Supabase spends about $300K a month on AI tools

Transcript

From the transcript Latent Space publishes with the episode.

1. 1:11:13

   [laughs] So we have, like, the auth product, and we've always been very insistent that they should make their own decisions, like, on what tools they use, like, how they behave as a team, like, the reporting structure, and the AI policy was kinda no different. It's like a lot of people asking, "What is the AI coding policy?" And we were like, "Well, just it's up to you as, as your team." I think we have started to consolidate a few bits, especially around cost.

2. 1:11:40

   But as much as we can, we'd like to keep that model of, like, let the teams decide how they wanna run things. That's it, very decentralized. Mm. [laughs] I, I would say, um, I am trying to figure out a number for company leaders to take home with them as, like, what is the market standard. And so, uh, what, what are data points you can hang your head on here? People say, uh, you know, the average, uh, person is spending, like, $20 a month, and then, uh, average employee is maybe $200 a month.

3. 1:12:07

   OpenAI, uh, you know, at the, the sort of 50th percentile is spending maybe three to five thousand a month, and the top percentile is in the maybe hundreds of thousands. Mm. Right? And so that is the rough idea. I think, uh, a good metric is basically what is your ratio of token spend to human spend, and it's r- like, for a lot of people, it's maybe 0.1, uh, tokens to, to one human salary. And, uh, you know, you can probably go up to one pretty safely, and if you're over 1.5, you're probably in trouble. [laughs] Something like that.

4. 1:12:37

   So I'll, I just, uh, say, uh, luckily, as a remote company, basically everything that's- Documented, yes... ever happened is documented. [laughs] So I can ask our AI how much we're spending. So currently, we spend about 300K per month. Yeah, uh, for 600 people. No, for- So like 450 people... 400. Four, 450 people. Yeah. Okay. Yeah. That, it gives people an idea. But, but what you've said about, like, scaling the cost in terms of related to the number of employees makes sense also for us because we are hiring very quickly.

5. 1:13:08

   So if we end up getting a lot of leverage out of just, like, actual spend- You should be, right?... then yeah. You, you need a whole bunch of serverless employees that can spin up and down as, as you need, right? Yeah, exactly. Wh- which is like, yeah. Like, you know, I, I just had a CTO of, uh, you know, a- another top startup say, like, he was worried about his company spend. And I'm like, "Well, actually your- that's your investment in everyone else, and probably is very skewed," right?

6. 1:13:28

   Like there's, there's like two or three people that are super insane at your company, and everyone else is like probably not using AI enough. Yeah, yeah. Yeah. [laughs] That's exactly what we see as well. [laughs] We've got a leaderboard, and people... We put the leaderboard up, and usually people are a little bit ashamed, so they kind of like- [laughs] Yeah, 'cause you're wasting money probably. Yeah.

7. 1:13:44

   [laughs] And that's also the wrong focus. Like, the focus shouldn't be on how much you're spending, but, like, how much end-to-end... Like, in terms of systems thinking- Yes... you should be thinking, like, in terms of how many end-to-end processes are you actually solving that where you no longer need a human in the loop, or collapsing down long-running human problems that suddenly become very short problems.

8. 1:14:04

   So that's where we're spending most of our investment on, like, how we're using AI. Luckily, you know, as you saw, like, every single thing that's ever happened in Supabase is documented because we're an async company. Yeah. So building a company harness or, like, an agentic harness around Supabase is basically just formalizing. You can point AI at it and say, "What are our principles?" And it'll docu- d- document down all our decision-making principles for the whole history of Supabase. So that's where we're spending a lot of our time.

9. 1:14:33

   Before we move on from people, we do have, I think, now, like, 50 open roles. [laughs] Yeah. And we are fully globally remote, so we don't care where you are. If you're very good- Same, same salary everywhere, right?