AI, Data Engineering, and the Modern Data Stack

**Tristan Handy** (0:00)
If you like literally take all the tasks that Data Engineer does every day, and you write them on a list, and you say, which one of these are things that a highly trained, highly paid human being should be spending their time on? It's like, a lot of them, they shouldn't. Pipeline failures happen, and yet, inevitably the cause of those failures is kind of dumb. It's like not that interesting. Agents are like quite good at identifying the problem and proposing a fix. I expect to see a lot of automation of data engineering tasks over the coming 12 months. Jevons Paradox is coming into effect pretty hard right now. Analytics always expands to fill the available budget. You want to continue to improve the price to performance ratio, not so that at the end of the day, like people can stop doing things, but so that they can do more things.

**Matt Bornstein** (0:54)
Thanks for listening to the a16z AI Podcast. We have another great discussion for you today, this time featuring dbt Labs co-founder and CEO Tristan Handy, along with a16z general partner Jennifer Li and partner Matt Bornstein. If you're active in the world of data engineering, there's a good chance you're familiar with dbt. But if you're not, here's the very short version. dbt helps its users build data products using the rigor and best practices of software engineering. And as Tristan points out during the episode, it counts more than 1 million users across more than 70,000 organizations.
However, this discussion isn't really about dbt. It's about the major changes in the data world brought about several years ago by the concept of a modern data stack and, more recently, by the advent of generative AI. The three start off on the topic of where AI can really shine in the world of data analytics and data engineering before getting into the rise and plateau of the modern data stack. They also cover the lessons data engineers can still learn from software engineers and, and finally, what we should make of a spate of acquisitions and product announcements across the data infrastructure market. And you'll hear it all after these disclosures. As a reminder, please note that the content here is for informational purposes only, should not be taken as legal, business, tax or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any a16z fund. For more details, please see hay16z.com/disclosures.

**Tristan Handy** (2:23)
I don't believe in the idea that you're going to do analytics by asking a model to write SQL. It's not that interesting if you can write a well-formed SQL query.
The hard part of analytics is what data analysts are doing is they are socially constructing truth inside of an organization. There is no such thing as revenue in an abstract sense. It is just, what do we all agree is the way that we measure revenue. And a model just doesn't have access to that unless you give it very specific instruction and you would do that through metadata and in a best case scenario, you would do it through something called a semantic layer. A semantic layer would actually give the model exactly the metadata required to construct the SQL query in a way that everybody in the organization agrees it should be constructed. We acquired a company called Transform, I think it was two and a half years ago, and now it's integrated into the dbt platform and we built an MCP server that kind of exposes this functionality. And when you go to any MCP enabled language model and you ask it questions about your business data, it gives you correct answers. And the funny thing is that like, there's a bunch of people that kind of play around with that, but it hasn't crossed into the mainstream. Like, there's a ton of curiosity around this and... But still, people are using the BI tools generally that they have been using.

**Jennifer Li** (3:54)
Let's break it down to like, what are the tasks an analyst is doing today? And what are the pieces that actually these models have capabilities to serve? I think, you know, even compared to a year ago, the capability of writing SQL is nice and day. And I recently asked ChaiGBT to build a chart for me with this very complex Excel sheet, actually. You know, you need to sort of do a couple of paper tables and like paint this chart, like take out a couple of rows and columns as well. It did a great job at painting this chart. I was very surprised and impressed. I did a couple of spot checks of like, if this is still correct data points and they're correct. And that sort of gave me more, you know, hope of where AI can be applied to maybe in the final step of visualization. But there's also data cleaning work to do for analysts. There is this organizational social work to do, which I don't believe ever, maybe through a few agents working together, they can get gather some truth. But which are the pieces maybe we can speculate now, ready to be automated and which are the pieces still, I think we require a human to come in and do the work.