Your cloud SSD is sitting there bored, and it would like a job.

Today we're putting into work with disk cache, a simple practical cache built on SQLite

that can speed things up without spinning up Redis or other extra servers.

Once you start to see what it can do, a universe of possibilities opens up.

We're joined by Vincent Warmerdome to dive into Discash.

This is Talk Python to Me,

episode 534 recorded December 19th, 2025. Talk Python to me. Yeah, we ready to roll. Upgrading the cold. No fear of getting old. They sink in the air. New frameworks in sight, geeky rap on deck.

Quark crew, it's time to unite. We started in pyramid. Cruising old school lanes. Had that stable base. Yes, sir. Welcome to Talk Python, the number one Python podcast for developers and data scientists. This is your host, Michael Kennedy. I'm a PSF fellow who's been coding for over 25 years. Let's connect on social media. You'll find me and

Talk Python on Macedon, Blue Sky, and X. The social links are all in your show notes. You can find

over 10 years of past episodes at TalkPython.fm, and if you want to be part of the show, you can join

our recording live streams. That's right, we live stream the raw, uncut of each episode on YouTube. Just visit talk python.fm slash YouTube to see the schedule of upcoming events. Be sure to subscribe there and press the bell. So you'll get notified anytime we're recording. Vincent, hello. Michael, Michael, we're back. Awesome. Awesome. We'll be back with you. Yeah. This is almost the sequel to the last time you were on the show. So it's going to be fun. Yeah, so SQL in this case, not the query language, like an actual sequel of events. Yes. Yeah, you can correct me if I'm wrong, but I think what happened is you had me on a podcast a while ago,

talk about a course that I made and a big chunk of the course that we were in very enthusiastic about was about this tool called Discash. And then we kind of came to the conclusion, well, we had to cap it off. Maybe it's fun to do an episode on just Discash, since we're both pretty huge fans of it. I think that's how we go here.

I think that is how we got here as well.

And we're going pretty huge fans of it. I think that's how we got here. I think that is how we got here as well. And we're going to dive into this pretty, honestly, it's a pretty simple library called Discash, but what it unlocks is really, really sweet. And then I'm going to talk about a lot of different angles. And now, even though it's just been not that long since you're on the show,

maybe just give us a quick intro of who you are. Hi, my name is Vincent. I've done a bunch of data machine learning stuff mainly in the past. That's sort of what a lot of people know me from. These days, though, I work for a company called Marimo. You might have heard from us. We make very modern Python notebooks. We took some lessons from Jupiter and we take a new spin of it. So that's my day to day. But I still like to write notebooks and do like kind of fun little benchmarks and also stuff with LMs. And I've just noticed that for a lot of that work, boy, disk cache is amazing. And I also use it for web stuff. And I think that's also what your use case is a little bit more of. But yeah, like in notebook land, you also like to have a very good caching mechanism. And on the Remo side of things, we are also working on different caching mechanisms, which I might talk about in a bit. But just for me, the bread and bother, the thing I've used for years at this point is disk cache whenever it comes to like that territory. Yeah, it's funny. This was recommended to me for Python Bites as a news item over there

quite a while ago, like years ago. And I'm like, oh, that's pretty interesting. And then I saw

you using it in the LLM Building Blocks course and it just it just unlocked from me like, oh my,

oh, this this is something else.. And so since I've been doing a

bunch with it, and I'm a big, big fan. I've been on this, like, trying to avoid complexity,

but still getting really cool responses, performance, et cetera, out of your apps. And I think

this, you know, this is a really nice way to add multiprocess super fast caching to your app without involving

more servers and more stuff that's got to get connected and keep running and so on.

...