00:51:47 · John Coogan, Jordi Hays
[00:51:47] S1: Yeah. [00:51:47] S2: That that that you're not building. I was talking to some people at a at a like a financial institution this week and obviously every finance company in the world is dealing with people like sending their agents to like poke around and do things in their product and they're like scared about that for for a good reason. But to... I was telling them this is just a signal that like you need to do a lot more Mhmm. In your product to to have that same functionality and I think that's what you guys have been doing for like, you know, the last two years. It's like make sure that that the the same, like, sort of, like, intelligence is available wherever you're sort of, like, starting a project from.
00:52:23 · Praveer Melwani, John Coogan
[00:52:23] S4: I also think you have to think about, you know, what are the things that you can do with lower latency, higher quality, and lower cost on your platform that makes it better for for a user to feel like they should be leveraging your first party first party, you know, AI use cases. So we've been developing a first party model now for for some time. You know, we doubled the amount of actions that are powered by that first party model over the course of the quarter. And we hold a whole bunch of unique context of our file system of, you know, the the Figma architecture so that we can actually build a smaller model that's lower latency can can work through those tasks much faster. So I think you have to think workflow when you think about, like, what are the things that you really want to own and where are there areas where it's it's helpful to, you know, almost use the MCP as an acquisition vehicle or rather part of part of the process versus feeling it like like it's combative in any way. [00:53:09] S1: What is it like doing an earnings call when there's all this chaos going on but it feels like you have a pretty clean story? Like there's like the semi trade is like going is getting nuked and then there's the midterms and it feels like you're just presenting like the business is still working, we're still growing, everything's moving along. Do you have to answer questions for... To investors about what's happening in the market broadly, or are you able to kind of just stay focused on the core metrics?
00:53:09 · John Coogan, Praveer Melwani, Jordi Hays, Keller Cliffton
[00:53:09] S1: What is it like doing an earnings call when there's all this chaos going on but it feels like you have a pretty clean story? Like there's like the semi trade is like going is getting nuked and then there's the midterms and it feels like you're just presenting like the business is still working, we're still growing, everything's moving along. Do you have to answer questions for... To investors about what's happening in the market broadly, or are you able to kind of just stay focused on the core metrics? [00:53:38] S4: I mean, I think you definitely take the context of where the world is because you recognize what are the things that people... That are on people's minds and how you can orient the story in a way that, you know, is is both appealing to the way that they're thinking, but then, you know, also is additive in the framework of how they how they view Figma. So, yeah, we go through a bunch of fun back and forth just to to brainstorm what are the things and themes out there. [00:53:59] S2: Oh, sure. [00:54:00] S4: But but I think outside of it, you're right in that, like, our story is clean. I think we've been really, really clear that we're focused on the long term here. Mhmm. And we'll make investments. [00:54:08] S3: You
00:54:08 · Keller Cliffton, Praveer Melwani, John Coogan, Jordi Hays
[00:54:08] S3: You [00:54:09] S4: know, one of the things we did over the course of the quarter was reprice our credits. Sure. We actually doubled the allotment that folks get on our pro plans and grew the allotments on our organized price plan 60% because we knew that in doing so, we experimented our way in there, it would drive credit consumption and, you know, Jevons paradox was starting to play out for us. Yeah. Yeah. And so I think we need to be long term oriented. We need to kinda, like, stick to our guns there, and I think it's the consistency and the consistency of our our execution that'll, you know, that I I wanna be able to use as a track record to to to point back... To point investors back to, and we gotta keep blinders on it and keep going. [00:54:41] S1: I love [00:54:42] S2: it. How are you thinking about token budgeting? There was a headline this week that I guess Apple is sort of like doing reductions or or like adding a lot stricter limits on how much engineers specifically can spend Yeah. Via via API on a on a day to day basis. But what's what's been your philosophy and approach today? Like knowing Dylan, he doesn't seem like somebody who would wanna be like, you know, like he wouldn't wanna have blanket rules around some of this stuff. But I'm curious. It's it's a
00:55:13 · Praveer Melwani, John Coogan, Jordi Hays
[00:55:13] S4: tough challenge. Yeah. I think that I think you you need to take a little bit of an an iterative approach there. So, you know, setting soft limits, giving yourself an ability to really understand what's going on with your... You know, across your... Across the entire company and using that to feedback into the overall strategy and implementation of what you're doing. I think one of the reasons why we, you know, we raised our op profit guide for the full year by 30,000,000 is because we found opportunities to be more efficient both in how we're serving our, you know, our our our AI products for our users while also then running our business more efficiently. You know, we found, you know, some of our engineers were previously spending, you know, tens of thousands of dollars be able to run, you know, run similar similar tasks that they had for for fractions of that as they start to, you know, come up with different... Leverage different model types. Sure. Level... Leverage different architectures, and then they start to share that information with others. We just hired a CIO who joined us from Coinbase, a guy named Casper Sorenson, who I'm really pumped to to be working with. And Awesome. Do a lot of that same thought they had there at Coinbase where, you know, we we also are of similar similar minds. Yeah, we'll we'll be able to to see what we get up to over the next little bit. [00:56:16] S1: Awesome. Fantastic. [00:56:17] S2: Well, great to meet you. Congratulations on the the fantastic supporter. [00:56:20] S1: We'll talk to soon.
00:56:20 · John Coogan, D.A. Wallach, Jordi Hays, Keller Cliffton, Praveer Melwani, Max Siegelman
[00:56:20] S1: We'll talk to soon. Ad[00:56:21] S0: Yeah. Cheers. Ad[00:56:21] S1: Have a good one. Goodbye. Let me tell you about the New York Stock Exchange. I wanna change the world, raise capital at the New York Stock Exchange. And our next guest might be there any day now because Keller Cliffton over at Zipline been on an absolute tear. Back on the show for the fourth, fifth, sixth time. I can't keep track, Keller. How you doing? Congratulations. [00:56:42] S2: You Thank guys landed. What's the latest? Good. Yeah. [00:56:47] S1: Give us the latest We've [00:56:49] S3: been busy. I'm I'm standing in our factory which we just quadrupled in size and we Again? Quadrupled the... Yes, again. [00:56:56] S2: Four heads. Heads. Four heads. We... I'll show you guys around. [00:57:03] S3: This is basically like end of line. [00:57:06] S4: So, you [00:57:06] S3: can see aircraft that are waiting to go out into commercial operations here. Wow. You know, Zipline Zipline builds... And keep in mind... So in this space, we not only build the aircraft, but we build the droid, the small robot that we're using to deliver to people's homes. Yeah. We build all of the ground infrastructure, the docks, which is what charges the aircraft, and we build the drop boxes, which is sort of how we integrate with partners like Sure. Walmart or, you know, any of the other large buildings that we're integrating with. So all of these sub assembly stations are where we are... [00:57:39] S5: Let me try