Catalog & Cocktails: The Honest, No-BS Data Podcast
Oct 1, 2026 · 6 min · 3 segments
This is the takeaway episode with Tristan Handy, President and Co-founder of Fivetran / dbt Labs where he joins Juan and Tim to discuss how agents will build data infrastructure at every layer and…
Juan SequedaHost
Tristan HandyGuestWe started with like honest no BS, so agents are taking over the, the world of data.
What, what, what comes next? Um, and I think you rightly said like well, you know, that in...
Like the, the taking over of the data is the what's next, and let's like kind of let's, let's, let's double click into that.
And you had two major kind of insights or, or predictions kind of depending on how you look at them.
Like one of them is that agents are going to be responsible for the data infrastructure layer or, or layers.
Um, uh, and then two, uh, agents are going to be the primary consumers of said data layers, data stack.
And uh, one is more about kind of how the infrastructure evolves and its capabilities and, and, uh, and, and how AI helps to maintain it, the role of people, how they change.
And the second one being more about what does it mean to live in a world where you have infinite and unlimited data analysts that, that, that work for you, right? Uh, and so double clicking a little bit on that first part, data infrastructure.
We've got a lot of these paradigms now for how to represent, you know, data pipelines and orchestration jobs and semantic layers and things like that.
The code is being now heavily developed, uh, primarily developed by AI, right? And so data engineers, uh, leaning into things like Cloud Code, et cetera, to help them develop a lot of this stuff.
And if you in your role are not being given an opportunity, tying to your advice here, to be able to use AI tools to help you with your job, you need to find a way to get them into what you're doing or maybe find a role where you get to do that, right? Um, the technical skills are becoming less relevant, and because agents are doing more, you're able to zoom out and focus more on the sociotechnical intersections, focus more on the business problems.
Um, and uh, think about things in terms of, uh, a loop, right? And how you're actually building the factory for your organization that is helping to build out these layers to maintain that infrastructure.
Think about am I working in the process or am I working on the process, is, is what you said.
And so at the end of every day or the beginning of your day, think about yesterday or today.
Did I, was I in the process or on the process? And be thinking more about how you can be on the process.
So when you're spending time on the process you get to do things that you didn't do before because there wasn't any, any much, much ROI, right? It's the tool sharpening thing.
Like how much time should I spend on sharpen my tool versus building things? And when you're doing that like you wanna collaborate with people who have the same type of problems, and you do this in open communities which results in open tooling.
And kind of your point there is that there needs to be some shepherding by communities, it needs to be open.
It talked about how you're seeing gravity around the iceberg tier, the ingestion, transformation, semantic layers, metadata compute, visualizations, and then yeah, agents can go stitch this all together.
We started with like honest no BS, so agents are taking over the, the world of data.
What, what, what comes next? Um, and I think you rightly said like well, you know, that in...
Like the, the taking over of the data is the what's next, and let's like kind of let's, let's, let's double click into that.
And you had two major kind of insights or, or predictions kind of depending on how you look at them.
Like one of them is that agents are going to be responsible for the data infrastructure layer or, or layers.
Um, uh, and then two, uh, agents are going to be the primary consumers of said data layers, data stack.
And uh, one is more about kind of how the infrastructure evolves and its capabilities and, and, uh, and, and how AI helps to maintain it, the role of people, how they change.
And the second one being more about what does it mean to live in a world where you have infinite and unlimited data analysts that, that, that work for you, right? Uh, and so double clicking a little bit on that first part, data infrastructure.
We've got a lot of these paradigms now for how to represent, you know, data pipelines and orchestration jobs and semantic layers and things like that.
The code is being now heavily developed, uh, primarily developed by AI, right? And so data engineers, uh, leaning into things like Cloud Code, et cetera, to help them develop a lot of this stuff.
And if you in your role are not being given an opportunity, tying to your advice here, to be able to use AI tools to help you with your job, you need to find a way to get them into what you're doing or maybe find a role where you get to do that, right? Um, the technical skills are becoming less relevant, and because agents are doing more, you're able to zoom out and focus more on the sociotechnical intersections, focus more on the business problems.
Um, and uh, think about things in terms of, uh, a loop, right? And how you're actually building the factory for your organization that is helping to build out these layers to maintain that infrastructure.
Think about am I working in the process or am I working on the process, is, is what you said.
And so at the end of every day or the beginning of your day, think about yesterday or today.
Did I, was I in the process or on the process? And be thinking more about how you can be on the process.
So when you're spending time on the process you get to do things that you didn't do before because there wasn't any, any much, much ROI, right? It's the tool sharpening thing.
Like how much time should I spend on sharpen my tool versus building things? And when you're doing that like you wanna collaborate with people who have the same type of problems, and you do this in open communities which results in open tooling.
And kind of your point there is that there needs to be some shepherding by communities, it needs to be open.
It talked about how you're seeing gravity around the iceberg tier, the ingestion, transformation, semantic layers, metadata compute, visualizations, and then yeah, agents can go stitch this all together.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
