Asked how his view of where a query should run has changed, Guillermo Sánchez Dionis describes DuckDB as the point where laptops became serious analytics machines.
…systems. So if you want to query an iceberg table and a Redshift server or a Postgres remote server, that you can use the same system and that this system hopefully does things in a very, very, very optimized way so that you can actually get pretty good results from doing this type of federation.
So we've already talked a lot about how developers would love to work more from home. Don't. That's just the truth of the matter. So how else has your view of where a query should run changed over the last year? And I guess who or what can run that query?
Yeah, for sure. I mean, when I started working with data, your laptop couldn't do anything. Not because it didn't have some capabilities, but there was no system that could run in your laptop that actually used your laptop in a proper way. And I think when I discovered DougDB, it completely changed the game for me. I was just like, okay, I don't need to spin up a Spark cluster for breakfast and wait for it to run, which is another different thing, of course. You can just literally type DuckDB in your laptop, and then suddenly you have a system that is capable of doing crazy things within the comfort of your laptop. That I find very special. And I think it's completely changed the way I see analytics. And I think it also means that people that are not really interested in running a production setup or whatever, but they still want to do data analysis, they can do that. They can do that as long as they have a laptop that is half competent, they can do that with DACDB.
Wow. Okay. Half competent. Those will vary, won't they? So let's put on your prediction fortune-telling hat. So by November 2027, a year from now, what will be true about how developers connect to analytical databases that isn't necessarily true today?
I think hopefully the experience will unify. So what I mentioned before about one system being able to run queries across different stacks, that will probably be the case. And I know for a big part of it, DACDB and DAC SQL will be the interface, or that's what I'm hoping for at least.
Okay. Because we need things to be simpler as humans and agents interact with that database.