.⁠⁠⁠ In this episode you will learn: (03:05) Why metadata are the labels AI agents need (06:31) From “data brawl” to “garbage in, gospel out” (10:57) Who the “sin eaters” of enterprise AI are (13:45) What data quality rules are and how CLAIRE generates them (17:21) Three skills AI engineers need in the agentic era

In Episode #1029, Dr. Katie Malone (Host of Linear Digressions) joins Jon Krohn to explain how AI brought her podcast back from the dead. After nearly 300 episodes, Katie shut down Linear Digressions due to burnout, but better tools helped her relaunch it six years later. Along the way she has taught machine learning at Udacity and the University of Chicago and led the development of agentic AI platforms inside a company of tens of thousands of people. In this episode, she argues that people management and agent management are the same skill in different clothing, works through what AI slop and process slop are doing to organisations, describes the agent that now produces her show, and takes a pop quiz on three of her favourite data paradoxes. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1027⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:04:01) Why Linear Digressions stopped, and what changed enough to bring it back (00:15:01) Why people management and agent management are the same skill (00:24:32) The "Claude Code in a trench coat" agent that produces her show (00:41:39) Bainbridge’s ironies of automation, and why expertise gets rusty

In Episode #1028, Anton McGonnell (VP of Product at SambaNova) joins Jon Krohn to explain why the chips running most AI inference today were never designed for the job. Agentic AI has changed the computational profile of inference, with much larger inputs and far heavier caches feeding the token generation that follows, and that shift has exposed where GPU architecture struggles. SambaNova has raised over $2 billion to build an alternative, the reconfigurable dataflow unit, which lays a whole model out spatially across the chip rather than executing it kernel by kernel. In this episode, Anton discusses why the speed that matters is payback, and how speed and concurrency are what turn a fixed hardware cost into a six-month payback. He also walks through the trade-off every inference provider faces between speed per user and throughput per chip, what the RDU architecture changes about scaling and data center deployment, the economics of the new SN50, and why four out of five AI infrastructure leaders say they would pay a premium for faster tokens. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1028⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:02:41) Why agentic AI is reshaping inference workloads (00:08:39) How SambaNova's RDU differs from a GPU (00:17:54) The economics of the SN50

In Episode #1027, Dr. Dilani Kahawala (Co-Founder and CEO of Anna) joins Jon Krohn to explain what it takes to build an always-on AI assistant that busy parents will trust with their inboxes. Anna watches the email, school apps, WhatsApp messages and calendars flowing into a family's life and surfaces what matters, over text and voice, with barely any app to speak of. Dilani came to it by way of a Harvard physics PhD, McKinsey, and a decade of product leadership at Etsy, Meta and Atlassian, and says she has had to throw away most of what that decade taught her about how products get built. In this episode, she lays out the three hardest problems in building Anna, why the eval loop is the heart of the product, how a long-running agent differs from a turn-based one, and the brutal unit economics of consumer AI. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1027⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:10:01) The three hardest problems in building a consumer agent (00:13:23) Why a long-running agent is a different problem from a turn-based one (00:22:33) Why the eval and improvement loop is the heart of the product (00:26:42) The unit economics of always-on AI on a flat subscription

In Episode #1026, Jon Krohn breaks down GPT-6 Astra, OpenAI’s new flagship that its president has floated as a possible marker of AGI. Jon covers what the model is, what it costs, its state-of-the-art results across computer use, coding, abstract reasoning and science and the safety story, which for this release is unusually intertwined with capability. He weighs the AGI claim against Anthropic’s Fable 5.1 and lands, as ever, in a measured middle. Additional materials:⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/1026⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:14) What GPT-6 Astra is, and what it costs (03:58) The capability highlights that matter most (10:23) The safety story and the AGI question

In Episode #1025, Dr. Luis Serrano (Founder of Serrano Academy) joins Jon Krohn to explain the paper he co-authored on the curved spacetime of transformer architectures, in which attention stops being a lookup table and becomes something closer to gravity: words bend the space around them, and the embedding of "bank" visibly curves toward "river" as it travels through the layers of the network. In this episode, he recreates Eddington’s 1919 eclipse experiment inside a transformer, draws the line between an LLM workflow and an actual agent, explains why agent evaluation is a step harder than evaluating an essay, and gives the cleanest account of GRPO you will hear. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1025⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:10:53) What changed, and what survived, between the two editions of Grokking Machine Learning (00:27:48) Word gravity: how attention pulls "bank" toward "river" (00:42:12) Why RAG is an LLM workflow rather than an agent (00:51:23) The two-by-two that explains why GRPO powers reasoning models

In ICYMI Episode #1024, Jon Krohn tracks the gap between AI investment and AI return, from the technology side to the people side. Hear from Pete Johnson, Jerry Yurchisin, Priyanka Vergadia and Tristan Handy, discussing why four out of five organizations have the structures for AI success in place while only one in five sees the returns, which decisions should never be handed to a language model however confident it sounds, how to structure Claude skills so that your output stops being slop and why the semantic layer matters more, not less, now that analytics agents are the ones asking the questions. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/1024⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:56) Vector Search, Agentic Memory and Effective RAG (09:20) Mathematical Optimization in the Agentic AI Era (17:30) Anyone Can Write Code Now, So What Gets You Hired? (27:14) How dbt Won Analytics Engineering

In Episode #1023, Aishwarya Srinivasan (Co-Founder of The Gen Academy) joins Jon Krohn to work out where a competitive moat comes from once anything you can build in ten minutes, somebody else can build in ten minutes too. Ash came to teaching through Illuminate AI, the mentorship community she started in 2020, and now trains senior engineers and leaders to ship agentic AI in production; she is blunt that vibe coding lowers the floor without touching the engineering judgment that production demands. In this episode, she explains what a whole-system eval covers that a model eval misses, traces reinforcement learning from the algorithm she patented at IBM to its resurgence in agentic fine tuning and lays out the MIND framework from her TED Talk for living with AI. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1023⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:10:10) Why cheap code shifts the software engineering job rather than ending it (00:15:50) What a whole-system eval covers that a model eval misses (00:36:11) Why reinforcement learning came roaring back for agentic AI (00:41:23) The one skill Ash says matters more than any hard skill

In Episode #1022, Jon Krohn tackles the art of steering AI agents, deciding where your instructions should live so they get followed reliably without bloating every request. A sequel to Episode #1020 (where model size and effort set an agent’s horsepower), this one is about direction: the seven ways to deliver instructions, why a hook beats a prompt, the industry-wide agents.md standard, and three practical takeaways you can apply whatever your stack. Additional materials:⁠ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠www.superdatascience.com/1022⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (02:52) The seven ways to deliver instructions to an agent (06:42) Why a hook is a guarantee and an instruction is only a probability (13:00) Three takeaways for organizing your instructions

In Episode #1021, Tristan Handy (Founder and CEO of dbt Labs) joins Jon Krohn to explain how a study of about a hundred companies in 2016 became analytics engineering, and then became a tool that over a hundred thousand data teams rely on. Tristan coined the term, chose SQL when Spark was the fashionable answer, and spent a decade turning down acquisition offers because none of them were good for the people using dbt. He is now merging dbt Labs with Fivetran and taking on the presidency of the combined company, the first deal he says cleared that bar. In this episode, Tristan walks through what dbt does to your raw data, argues that the semantic layer matters more once analytics agents are asking the questions, explains the type safety behind the Fusion engine, and details how a 12-kilobyte skill file collapses a million-dollar migration into six weeks. Additional materials: ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.superdatascience.com/1021⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ Interested in sponsoring a SuperDataScience Podcast episode? Email [email protected] for sponsorship information. In this episode you will learn: (00:07:22) Why Tristan chose SQL over Spark, and what progressive complexity means (00:10:44) How a dbt project turns raw data into modeled tables (00:17:50) Why a decade of acquisition offers kept failing his one test (00:40:47) How 12-kilobyte skill files cut year-long migrations to six weeks