Real Alpha
Watch mode

The MAD Podcast · 2 min · Oct 8, 2026

LLMs tune databases 85% of the way in 15 minutes

Andy Pavlo

VP of Database Research, ClickHouse

Carnegie Mellon database professor who has led its Database Research Group since 2013 and co-founded OtterTune.

  • Matt TurckHost · General Partner, FirstMark

    Publishes the annual MAD landscape of data and AI; sold TripleHop to Oracle.

About this segment

Pavlo says LLMs changed his decade of self-driving database research because they were trained on tuning guides. They get most of the way in minutes, which helps the many teams running default cloud configs.

From “Andy Pavlo: What Happens When Billions of AI Agents Hit Your Database”, Oct 8, 2026. Watch the full episode

“the LLMs can get you maybe 85% of the way there and tuning certain things.”
Andy Pavloat 34:14 · The MAD Podcast34:14

Takeaways

  1. Earlier ML tuners needed hours of training and scarce production data
  2. LLMs reach about 85% of optimal tuning in roughly 15 minutes
  3. Many users run default cloud configs, so easy wins remain
Watch 2:02 from 33:51

In this episode

Full episode
  1. 10:43Agents could mean 10 to 100x more database queries4:37
  2. 23:38Text-to-SQL reached 99.5% at a bank, but at great cost2:11

  3. 33:51LLMs tune databases 85% of the way in 15 minutes2:02

Up next · Training DataGoodput, not FLOPS, is how Google holds itself accountable5:03

Transcript

Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.

  1. 33:57

    that you can then make good decisions on the production databases. And then the LLMs show up, and that really changed how you can approach the problem because they are basically trained on blog articles, documentation, best practice guides that are all available on the internet of how to tune these databases. And so on the research side, what we found is the LLMs can get you maybe 85% of the way there and tuning certain things.

  2. 34:24

    And obviously there's weird corner cases, bespoke algorithms can, can handle those. But like our models oftentimes would take hours and hours to train, assuming you had enough training data to produce like, you know, the, the pristine optimal configuration, but the LMs can come in just like in 15 minutes, produce something that was good enough. And that's good enough for most people. So are we there yet?

  3. 34:45

    are we in a world where database can largely be sort of self-tuned, self-optimized? I mean, the answer is yes, because it depends on how much improvement you actually want. Like I'm saying, like, the LLMs can get you 85% of the way there. Is there always more performance to squeak out? Yes. And some people are always going to care about the things, and then therefore the automated methods we built in the past would matter a lot more.

  4. 35:10

    But for most people, actually, one of the things I was surprised so much was, like, when we had the startup and trying to do this auto tuning stuff for Postgres MySQL, is how many people... That was auto-tuned, right? Auto-tuned, yes. But it's how many people were running with the out-of-the-box configuration from Amazon or Google, like, in the cloud platforms. And they would tell us, oh, we thought Amazon was tuning this for us.

  5. 35:29

    And like, no, they're not, right? Like, so there's enough people that are maybe running with a DOL configuration, which is that thing's just terrible, don't do that, where, like, an LLM could come in and take care of some little hanging fruit and provide some meaningful improvements for people in a short amount of time. So in the rest of the software world, we've seen the nature of the job of a developer profoundly change.