The Joe Reis Show

Tristan Handy (dbt/fivetran) joins me to discuss the tectonic shifts reshaping data teams, analytics engineering, and business intelligence in the age of agentic AI.

Tristan opens up about stepping back into "builder mode" after years as CEO, the parallels between the dot-com era and today's AI wave, and why data teams need to stop clinging to yesterday's work, such as dashboard maintenance and SQL ticket queues.

We also chat about the reality of context engineering, why traditional BI and point-and-click tools face an existential reckoning, how dbt charts approaches agentic abstractions, and why data practitioners must become deeply embedded business partners rather than pure pipeline mechanics.

The old data industry is dying. The new one hasn't quite been born. And in between, we're seeing some strange things.

In this episode, I explore the Great Data Interregnum, a period when the playbook, practices, and architectures of yesterday are losing their coherence, but whatever replaces it remains uncertain.

In this episode of The Joe Reis Show, I sit down at my house in Salt Lake City with Kyle, founder of Credible Data.

Before founding Credible, Kyle spent nearly 17 years at Google building core distributed systems—including the in-memory shuffle/join backplane that powers BigQuery—before moving to Gradient Ventures and spending four years helping integrate Looker into GCP.

We dive deep into the reality of semantic layers, the origins of Malloy, and why the current hype around semantic layers often misses the mark. Kyle breaks down why SQL conflates data modeling and queries, how agents are shifting data work from painful bottom-up ETL to deterministic top-down domain modeling, and why giving an LLM an entire enterprise schema creates nothing but "data slop." We also get into the future of data engineering, what analysts should actually be learning today, and why unlocking the "long tail" of dark business data is where the true value lies.

Connect with Kyle & Credible: credibledata.com

G​uest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

Ever feel like you've been lied to?

In Scott’s Tots, Michael Scott promises a class of third graders he’ll pay their college tuition if they finish high school. Ten years later they’ve held up their end, and he shows up to tell them he's broke.

A lot of people feel like those kids right now. They got the degree, learned to code, and followed the career plan they were handed, only to find a job market that feels empty and hollowed-out.

In this episode I use Scott’s Tots to look at who made these promises, why most of them were sincere but wrong, and what the consolation prizes look like today.

Guest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

In this episode, I sit down with Doug Meil, software veteran and author of The Rise and Fall of Explorys and IBM Watson Health. A decade ago, IBM Watson was the undeniable face of the modern AI revolution. Fresh off its historic Jeopardy win, IBM promised Watson would revolutionize entire industries and cure cancer. Instead, it became one of tech history’s most high-profile corporate implosions. Doug lived through the heart of it. His healthcare data startup, Explorys, was acquired in 2015 to serve as a foundational pillar for IBM Watson Health. What followed was a masterclass in what Doug calls the "paradox of inorganic growth"—an unravelling driven by corporate hubris, aggressive internal "blue washing," brutal cloud lock-ins, and leadership prioritizing marketing slogans over iterative product engineering. We dig deep into the untold engineering and product side of the acquisition, why enterprise mergers crush agile technical teams, and the enduring lessons today’s AI founders and leaders must take away before jumping headfirst into enterprise AI hype.Get Doug's book here: https://link.amazon/B05ATbGBL

Guest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

In this episode, I'm joined by Rob Collie, CEO of P3 Adaptive, early pioneer of Power BI and Power Pivot at Microsoft, and author of the new book Fair Game: Customizing AI to Your Business is Easier Than You Think. Most business leaders play with off-the-shelf ChatGPT or Claude and think it's magic until it completely face-plants on their actual company data. Why? Because the LLM is just one Lego brick. Delivering real value requires the unsexy, "boring" work: structured context, reliable data modeling, and robust semantic layers.Rob and I dig into why 15 out of 16 people bounce off Excel, why semantic layers are hot, if dashboards are dead, and much more.

A lot of people ask me for career advice. My latest advice is learn the fundamentals, build useful things that interest you, and stop outsourcing your direction to everyone around you.

But originality doesn’t automatically pay the bills, copying is part of learning, and “do what AI cannot” is lazy advice.

In this episode, I unpack the advice, its limits, and how to find your own way without betting your livelihood on it.

Guest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

The data industry rarely moves in a straight line. We swing from centralized to decentralized, batch to real-time, general-purpose to specialized, abstraction to control, and governance to self-service.

In this episode, I unpack why these cycles keep repeating, what forces push the pendulum in each direction, and why every successful architecture eventually creates the conditions for its own backlash.

We see a lot of hype cycles in our industry - Data mesh. Generative AI. Agents. Now context.

Every cycle brings a new label and a familiar promise: this will finally make your organization’s data useful.

The technology has improved enormously. So why do the same problems keep coming back?

In this episode, I explore how communication, incentives, and organizational politics shape what our tools actually deliver. Will context finally deliver the silver bullet to solve our problems? Is buying the next platform easier than resolving the messy organizational problems we all deal with? Listen to get my thoughts.

Hannes Mühleisen (co-creator of DuckDB) joins me to chat about the past, present, and future of DuckDB, Ducklabs (recently acquired by AWS), and more.

Note - parts edited due to AWS's request.

Guest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

Former Gartner VP Analyst and principal at SanjMo, Sanjeev Mohan, joins me to break down the reality of AI adoption, data foundations, and why enterprises are hitting roadblocks with agents and LLMs. From comparing the current AI boom to the dot-com era to unpacking his new concept, ECL (Extract, Context, Link), Sanjeev explains why moving data around is no longer viable and why "there is no AI stack, only one stack." We also dive into capability overhang, artificial guardrails, the race to standardize semantic context, and why writing books and understanding core primitives still matter in the age of generative models

Guest policy: The Joe Reis Show is strictly invite-only. Unsolicited guest pitches, PR agency solicitations, and booking requests are not accepted or reviewed.

This week at dbt Summit, Tristan Handy (co-founder of dbt) asked me why I write books. It's a good question, especially today when reading is declining and any answer is seemingly at your fingertips.

In this Freestyle Friday, I discuss why I write books, why they still matter, and much more. If you're interested in writing a book, this episode is definitely a must-listen.

I talk to executives and practitioners around the world about AI and what comes next. Everyone has predictions and pontifications, and everyone seems to think someone else has the answers.

But the more conversations I have, the clearer it gets that nobody knows what's next.

In this episode, I explore that uncertainty and how we move forward anyway.

Because in the end, nobody knows...

In this episode of The Joe Reis Show, I catch up with my good friend Christophe Blefari, co-founder of NAO, to dig into what's actually happening on the front lines of AI agents and data engineering. We talk about the evolution from building SQL IDEs to agentic analytics harnesses, how prompt-driven workflows are reshaping data teams, and why age-old problems like data discovery, modeling, and documentation are suddenly roaring back. We also dive into the risks of "token slop," whether data and software engineering roles are destined to merge, Christophe’s take on DuckDB’s acquisition by AWS, and the realities of building AI tools across Europe and the US.Nao: https://getnao.io/

In this episode, I sit down with Sami Hero, CEO of Ellie.ai, to talk through the evolution of data modeling and why foundational conceptual modeling is more critical now than ever. With the explosion of AI agents, LLMs, and semantic layers, many teams assume they can bypass data modeling and just let an AI infer meaning directly from raw tables.

We dig into why that mindset leads to bad decisions, how companies end up with 15,000 Snowflake tables and thousands of unmaintained dashboards, and why defining core business concepts is the real prerequisite for reliable AI. We also discuss the resurgence of ontologies, lessons from old-school methodologies like MERISE, and how to balance speed with sound architecture.

AI agents are being pitched as the “death of the data team,” with CEOs supposedly able to do the work themselves. We’ve heard versions of this promise before, from self-service BI to low-code analytics.

In this episode, I revisit that history and ask what agents actually change. Who prepares the data, resolves conflicting definitions, and checks the answers? What does the hiring research tell us about demand for experienced people and opportunities for juniors?

In this episode of the Joe Reis Show, I sit down with David Aronchick, co-founder of Expanso and one of the early pioneers behind Kubernetes, Kubeflow, and the CNCF. We dive deep into the evolving data landscape and discuss why the architectural pendulum is swinging back from pure cloud centralization toward distributed edge compute. David breaks down the real operational nightmares of edge data collection, why treating your bronze tier as a "toxic waste dump" breaks downstream pipelines, and how applying schema and context upstream makes your data truly AI-ready. We also get into the critical need for data provenance, data bills of materials, and what agents actually need to communicate reliably.Website: https://expanso.io

I get asked all the time whether I’m still “hands-on.” The short answer: not in the traditional sense. My work has shifted toward writing, teaching, advising, investing, building companies, and working with the people building what comes next. In this episode, I argue that technical relevance isn’t about staying trapped in the same operating mode forever, and that AI is drastically changing what “hands-on” even means.

------------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit ⁠⁠⁠⁠⁠fivetran.com⁠⁠⁠⁠⁠ to learn more.

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

On this episode of The Joe Reis Show, I’m joined by Dan Bennett, Head of Technology for the Enterprise Data Organization at S&P Global.We get into what it actually looks like to build with modern AI coding tools like Claude Code. Not just as a toy, but for writing production-grade C++. Dan walks through how he built an open-source RDF extension for DuckDB, why rock-solid test coverage is non-negotiable when working with LLMs, and why engineering leaders need to keep their hands dirty to understand where this tech is headed. We also react to the breaking news of AWS acquiring DuckDB Labs, talk through the shift from "human-in-the-loop" to autonomous agents running in headless VMs, and dive into why data semantics across organizational boundaries remains one of the hardest - and most important - unsolved problems in our industry.Website: https://nonodename.com/

I discuss the The Mixed Model Arts Manifesto. That's it. That's the show.

Will AI agents solve the data industry's hardest challenges, or are we just amplifying existing chaos?In this episode, I chat with Oliver Laslett, Co-founder and CTO at Lightdash, to explore the frontier of agentic analytics, data modeling, and modern software engineering practices. We dive deep into how AI is shifting daily workflows from writing SQL and DBT models to higher-order system thinking, why the hardest data problems remain human and organizational, and the cultural differences in tech optimism between London and San Francisco.Oliver also breaks down why permissions, curation, and data lineage are still the true bottlenecks, and why high-ownership communication matters more than ever in an era of AI slop.

It feels like everything is converging right now. Data forms and systems are converging. Agile and Waterfall are starting to look strangely compatible. Software engineering, data engineering, and AI engineering are bleeding into each other.

In this episode, I explore why AI is collapsing boundaries we spent decades creating, and why semantics, architecture, specification, and judgment may matter more as a result.

-----------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit ⁠⁠⁠⁠fivetran.com⁠⁠⁠⁠ to learn more.

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

Are massive frontier AI models actually worth the cost?In this episode, I sit down with Daniel Svonava, Founder of Superlinked, to break down where AI is actually heading in production. We dive into why the hype around massive frontier models is shifting toward targeted, small open-source models, and how to orchestrate smart agents without blowing through your budget.Daniel shares how Superlinked is building open-source inference infrastructure, why small models are rapidly closing the performance gap, and how you can engineer faster, cheaper, and more control-driven AI systems.

Every so often, the tech industry declares something dead. Data warehousing is dead. Data modeling is dead. Semantic layers are dead. SQL is dead. Now AI is supposedly making entire professions obsolete. I get it. Ragebait gets the clicks and comments. But it's also lame and disingenuous.

In this Freestyle Friday, I unpack the logical fallacies behind these claims, why technology almost always evolves instead of replaces, and how separating fundamentals from shapeshifting implementations gives you a much clearer view of where the industry is actually headed.

-----------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit ⁠⁠⁠fivetran.com⁠⁠⁠ to learn more.

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

In this episode, I crashed Pattern's office to sit down with Jacob Miller (VP of Platform Intelligence) and Jeremy Mumford (Lead AI Engineer) from Pattern to discuss their newly co-authored book, "Architected Intelligence." We dive deep into the reality of building scalable AI systems, explaining why companies cannot simply put an LLM in front of everything. We also explore the importance of curating data to avoid the "semantic swamp," the reality of adapting organizational structures to external marketplace algorithms, and the collaborative process of writing a tech book using AI as an "anti-sycophantic" sparring partner. We also discuss the booming tech and startup scene in Utah, the future of autonomous cloud agents powered by budget-friendly AI models, and more.Buy Architected Intelligence: https://amzn.to/4haOqAr

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

I'm getting a ton of emails from individuals and PR agencies asking to be on my podcast. Almost all of it is AI slop spam of the most bland and uninteresting kind. "I have an AI startup solving X." Yeah, you and everybody else... my Goldendoodle also has an AI startup.

As a reminder my podcast is invite-only.

If you or your PR agency sends me these types of messages, I will ignore them or berate you for spamming me with this nonsensical and offensive slop.

Marshall McLuhan predicted a post-literate world in The Gutenberg Galaxy. That world may now be arriving: humans are reading less, AI is reading more, and engineers can increasingly produce code, schemas, documentation, and infrastructure without fully understanding what they created. Perhaps this describes you? ;)

In this Freestyle Friday, I look at the gap between production and comprehension, why judgment is becoming more valuable than cranking out code and diagrams, and how software and data engineers may evolve from writing code, shipping widgets, and building pipelines to creating the context and semantic infrastructure that machines need to act.

-----------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit ⁠⁠fivetran.com⁠⁠ to learn more.

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

I sat down with Sharon Goldman - AI journalist, formerly at VentureBeat and Fortune, now running her own platform, Ground Level AI on Substack - for a conversation that went a lot further than I expected.We cover a lot of ground about the AI backlash, data centers, AI hype vs. reality, and the future of writing.

Check out Ground Level AI: https://www.groundlevel-ai.com/

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

It's been four years and a day since the Fundamentals of Data Engineering was published. The big mental models of the book are the Data Engineering Lifecycle and its Undercurrents. I also discuss what would change about the book if I write it again today.

Thanks to everybody who has supported the Fundamentals of Data Engineering over the years. It's super cool to see the massive impact that the book has had on the data industry and its practitioners around the world.

Are AI agents replacing traditional data engineering tasks?

I sat down with Matt Glickman, CEO & Co-founder of Genesis Computing and former Snowflake and Goldman Sachs executive, for a conversation about where data engineering is actually heading. Most organizations today are stuck in "demo mode" or using LLMs as expensive code completers, simply writing code faster to feed legacy data architectures. But if you look at the industry from first principles, humans were never meant to be data machinery.

In this episode, Matt and I break down how true agentic data systems work, why current enterprise architectures will fail under agent-scale query volumes, and what it really takes to capture institutional knowledge before it leaves the building.

We also cover Matt’s front-row seat at Goldman Sachs during the 2008 financial crisis, his early days turning Snowflake into an enterprise giant, and how AI-native companies will outpace traditional enterprises. If you're building, architecting, or leading teams in the AI and data space, this is a conversation you don't want to miss.

The Joe Reis Show is invitation-only. Unsolicited guest pitches and PR outreach are not considered.

Here's a new show where I quickly cover a mental model that applies to your work as a data leader or practitioner.

Kicking things off is Conway's Law.

It's been a crazy summer for AI. AI enthusiasm remains high, but the market is starting to ask far harder questions. Semiconductor stocks are in a bear market. IBM’s shares fell sharply. Data-center projects are facing political and community resistance. Enterprises are adopting AI, but many are struggling to turn experimentation into measurable returns.

In this Freestyle Friday episode, I look at why these signals do not necessarily mean an AI winter. Instead, they suggest that AI is moving from a "possibility market" into an "accountability market", where capability, economics, infrastructure, and organizational reality finally have to reconcile. Put another way, things are going to get real.

-----------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit ⁠fivetran.com⁠ to learn more.

-----------------------

Sponsor: Revefi

Save serious money on your cloud costs with Revefi’s new autonomous AI DBA, a tool built to handle the gritty reality of cloud data management so you can stop babysitting your infrastructure. Start saving money today at ⁠⁠revefi.com/ai-dba

A short rant based on my new article this week, "The Database Is Not the Data Model"

"In discussions with data practitioners, I keep seeing the same confusion. Someone pulls up a DDL file, a folder of dbt or SQL, or an ERD reverse-engineered from a Postgres instance and says: “Here’s our data model.”

Not to be pedantic, but that’s a schema. Schemas are great. But data modeling is more than just schema design."

Let’s dive into the difference in this podcast and the article.

Article: https://practicaldatamodeling.substack.com/p/the-database-is-not-the-data-model

In this episode, I chat with Steve Brown, an AI futurist, independent consultant, and former professional at Intel and Google DeepMind. Steve shares his insights on helping global leadership teams decode the future of artificial intelligence.

We dive into the common pitfalls of corporate AI adoption, including why so many AI initiatives fail due to poor communication and a lack of cultural integration.

Steve outlines a definitive three-step path to AI transformation, emphasizing the critical difference between 20th-century cost-cutting and 21st-century labor amplification. The conversation also explores how AI is reshaping middle management, the distinct differences between "AI-first" and "AI-native" businesses, and the staggering acceleration of physical AI and humanoid robotics.

In this Freestyle Friday episode, I break down results from the June 2026 Pulse Survey on organizational dysfunction among data engineers. I also dig into why data product ownership (or lack thereof) is one of the fundamental issues standing between companies and success with data and AI.

-----------------------

Sponsor: Fivetran

With the rise of AI and agents, having centralized, trustworthy data is the absolute foundation for building tools and training models. Fivetran automates your data pipelines, removing the need to build fragile connectors so your data arrives clean and reliable. By handling the messy infrastructure behind the scenes, Fivetran allows your team to focus on building the future. Visit fivetran.com to learn more.

-----------------------

Sponsor: Revefi

Save serious money on your cloud costs with Revefi’s new autonomous AI DBA, a tool built to handle the gritty reality of cloud data management so you can stop babysitting your infrastructure. Start saving money today at ⁠revefi.com/ai-dba

I walk through exactly how I use AI for idea generation, review, editing, and drafting on my upcoming book, Mixed Model Arts. This is also the first in a new Wednesday format where I answer listener questions about what I'm building.

--------------------

Mixed Model Arts will be out soon. If you want to get a good deal on the e-book and on-demandcourse, please go to Practical Data Modeling.: https://practicaldatamodeling.substack.com/

Kirill Bobrov, a senior data engineer at Spotify and author of the blog Luminousmen, joins me to talk about his viral "Drunk Post" repost, whether the AI boom is real infrastructure or another bubble, why engineering judgment can't be automated away, and the process of writing a technical book on concurrency.

Luminousmen: https://luminousmen.com/

With AI reshaping employer expectations, junior candidates in data are facing real anxiety about the job market. I share the technical and personal skills I'd focus on, my top five traits I look for in a candidate, and how to build a network from scratch.

-----------------------

Sponsor: Revefi

Save serious money on your cloud costs with Revefi’s new autonomous AI DBA, a tool built to handle the gritty reality of cloud data management so you can stop babysitting your infrastructure.

With a five-minute, zero-touch setup, it deploys 18 specialized agents across your data estate to automatically manage FinOps, performance tuning, and data quality. If you want to cut your cloud costs by 30% to 70% and get back to actual data architecture, check out what they are building at revefi.com/ai-dba

Maxime Beauchemin, creator of Apache Airflow and Superset, joins me to talk about his path from data orchestration to the frontier of AI. We cover his "Claude Code moment," his new agentic workflow tool Agor, and what a software-utopia future might actually look like.

Standing on the high plains near South Pass, Wyoming, where the Oregon and Mormon trails once crossed, I look at the remnants of the 1800s gold rush and draw a direct line to the current AI boom.

Prukalpa Sankar, founder of Atlan, joins me to unpack the piece most AI projects are missing: context. We get into building an enterprise "second brain," why agents get abandoned in testing hell, and what it actually took to rebuild Atlan as an AI-native company.

Juan Sequeda joins me straight off a month of travel to unpack the real state of AI agents in the enterprise, the danger of a "semantic swamp," and why pragmatism beats pedantry every time in data architecture..

🎙️ SPONSORS

Revify - surprise Snowflake bills? One customer cut theirs 50% in 48 hours.→ https://revify.com/demo

Ryan Dolley and I record from the historic Guardian Building in Detroit to talk about building a data career away from the coastal AI bubble, and why Detroit's comeback energy makes it a city worth watching.

🎙️ SPONSORS

Revify - surprise Snowflake bills? One customer cut theirs 50% in 48 hours.→ https://revify.com/demo

Back from months on the road across Asia, Europe, and the US, I unpack the mood I kept running into: uncertainty. Energy shocks, supply chain chaos, and AI upending the playbook for vendors and practitioners alike. Plus updates on my book and an upcoming Salt Lake City conference.

------------------

This episode is sponsored by Revefi,who gives youfull cost and performance visibility for Snowflake by warehouse, user, and workload. One team cut Snowflake costs ~50% across 711 warehouses in under 48 hours.

Book a demo at revefi.com/demo.

------------

Timestamps

0:00 — Intro & travel recap — Sets the stage: months of globe-trotting across Asia, Europe, and the US

1:10 — Global uncertainty & resource scarcity — Fuel/water shortages in Southeast Asia, flight cancellations in Europe, ripple effects of geopolitical tensions

5:30 — AI dominates every conversation — The #1 topic at conferences worldwide; vendors facing existential questions and forced to rethink everything (Atlan pivot, DuckDB agent idea)

10:14 — AI's impact on workers at every level — Senior practitioners gaining superpowers, juniors worried about jobs, leaders expected to do more with less

17:51 — Key takeaway: everyone feels behind — Even top AI insiders are uncertain; give yourself grace, upskill, and consider building something for yourself

20:38 — Announcements — Book drops July 27th, course coming, Practical Data Community Newsletter live, fall travel schedule (London, Paris, possible Salt Lake City conference)

Chris Tabb, founder of LEIT Data, joins me live at Confluent Current London to talk honestly about how AI agents are reshaping the consultancy business model, from billing structure to internal knowledge management.

Eric Weber makes the case that the feeling of falling behind in AI is mostly manufactured. We're miscalibrating what's true for the top 0.1% in San Francisco against everyone else. We get into judgment as the real bottleneck, and why so many people are quietly leaving corporate roles.

Tristan Handy and I unpack a shift coming to data infrastructure: within 12 months, the primary consumers of data won't be humans, they'll be AI agents. We dig into compute costs, "context stores," and why the industry shouldn't just build horses with wheels.

NOTE - Sorry for the edits in this video. I used Descript to edit out the umms and uhhs, and it was a bit too aggressive. Will make it less jarring in future videos. Thanks.

🎙️ SPONSORS

Fivetran - stop cobbling pipelines together. Set it, forget it, scale as you grow.→ https://fivetran.com

Revify - surprise Snowflake bills? One customer cut theirs 50% in 48 hours.→ https://revify.com/demo

Sanjay Agrawal, CEO of Revefi and co-founder of ThoughtSpot, joins me to talk about how AI agents are quietly driving unpredictable spikes in cloud data spend, and why most FinOps practices are built for a world that no longer exists.

Zach Wilson and I happen to be in Stockholm on the same night, so we sat down to talk about what it actually takes to be a data engineer in 2026.

Josh Wills has spent 25 years writing data pipelines. He uses coding agents daily and keeps hitting the same wall: they jump to conclusions and ship pipelines nobody understands. We get into the $200K vibe-coded pipeline problem, and why politics is the last job LLMs will ever touch.

Everyone says if you're not maxing out AI every second, you're getting left behind. I disagree. I talk about minimizing AI use for deep cognitive work, my new minimalist travel setup, and the skills that will actually get you left behind.

Shinji Kim, founder of SelectStar (acquired by Snowflake), joins me to talk about the deal, and why in a few years we may stop calling these things "data catalogs" at all as the category evolves into a living AI context layer.

Recovering from jet lag and hiking the hills of Salt Lake City, I break down a conversation with a VC friend about what's actually defensible in the age of coding agents, and what isn't.

George Fraser, CEO of Fivetran, breaks down why the "modern data stack" has evolved into "open data infrastructure," and why data gravity is the most overrated concept in data management.

From Shibuya Crossing in Tokyo, I make the case that the 40-year adoption curve of the electric grid is a better analogy for the AI boom than the dot-com bubble, and why most AI failures come down to the organization, not the technology.

Bob Seiner has worked in data governance since before it had a name. We dig into his Non-Invasive Data Governance approach, why he pushes back on the term "data enablement," and the state of data security in the age of autonomous AI agents.

From the hillsides of Phuket, Thailand, I push back on the idea that teams "don't have time" for fundamentals anymore, and make the case that AI actually makes data engineering more critical, not less.

Wes McKinney returns to talk about his complete reversal from AI skepticism to being fully locked in on coding agents, why he thinks Go has become the best language for AI agents, and how he's building his own local data sovereignty tools.

In this Freestyle Friday episode, I catch up with Eric Weber after our recent walk through downtown San Francisco. We dive deep into the very real fear and identity crises sweeping through the tech industry as AI accelerates. We discuss how packing a year of change into a single week is disorienting workers and how the constant hustle culture in SF might finally be hitting its threshold.

We also get into the darker side of this shift, including the "reverse centaur" effect where humans are reduced to parts of a machine. Are white-collar engineers about to face the Amazon warehouse treatment through token consumption leaderboards? Eric also shares why he took a step back from leadership, his focus on writing, and the importance of genuine human connection right now.

Eric Weber: https://www.linkedin.com/in/ericweberdata/

I recently sat down with Amit Prakash, the brilliant mind who co-founded ThoughtSpot and led AI teams at Google and Microsoft, to talk about a massive shift happening in the data world.For decades, we’ve been forcing the "messy reality" of business into rigid database tables, losing about 90% of the actual information in the process.Amit is now building Ampup to flip that script. We dive deep into how he’s using dynamic ontologies to extract high-fidelity insights from unstructured data, like the nuances of a 60-minute sales call—to drive massive ROI.In this episode, we explore:- The "SaaSpocalypse" and the Future of Agents: Why the early stages of the sales cycle might soon be a "dance" between AI buying and selling agents.- Sales as Athletics: Why high-stakes negotiation is more like football than a desk job.- The "Business Brain": Moving beyond simple CRMs to a central strategy engine that understands every department’s unstructured data.- Human-to-Human Trust: Why large contracts will always require a human touch, even in an AI-saturated world.- Amit’s perspective on how AI can deliver real value to the GDP by fixing the "distribution bottleneck" of innovation is a must-listen for anyone in tech, data, or leadership.

In this episode, I sit down with Chris Gambill, a data strategy and engineering leader, fractional consultant, and career coach. We dive into the realities of the data engineering job market in 2026, exploring what it takes to stand out, the massive shift AI coding tools are causing, and why mastering the fundamentals of data engineering remains crucial.

Chris shares his unfiltered thoughts on coaching career switchers into data engineering , why finance professionals make great data engineers , and the exact resume and portfolio strategies hiring managers are actually looking for. We also get into the weeds on the latest AI development tools, comparing GitHub Copilot, Claude, and Codex.

If you're looking for solid, no BS advice on the field of data engineering in 2026, this is a great discussion!

Gambill Data Engineering: https://www.gambilldataengineering.com/

LinkedIn: https://www.linkedin.com/in/databasemanagement/

In this episode, I sit down with Gowtham Chilakapati, an analytics veteran of 18 years and Executive Director at Humana , to pull back the curtain on the reality of Agentic AI in the enterprise.We dive deep into the recent wave of tech layoffs—like the news of Block cutting 40% of its workforce —and debate whether AI is truly driving these decisions or simply serving as a convenient excuse for broader management failures.Gowtham shares his firsthand experience navigating an astounding $1 billion AI investment during the early adopter rush of 2024. He details the chaotic first six months of that initiative and the multi-dimensional framework his team developed to measure true return on investment beyond the traditional, and often flawed, software implementation mindset. From the massive risks of pasting PII into LLMs to how AI prototyping is finally bridging the historic gap between product and engineering teams, this conversation is a masterclass in pragmatism for anyone looking to cut through the AI hype, especially in highly regulated industries.

The new Practical Data Community Pulse Survey for March 2026 just came out, and I unveiled some of the findings at yesterday's Undercurrent event in San Francisco.

The short version is: AI is here to stay. Everyone's using it, but the hard parts we've always dealt with as an industry still remain unresolved. Listen and find out why.

In this episode, I sit down with Jake Ward, founder of the Application Developers Alliance. We dig into the AI "Frankenact," aka the EU AI Act, and why policymakers regulating tech they fundamentally misunderstand creates a cold wind for software innovation.Jake drops some harsh truths about why giving developers a voice in Washington is harder than it looks, why collective bargaining and developer unions probably won't work, and how bad policy is forcing companies to build for compliance rather than ship great products.

The data job market is evolving, but it's still there. In this episode, I give my thoughts on the data job market, ways to navigate it, going solo and having a Plan B, and more.

In this episode, I sit down with Demetrios Brinkmann (godfather of the MLOps Community) to talk about the absolute Wild West of AI right now.

We cover how fast coding agents are changing the game, the reality of "vibe coding" your own CRM , and how Demetrios's community saved $20,000 just by ditching bloated enterprise tools.

But we don't just talk tech. We get into the weeds on the content creation pipeline, from the bizarre rise of AI OnlyFans to the "Doorman Paradox" of automated content. Finally, we spill some serious inside baseball on the tech sponsorship game, calling out the sheer audacity of heavily-funded startups expecting free labor from communities , and why protecting your reputation is worth more than any quick paycheck.

In this episode, Matt Housley and I reunite for a Friday catch-up, bringing back some of that classic Monday Morning Data Chat energy. We dive into the absurdity of the "buzzword industrial complex," and why declaring it the "Year of Context" is mostly just industry hype, per usual.

We also tackle the chaotic reality of deploying AI agents (including the ultimate YOLO, OpenClaw) without proper data governance, the Anthropic class action lawsuit regarding copyright, and why regional conferences like DataTune are awesome. Finally, we discuss the shifting landscape of media, the death of traditional book publishing models, and the rise of the independent, niche creator.

The white-collar tech industry isn't what it used to be, and anyone could be on the chopping block at a moment's notice. With tens of thousands of highly skilled people getting laid off from Big Tech on a seemingly bi-weekly basis, competing in the traditional job market is brutal right now.In this episode, Jody Hesch and I discuss why building a freelance data consulting business isn't just a career pivot—it is a necessary Plan B. We break down the exhaustion of constantly reinventing yourself and navigating new team dynamics every time you switch full-time roles. We also explore the counterintuitive reality that by going freelance, you only have to build your network and reputation once to create a repeatable motion. Whether you are actively looking for an exit or just realizing that the gig economy is coming for data engineering, this conversation covers the realities of making the jump.

In this conversation, Paul Blankley and Ryan Janssen, founders of Zenlytic, drop in to discuss the massive shift in how we build software and handle data. We trace their journey from studying early NLP and Transformers at Harvard right when the BERT paper dropped, to building a company that relies on cutting-edge LLMs. As far as I know, they're the first to use LLM's for analytics.

We dive deep into the reality of the agentic era: engineers are no longer writing the bulk of the code; they are managing agents, verifying outputs, and maintaining ridiculously high standards. We also explore why the industry needs to embrace "net negative scaffolding" as models get smarter, and why having good "taste" might be the ultimate human moat left in tech.Bonus:

To prove that software development is changing faster than ever, we literally "vibe coded" a brand-new CRM called "Slop Force" in 20 minutes during this episode. Zenlytic: https://www.zenlytic.com/

We often hear about the AI skills gap, where people need to get training on the latest AI tools. There's also the AI competence gap, where people might not have the skills or competence in a field, and use AI to mask over those shortcomings. The results are what you expect - chaos. In this episode, I unpack these two gaps, and do my usual ranting about learning the fundamentals and investing in oneself.

----------

🚨

Also, if you happen to be in San Francisco on March 26th, please join me at Undercurrent, a small and tech-focused conference for data engineers and architects. No sponsors, no salespeople, no bullshit. Just great technical discussions all day.

Register here: ⁠https://cnfl.io/3ZCTaVx⁠

In this conversation, I sit down with Tim Delisle and Chris Crane, co-founders of 514, to discuss bridging the gap between software development and data engineering. We cover their experience leading global data engineering at Nike and why software teams are increasingly taking ownership of heavy analytical workloads.We also dive into how they are building the Moose Stack to give developers a local-first, code-first analytics experience. Finally, we explore how AI co-pilots are acting like an "army of interns" to fundamentally change how we write code , and why the "personal data lake" might be the future of privacy and local compute.Check out 514 & The Moose Stack: https://www.fiveonefour.com/

Sadie St. Lawrence joins me to unpack her concept of the "AI Orchestrator," explaining how it shifts our mindset from being a musician to a conductor in the age of AI. She shares insights from her work at the Human-Machine Collaboration Institute (HMCI), detailing how her team is building AI-powered solutions and tackling complex problems. We also chat about the common pitfalls in AI adoption, from unfounded fears to "work slop," and why foundational systems thinking remains paramount.

This week, I published an article called "2028, the Great Data Reckoning," which got a ton of response. Although I originally meant it to be satire, when I re-read it I felt like it was actually a glimpse into what's happening in our field right now. In this episode, I chat about the implications of the Great Data Reckoning on practitioners, leaders, and founders.

Article: https://joereis.substack.com/p/2028-the-great-data-reckoning

----------

🚨

Also, if you happen to be in San Francisco on March 26th, please join me at Undercurrent, a small and tech-focused conference for data engineers and architects. No sponsors, no salespeople, no bullshit. Just great technical discussions all day.

Register here: https://cnfl.io/3ZCTaVx

In this episode, I sit down with Prashant Sridharan, a 30-year veteran of developer marketing who has shaped go-to-market strategies for tech giants like Sun Microsystems, Microsoft, AWS, Facebook, and Twitter, and currently runs product marketing at Supabase. We dive deep into the origins of DevRel and how marketing to developers has evolved in an increasingly noisy, AI-saturated landscape.Topics covered:- Transitioning from massive tech companies to the fast-paced startup world - How to genuinely measure the success of Developer Relations without ruining communities - Using AI tools like Claude to accelerate mechanical marketing tasks while preserving authentic storytelling - The shift from traditional SEO to GEO (Generative Engine Optimization) for developer tools - The thrill of live, unscripted coding demos and stories from sharing the stage with Steve Ballmer - Prashant's upcoming fiction novel, The Midnight Coders Children, and the craft of writing Find more from Prashant at StrategicNerds.com and check out his non-fiction book, Picks and Shovels: https://amzn.to/4cJ2TRO

For 40+ years, the data industry has tried to teach good practices and get adoption, often in the same way. And for 40+ years, that approach keeps failing over and over. Based on the recent Practical Data Community Survey, practitioners face challenges like time pressures, lack of direction, and lack of clear ownership. Do we need to try something else as an industry? Or do we continue to be the poster child for the definition of insanity - doing the same thing over and over, yet expecting different results? I hope not.

Why are we still using row-based protocols like ODBC and JDBC in a column-oriented world? In this episode, I sit down with Ian Cook, co-founder of Columnar and a long-time Apache Arrow contributor, to discuss the critical infrastructure changes needed to speed up modern analytics and AI.

We dive deep into the technical bottlenecks of legacy standards - specifically the "serialization tax" of converting columns to rows and back again - and how ADBC (Arrow Database Connectivity) solves this by keeping data columnar from end-to-end. Ian also shares his insights on the intersection of tabular data and LLMs, why AI agents need better access to OLAP systems, and the tension between vibe coding speed and the stability required for critical open-source infrastructure.

The 2026 Practical Data Community State of Data Engineering dropped this week. It's full of some obvious and very counterintuitive information about the state of data engineers around the globe, in all sizes and types of organizations. Check it out!

Also, I talk about the book writing process, where I messed up on this latest book, it's progress toward publication, and more.

Survey: https://joereis.github.io/practical_data_data_eng_survey

---------------------

This episode is brought to you by Ellie.ai

Ellie makes data modeling as easy as sketching on a whiteboard—so even business stakeholders can contribute effortlessly. By skipping redraws, rework, and forgotten context, and by keeping all dependencies in sync, teams report saving up to 78% of modeling time.

Check out Ellie: https://ellie.ai/

I sat down with Paul Dudley (CEO) and Ricky Thomas (CTO) from StreamKap to catch up on where the world of streaming data is heading—and things have changed fast since we last spoke.

We dive into the concept of "vibe coding" and how AI is radically accelerating how we build software (I even share a story about building a data analysis tool in an hour). But the real meat of this conversation is about the intersection of streaming data and AI agents. Everyone is building agents, but without real-time context, they’re flying blind. We discuss why streaming is a missing link for agentic workflows, the shift from dashboards to automated decision-making, and why SaaS companies are racing to build walled gardens around their data.

We also get into the nitty-gritty of the UK vs. US tech markets, the resurgence of PR in the AI era, and StreamKap’s upcoming move into the Snowflake native app ecosystem.

Streamkap: https://streamkap.com/

This week was a doozy with new AI releases, the stock market, and more. It really feels like this was the first tremor in AI's impact on the SaaS market. What's do I think is next? Listen and find out.

In this episode, I sit down with Mike Driscoll, founder of Rill Data, to discuss the evolving landscape of business intelligence and data engineering. We explore why the industry keeps "rediscovering" old concepts like the semantic layer and how the rise of AI agents is forcing us to rethink how we structure data.

Mike shares his insights on the "shape" of analytics, debating whether conversational interfaces will replace dashboards or simply complement them. We also dig into the growing demand for data engineering, the importance of watermarks and temporal semantics, and why data visualization remains a critical tool for "trust but verify" in an AI world.

Rill Data

Mike’s Podcast: Data Talks on the Rocks

As I use AI, I'm finding that I create MORE work for myself, not less. One task completed means five more to do. This is the paradox of today - AI might actually mean more work, not less.

I talk about this, the Data Day Texas final episode, and more.

Check out the review I did of Cube's new analytics agent: https://www.youtube.com/watch?v=p3frGJOUl1E

(Thanks to Cube for partnering on the review)

Lak Lakshmanan had a successful career in Private Equity and Big Tech, but he realized he couldn't just "coach the game" while the rules were changing. He had to get back on the field play it. We discuss vertical AI, the "foolhardiness" required to start a company , the reality of the AI technology wave, and why sitting on the sidelines is the biggest risk of all.

LinkedIn: https://www.linkedin.com/in/valliappalakshmanan

Generative AI Design Patterns (book): https://amzn.to/45v0xBO

In this episode, I talk about how I'm kind of living in a bubble of cool tech and AI, and how the 99% of businesses out there are still grappling with the same old data and tech problems they've always dealt with.

I also talk about how me and my friends are using AI to automate the boring stuff and scratch our own itches.

In this episode, I sit down with science fiction author, activist, and journalist Cory Doctorow to unpack his viral concept of Enshitification, the three-act tragedy of platform decay:

1. be good to users

2. lock them in

3. extract value from users to feed advertisers and shareholders

We also dive into:

- The AI bubble: Cory’s case that parts of the sector are propped up by aggressive accounting and incentives, not durable value.

- The “Reverse Centaur”: How workers (from Amazon drivers to radiologists) are being reorganized to serve machine workflows, rather than machines serving humans.

- Software engineering vs. “vibe coding”: Why autocomplete isn’t engineering, and why AI can’t replace process knowledge and domain context.

- The Post-American Internet: What happens when the U.S. weaponizes platforms, and the rest of the world builds alternatives.

About Cory Doctorow: Cory is a multi-time international bestselling author, special advisor to the Electronic Frontier Foundation, and creator of the blog/newsletter Pluralistic.

If you got value from this conversation, hit Follow and share it with one person who cares about the future of tech.

Tech is full of smart people with smart ideas - enterprise data models, ontologies, data mesh, proprietary AI strategies - that repeatedly fail to gain traction. When they fail, the blame usually goes to "stupid users", "lazy and immature organizations." Perhaps, but I don't think that's the whole story, and if you adopt that mindset, you're sure to keep failing.

I think there's more to the story. Listen and find out...

In this episode, I visited the Hex office and sat down with Barry McCardle (CEO of Hex) to talk about the massive shift we’re seeing in the data stack. Countless companies have spent decades buying BI tools in the hope of "self-serve Nirvana," yet most dashboards still raise more questions than they answer.

Barry and I dive into why the traditional dashboard is becoming a "jumping-off point" rather than a destination, and how AI agents are finally closing the gap between having a question and getting a sophisticated answer.

We also discuss building tools people love, "commitment engineering", Barry's story, and much more.

What status game are you playing? Are you trying to outcompete others, or playing your own game? In this episode, I talk about status games in data and careers in general.

The technology industry is prone to moving fast and forgetting its history. This is a shame because our industry is built on the shoulders of many giants, often long forgotten.

Bill Inmon, Roger Whatley, and I discuss the history of technology and computing, covered in their new book, From Stone to Silicon. We talk about the big people and moments in technology and computing, and much more.

From Stone to Silicon (book): https://amzn.to/4pLfqat

Welcome to 2026! In this spontaneous Friday AMA, I take listener questions on ontologies, the “leaky abstractions” of AI coding tools, why the “button pusher” era of engineering is a professional dead end, and the shifting landscape of data engineering.

I also provides an update on my upcoming book, Mixed Model Arts (launching in March 2026), and discuss the unexpected convergence of library science, ontologies, and traditional data modeling, something not on my 2025 bingo card.

Great turnout, especially for no notice. Thanks to everyone who showed up!

Happy 2026! In this episode, I rant about whether vibe coding and AI coding agents makes the Law of Leaky Abstractions obsolete, making your first dollar (or whatever currency), and more.

The Law of Leaky Abstractions: https://www.joelonsoftware.com/2002/11/11/the-law-of-leaky-abstractions/

If you like this podcast, please take 10 seconds and give it a rating or review on your podcast platform of choice. It will go a long way to giving the show more visibility. Thanks!

2025 is nearly gone, and in this episode, I give some thoughts on what I think might happen in 2026. I also chat about this week's surge of interest in Kimball vs Inmon (and the podcast I tried to organize with them) and much more.

“What I built today might be obsolete tomorrow.”

This is something I heard this week from a developer, and this is not uncommon given the warp speed nonstop advancement of AI models every week. We used to measure the rate of change in months or years. Now it’s days or weeks.

In this episode, I talk about why writing code is rarely hard part, and why having good taste and shipping things that people love is the most important things we can do.

In this episode, Nik Suresh returns to the show to discuss his first year running a bootstrapped services company. And no, he probably will not throat punch or pile drive you.

Nik explains why he moved away from hourly billing to fixed pricing, why writing code is often the least profitable part of a project, and how to spot "status games" in the tech industry. We also dive into the current state of AI, why bad leadership is the real problem behind failed tech initiatives, and trade stories about MMA and boxing.

We debunk the myth that starting a business has to be miserable, explore the performative nature of "hustle culture" in Silicon Valley, and break down why engineers often struggle with consulting sales.

Oh yeah...ontologies. In this mini-clip from Matt Housley and I, we chat about why ontologies are super popular now.

Had an interesting discussion with my 15 year old son. He and his friends see white collar work as “cooked.” They see it as a rat race where the work is increasingly insecure, abusive, and meaningless. Then there’s the looming question of AI…

Instead, they’re interested in careers they find meaningful and not as exposed to whatever AI does to work. And if they own a company, they’ll just hire “clankers” whenever that moment arrives.

I’m excited that these kids are looking at what’s happening right now, questioning if it’s their path, and choosing a life that’s fit for them.

More broadly, especially in the age of AI, I think some of the most important conversations we need to have is over what we find valuable and meaningful, making a living and the nature of work, and the nature of community.

It's Friday! Matt Housley and I catch up to discuss the aftermath of AWS re:Invent and why the industry’s obsession with AI Agents might be premature.

We also dive deep into the hardware wars between Google and NVIDIA , the "brain-damaged" nature of current LLMs , and the growing "enshittification" of the internet and platforms like LinkedIn. Plus, I reveals some details about my upcoming "Mixed Model Arts" project.

In this episode, I sit down with Mark Freeman and Chad Sanderson (Gable.ai) to discuss the release of their new O’Reilly book, Data Contracts: Developing Production-Grade Pipelines at Scale. They dive deep into the chaotic journey of writing a 350-page book while simultaneously building a venture-backed startup.

The conversation takes a sharp turn into the evolution of Data Contracts. While the concept started with data engineers, Mark and Chad explain why they pivoted their focus to software engineers. They argue that software engineers are facing a "Data Lake Moment, "prioritizing speed over craftsmanship, resulting in massive technical debt and integration failures.

Gable: https://www.gable.ai/

I meet a lot of people who want to accomplish major goals next year. Then the year comes and goes and most people are still waiting to get started.

It's almost December. Rather than wait until the New Year to get going, use December to plan how you'll execute on "that thing" you're itching to accomplish. Time waits for nobody, so get going.