Dive Radio: Bumpy Takeoff

[00:00]. [00:01] Thank you. [00:03] - [00:05] - [00:07] Thank you. [00:13] Thank you. [00:15] Thank you. [00:20] - [00:21] - - [00:28] Thank you. [00:30]. [00:34]. [00:40] Thank you. [00:43] Thank you. [00:45] Thank you. [00:46] Thank you. [00:53] Thank you. [00:55] Thank you. [00:57] Thank you. [01:27] Thank you. [01:35] *Loud music* [01:37] you [01:41] - [01:54] *phone rings* [01:56] I'm going to go to the next one. [02:03] *Loud noise* [02:08] *knock* [02:09] Oh! [02:11] Thank you. [02:12] *knock* [02:12] Oh! [02:14] - - [02:20] What? [02:23] - - [02:35] Hmm. [02:36] Thank you.

[03:06] - - [03:15] and desire for people to just see the flows and the different experiments that people are running and the different learnings that others have. The tooling kind of increases our capacity in the day. We have all the things we can do now because the tooling makes those kind of rope things easier. [03:36] I'm able to go through the spectrum of explorations that are so much larger than I could do in a day [03:45] Agents are using tools. What does that look like and what does that mean? Show up, try things, be curious.

I think the future is bright, but it has to be built. I always see things that I didn't plan on seeing. I'm making micro decisions that I didn't have a grid for a year ago, and it's only because of the ability that I can see. [04:15] that aren't fully automated. The ability to effectively multitask at a really high level is one of the core skills of operating in this, like, AI world. No preconceived notions about how things should be done. This is the moment in time right now to free the rewards of the public relations.

What do you do with a question, what if? Hey, guys. Hi. What's up, design fam? This is Chloe Park. Jenny Wen. This is Soleil. [04:45] And you're listening to the Dive Radio Show. [04:53] That worked. [04:54] We did it. We actually did it. [05:01] You don't even know what the last 90 minutes has been like. [05:04] The last 90 minutes, I'd say the last 24 hours. Yeah, let's call it maybe 36 hours. [05:10] So hold on let's let's talk about it. What have the what have the last 24 hours been like for you because you have a lot of stuff you're doing right now?

[05:18] I mean, there's a lot of things going on. We're actively trying to release a new version of InFlight. So I have Fable just spinning on one window, and then we're working through all of the audio stuff for the first time. And really, at the end of the day, all of the stress-- I do blame you, because we had a basic set of logistics that I thought you were kind of taken care of, but you stayed up all night working on making AI videos. So what's actually funny is I did-- [05:46] I have never used an AI video tool before, and I use Kling on that.

[05:50] intro we just saw last night. [05:52] And I didn't test the timer. And I completely covered the entire interesting part of that video with the countdown timer. So W for me. [06:02] Um, [06:03] On top of that, Fonz has been helping us with some of the refresh that we kind of want to show. [06:09] and get some opinions on. And that poor guy. [06:12] He started on Monday. [06:13] And we're like, hey, we need this by Thursday and... [06:17] our feedback was pretty... [06:19] Thank you. [06:20] Pretty thorough. I took a screenshot of something that he's just like writing notes and all these Figma files and we'll show it because it's kind of fun.

But it's just iconography inspired by deep sea creatures sticking to the side of a ship. [06:36] So if you think your client briefs are annoying, you have no idea what he has been through the last like 72 hours. [06:44] Yeah, yeah. Poor Fonz, who's also just an absolute pro. Yeah, I do want to call out if you're if you're joining us, this is obviously the first live stream. [06:54] of dive radio and uh [06:56] Very bumpy. And you're going to be along for the ride as we kind of evolve this thing over the...

[07:01] course of time. [07:03] So thanks for bearing with us. [07:05] Throw tomatoes. We like them. It was tempting. It was tempting to push back a week and just get all of the ducks in a row. But then we kind of thought about what would be fun. [07:14] And so this is almost going to turn into a behind the scenes pilot where you can see like we have a bunch figured out. We're going to try some stuff, but also. [07:23] So much is like in progress and we're actually having a blast working on it and thinking about it.

And so we're kind of just going to pull back the curtain and show a little bit about. [07:31] what's happening behind the stream as well and and before we do that i i'm curious to hear from you i i know why this live show is really interesting to me but [07:42] Why is this like an interesting thing to you? You've already got dive club, the podcast. [07:47] Thank you. [07:48] You know, I don't actually need a lot of logic behind it. I really follow what brings me life. And it's easy for me to identify what that is.

And I've only done a few live things. I've done stuff with Kyle Zantos, who I think is in here. I've done stuff with James McDonald before. Every single time I... [08:06] just, [08:07] jump on a live stream and hit record, I walk out super energized. And I used to do a lot of teaching, too. I would do Figma training stuff. And so I think maybe there's a part of me that just [08:17] likes this. [08:18] And you don't get it in the same way with the recording and having things be more polished and [08:24] Oh, brings me life.

[08:25] What do you want it to be? [08:28] *laughs* [08:32] Maybe I just want to talk more. Maybe this is really the subconscious answer. I'm just like, ah, I can actually just have like 50/50 back and forth. [08:40] Man, there's so much to talk about at all times. And the... [08:45] it's not even like just new cycles, right? Like the rate of change in tech and definitely design is so [08:51] fast and accelerating rapidly that it [08:54] It's hard to actually keep up with that in a [08:58] productized [09:00] podcast.

[09:01] And so to just be able to [09:03] look at cool things together process it riff like just have a sounding board i'm thinking about all types of ideas even as i'm doing the interviews which i'm sure that you are doing too and so creating a space to kind of almost [09:16] digest publicly, [09:18] I think, I hope that that's going to be [09:21] really valuable not only for people listening but also for us too [09:24] I heard Anthony Bourdain, a clip of him a while back. Yo, Tommy, bump your audio.

Bump your audio. Get that mic a little bit closer. We have a couple people saying this is – [09:33] you're a little bit soft. [09:36] Hello, hello. [09:38] Yeah, there's a clip of Anthony Bourdain and he talks about just how people who have the guts to be curious. And I was like, that line stuck with me. And I feel like we're in that phase right now. I love talking to everybody who's going through different things. [09:53] layers of everything that's changing, but I really want a place where I can kind of escape [10:00] and hang out with all the designers who have the guts to be curious.

We're kind of exploring this crazy world and it was funny, we were talking about like, "Oh, we should show workflows and these things that we're using in the work we're doing." [10:12] And you said something I really liked, which was, yeah, but let's not call them workflows. [10:17] I think it's better to call them ideas. And I think that's... [10:21] Smart because like we don't we don't have prescriptions like here's just like I could walk you through the wild process that I went through to create that intro or that radio of dive dot radio.

[10:34] And it feels pretty good. [10:36] There are things I'd do differently. I don't think I'm ready to prescribe a workflow to it, but there's definitely some ideas of ways that I would repeat that. [10:44] and things I'd do different. And I feel like there's a lot of that missing. [10:48] in the noise. There's a lot of hype, there's a lot of people making big claims. [10:53] I kind of just want to see how people are messing around. [10:56] And that's hopefully what a little bit of this show is. [11:00] Yeah, even just creating space to show how we are messing around too.

It's like if you put something in a, [11:06] like a YouTube, there's like a permanency. [11:09] to the podcast right but if you try to make content that is up to date it's like the opposite of evergreen it becomes out of date in like 90 days tops you know but i'm constantly tinkering and trying things and like looking at other tweets honestly like i come up with so many twits bookmark things like oh my gosh i have to try that you know try it okay how did it work cool like does this become a space to share that kind of stuff maybe we'll see you feel like this new twitter update that x algo where you can now see all the people you follow is that useful to you [11:39] Things going to stick around?

You think that's going to create just like... Yeah. Yeah. [11:44] I saw so many people say that they were planning on leaving, but then decided to stay because of it. That's interesting. My only question is, what the hell were they waiting for? Like, what? You're telling me this was a switch? You could just flip the switch? Unbelievable. Yeah, yeah. I want to say hi to everyone who's jumping in here, guys. Josh, I see you. I see Josh Puckett on Contrad. What's up, guys? I see Janelle. Yes, dude. Jeremy. Jeremy. [12:07] Eric, good to see you Chris.

[12:10] Chris doing some crazy magic at the Detach party, by the way. If you came to Detach this year, our party we threw at Config... [12:19] That was a pretty good time. And what that taught me, and that was our third year throwing it, but what it really showed me is I want to do so many things. [12:28] more IRL things. [12:30] Small events, big events, parties, hikes, chats. I wanna do one of these shows. [12:37] in real life. I think that would be really cool. [12:41] I'm down, man. [12:42] Let's do it.

[12:44] First, we have to figure out what the show is, but then, yes, we'll do it live. And it's my hope that this show every Thursday, about 90 minutes, becomes a staple. You know what? Dive Radio, I liked this name because I drive home from work way back in the day, and there was a couple select radio programs that would come on in the late afternoon time. [13:04] And they'd always have like segments and they were really cool. And I'd love to evolve that into this. One of the things that I really want to do.

[13:11] Because we already do interviews with people. We don't want a show and we're still doing more interviews with industry folks. Yeah. [13:18] You can catch that elsewhere. I want to talk to the community here. We're going to do call-ins. Admittedly, with the delay we had today, it was trying to solve the call-in. We're very close. We don't have the call-in solved today. We will have call-ins solved by next week. So bear with us. But it doesn't mean we can't answer questions in chat. I'm looking at it here, guys. [13:39] remote work IRL party what's up Kyle good to see you man and Kyle Shepard too [13:43] Did we turn off the chat?

We did not turn off the chat. [13:46] And I would say this too, we're streaming to both of our YouTube channels, both of our X channels. So we see everything that's coming through. [13:55] but you might not see some of the community, I would probably recommend if you want to jam with each other too, go to the Dive Club YouTube stream. It's probably a pretty good... [14:06] Pretty good group. [14:08] Dude. [14:10] It's been about a week. [14:12] and change since I think we've been able to compare fable five and GPT 5.6.

And there's already rumors of like, [14:21] opus five right around the corner and all these things and by the way we are going to talk a lot about ai how can you not [14:27] How can you not right now? There's so many things to talk about with it. [14:31] What's your daily driver, man? How are you feeling? [14:35] He was funny with all the Fable stuff came out and... [14:39] I blitzed through my credits twice. Yeah. So I think there's actually like a new skill to figure out how to be more token efficient.

And I've built like, [14:47] delegate skills where I can like have Fable spawn sub agents and report back. That's working pretty well. [14:54] I have had multiple times where I've just been in a loop and I can't solve something, but I'm still using like I like called. I've been using 4.8 Opus as my daily driver. [15:03] I've started doing more GPT. So you're using 4.8 as a daily driver. [15:10] It's been my daily driver for so long, and now I'm like actively going like 50-50. [15:15] between that and GPT. Like it's very, it's very clear.

I think that five, six is, [15:21] I probably should just make a shift to do that. [15:23] um [15:24] I pit them against each other right now. This is kind of what I've been doing a lot of. Did you see – What about you? Did you see – Well, I mean I – [15:31] it's hard because I'll say this 5.6 seems to be I really like it it's obviously very efficient like I feel like I'm just not hitting my limits and I'm blowing through them on fable. [15:43] Yeah, same. [15:45] I feel like I get really...

[15:47] I feel like I'm on like mushroom trips with fable. Like all of a sudden it'll give me some insight and I'm like, Whoa, whoa, [15:53] That's really... [15:55] interesting, whereas 5.6 feels like I'm working with an extremely competent engineer. [16:01] Yeah, I think that's secret. You know what I mean? Like one is like I go to the village noble and the other is like this is the engineer in the room. But I will say I've started having more open ended brainstorming because I'm having so much success thinking through more meaty challenges with Fable.

[16:19] Like really open ended. Let's just do you orchestrate with fable and then and then get like do the grinding work with 5.6. That's a that's a workflow that I've seen a lot of people do that I've been doing lately. [16:31] Yeah, I have been doing that. I've actually also just been leaving it up to Fable, too, about what it wants to spawn. But yeah. [16:39] I don't know, I find myself chatting with Fable more like I'm having more [16:44] brainstorm me back and forth than I was before because I feel like the output from Fable is so good.

[16:50] Like, it's so good. [16:52] I want to hear from the chat what you guys are rocking with as your daily driver. Where's your preference at? Have you had any of your – [17:00] code base is deleted. Have you seen this? [17:03] I did. [17:04] You feel for those people, man. At first I thought it was a joke. They're like, oh, it deleted all my root files. And then they're like, no, it deleted my whole project. [17:12] Then I saw... [17:13] open AI say we're looking into this and I thought, [17:16] Well, that's kind of a...

[17:18] That's kind of a weird thing to suddenly be popping up as a... [17:21] recurring issue. I don't know how recurring it is or if I saw one tweet and then I saw all the tweets around it. [17:27] But I thought that was kind of interesting that that was happening. You saw Grok getting a lot of... [17:32] Uh... [17:33] I guess they were uploading code bases. [17:36] And then they decided to turn it off. [17:39] in some weird way. So a lot of interesting stuff getting more and more compelling to try and get local models working.

I installed a local model a week ago. [17:48] Oh, how's that going? [17:50] Thank you. [17:51] There's your answer. I've got a Mac Mini. Yeah, I'm not in love with it yet, but it's like I could see it because listen like [17:59] The frontier models are still just like they're just on a different level. [18:03] You know what I mean? And I want I want to believe in local and a lot of their estimates that local models are going to be at like the four point eight level within the next six months.

Who knows? [18:13] Um, eventually though, it will be extremely compelling and, um, [18:17] I'm worried, so I do a bunch of video stuff, and so I've got, [18:22] like 20 terabytes of footage that I have to store. [18:25] And I... [18:28] I'm realizing whether it's buying a server for storage, they're buying like compute locally. [18:34] We're going to need to build some stuff there. Like the Mac minis are already out of like, they're already either out of stock or they're extremely higher price than they were. [18:42] I want to just build a rig and just be ready to install the local models when they're actually worth using.

[18:48] I've had a Mac Mini sitting unopened on my desk for like four months. [18:53] Because my co-founder, Kyle, just sent it to me in the middle of all the open call-out buzz. And... [18:59] I don't know, man. I just didn't want to die. I just didn't want to dig in. It just felt like a lot, but I just set it up. This last weekend, I'm like, I have the base foundation. I'm like, okay, now what? I think a lot of it is kind of preparing for local, though. You're going to learn the UX challenge.

There's like a weird challenge here that now when you're connected, there's this, you're connected to multiple devices. You're on your primary device, and then you're [19:22] screen sharing into your Mac mini and you're using the apps in there too. And it creates this like, and apparently like there are people who have multiple Mac mini, so they're having to do this. And I'm like, [19:33] What a weird... [19:35] It's weird. Challenge. It's a weird, it creates this weird UX thing that like we need to solve somehow. I don't know. Yeah. I was even like, I use clean shot all the time.

And so I can, I can seamlessly move my mouse between my laptop and then my monitor, which is now doing the Mac mini. But like the screenshotting is a different. [19:53] UX on both of them, it's so trippy, man. It took me a second to even just orient myself. [20:00] I think I mostly hated it, but I believe in the importance of it in the long run. So I'm just like forcing myself to go through it. [20:08] all of the logistics of getting it set up. So I'm curious if there's anything that's caught your eye in the last week or so.

For me, I've been messing around. And I know, like, you use Super Whisper? What voice UI do you use? Yeah. Super Whisper? [20:23] I use Super Whistler. I do too. I use it for everything productivity related. And GPT Live came out, and so I've been messing around with that. [20:33] And I like it. Like it's really, it's fast. It moves. It's only at 5.5. So I'm not using it for work-related things, [20:40] But I am like I had it help me in the kitchen and I was cooking stuff. And it was funny.

I had this moment where I'm just like yapping with with it. And it's joking with me. We're going back and forth. I'm like, oh, this feels like. [20:51] This feels like the movie Her. [20:53] Because it's like, we're jamming. And I make this like wonderful sandwich, breakfast sandwich that we put together. [21:00] And so I have Alexa set up over here. [21:05] And while I'm talking to it, I turn to Alexa and go, hey, Alexa, set a timer for five minutes. And then the thing's like, no, no, no, you need to set it for 10 minutes.

But at five minutes, I'm like, oh, I'm like, Alexa, actually change that to 10 minutes. And Alexa just bricked like that task was just not feasible. And I was like, oh, darn it. I forgot. You're dumb. You're a dumb robot. And so I had to like, you know, reverse the steps. But what it made me realize is the UI, the voice UI, I really, really like I'm still going to use Canvas. [21:35] there's still gonna be a need for screens [21:37] I think there's a world not far from now where voice UI is actually my primary interface for most technology.

[21:44] I think we're so close to it. And now OpenAI is talking about their first hardware release, that speaker... [21:52] Voice puck. [21:54] And [21:55] Do you know when they said that they're thinking about releasing that? I didn't see a date. I'm not sure if there was one. I'm going to look this up real quick. It's so, just like a quick sidebar too, like if anybody here has not [22:07] just played with the new voice mode in GPT, it's so worth 10 minutes of the time. Like it really is trippy.

It's super cool. [22:14] last night, so we're having all these back and forths on just the million things that we're trying to figure out to make this happen. [22:20] And I got to pick my mother-in-law up from TJ Maxx, 9.30 m. And, you know, high on top of the totem pole of things to do. And I'm sitting in the parking lot waiting for her. And I got GPT going and I'm just talking to it. And it's crazy. [22:34] It's shockingly good. You know, it's right away. You're like, okay, this is a little bit trippy, but I'm, I'm generating images to mood board.

The thing that we talked about where like, maybe the final segment is the phone on the bottom of the ocean floor and people can call in. And so I'm, I'm. [22:50] playing with that and I'm playing with like [22:52] black photography stuff as well. And I'm just talking with GPT and then you can swipe down [22:59] and review the images so it like really felt like a creative director right like i'm having a conversation the images [23:06] me [23:07] in voice mode and I did it intuitively. Like it just felt like, yeah, I should be able to pull this down.

I pulled it down and it actually closes that window and you're back in your chat. So you can scroll up through your chat history while you're talking. Yeah. It's awesome. But my chat history is all images. [23:22] So I'm like reviewing imagery and having a verbal critique session about this [23:29] concept that I'm making in voice mode on my phone in a TJ Maxx parking lot like, damn, this is the future. [23:35] It was awesome. Yeah. Yeah. And I've been using it for other like super unproductive things. I play a lot of Dota 2, which is like a 4D.

[23:44] chess computer game. And so like I, I wanted coaching. And so I have it live and I'm talking to it while I'm talking to my teammates, playing this like super complex game with all these different variables you have to manage. And it's given me all these tips and it's coaching me through this whole thing. And I recorded that session and then I recorded and it was funny. It made a, I've been playing this game for a decade plus. Yeah. [24:05] And it made a recommendation. And I was like, no, no, no, no, like that's wrong.

And it was a very like nuanced recommendation. [24:12] And it was like, just trust me, do this. You're going to need it for this thing. [24:16] And then five minutes later, I did it. I did it for the thing it mentioned. My teammates go, that was one of the big-- that you probably just won us the game because of that Lotus move. And I was like, that's crazy. That's pretty cool. It was just like a cool little moment to show that this is happening. [24:31] Um... [24:33] You see these you see these talk socks.

[24:36] Have you seen these? [24:37] This made me think of you. [24:42] And I want to show, it made me think of you because [24:46] - - [24:47] I know that you're walking around the house [24:50] doing this. Oh yeah. [24:52] Oh, yeah. [24:53] You've seen these? This is ridiculous, by the way. [24:58] Thank you. [24:59] This isn't like there's a real prototype of it somewhere. [25:02] Like this is a thing that people are actually using. [25:06] You wouldn't use this? [25:08] No, it... [25:10] It's... [25:10] No, it makes me physically uncomfortable to look at that thumbnail.

[25:14] Wait, how much is this? [25:16] $2.99. [25:17] What a steal 292 299 by the way my sound not working on this one. I thought we solved this did we not? [25:25] Thank you. [25:25] That sounds about right. [25:27] Sounds about par for the course here. Yeah. [25:30] It's okay. We'll get there. Yeah, these noise-canceling masks. This is ridiculous, dude. [25:35] Thank you. [25:35] Do you feel rude to [25:39] when you are [25:41] talking to your AI, do you try to like sneak away out of earshot? [25:47] Yeah, I didn't realize this until I was working in person out of South Park Commons.

I was in their SF offices. [25:54] And I didn't realize how much I was talking to AI until [25:57] that [25:59] Weak. [25:59] And then I was like, I feel so uncomfortable. [26:02] I didn't have one of the little like, you know, whisper mics kind of thing. Um, [26:07] But... [26:09] I don't know. I found myself just like sneaking into closets in order to like talk with AI. I don't know. I do feel I do. I don't feel rude, but I feel self-conscious. I'm also the kind of person that would like never answer a phone call in an airplane.

You know, that just gets me inside. I just thinking about that idea. [26:28] Are there moments you prefer to type verse talk I get I get like I'll sometimes not do a task because I'll be doing three other tasks and it's like oh I have to type that one out like I'll I'm not gonna do that right now then because for some reason I can't like if I'm not at my phone and I'm at my computer I don't even like to have to like. [26:44] I got to open Super Whisper.

I just I'm always on my phone yapping. You know this. [26:48] We were debugging this even before. Like in the 30 minutes, we were working through all these different audio issues, and it's just-- [26:55] Us talking with AI, both trying to solve our own issues, [26:59] It's a weird little glimpse into the future. Thankfully we don't have TalkSox though because that [27:04] feels like a Black Mirror episode that I don't like. Can I just, yeah, that does. And actually I want to call out Brad just said, I don't want people to see how dumb my prompts are.

Yeah, exactly. [27:14] How are your how dumb are your prompts? [27:17] I contradict myself in the middle of the problems all the time. I'm like, da-da-da-da-da, do this, do this, do this. Actually, wait. [27:23] "No, don't do that. Let's do this instead." I just hit send on that all the time. [27:28] I love working with AI because I have the freedom to not find the right words and to be super verbose and to just sound like a complete moron. And then I'm like, oh, we're going to go live and show people how we do our things.

I'm like, oh, no, that's terrifying. Like I misspell almost every word. [27:47] It's a mess. [27:48] I feel you, Brad. I feel you. [27:51] Um... [27:52] Are you a car guy by chance? [27:55] Thank you. [27:56] Is this your lane? [27:59] Depends on what you mean by car guy, because when I hear car guy, I think like appreciation of the decades and an understanding of how anything in an engine works. [28:08] which I am not. [28:10] Am I interested in some of the more futuristic concepts that remove me from a driver's seat?

Yes. [28:17] So I think that makes me not a car guy, probably. [28:19] Are you rocking the... [28:23] I saw this. [28:25] I... [28:26] I love this. [28:29] What do you love about this? [28:30] It looks... [28:34] Pull. [28:35] on something that is carrying around. It's like moving me around my vacation resort. [28:41] This is what I mean by being a car guy. It's like, [28:43] I think this is kind of fun. [28:45] But I wouldn't drive a charger or anything that made noise. [28:49] Is this... [28:51] Is this like a road legal?

This is a road legal car, I'm assuming, right? Yeah, yeah. I was assuming so. It looks like a golf... [28:59] Cart, hybrid meets. Yeah. Yeah. [29:02] There was another one of the golf carts that went super viral like 10 days ago too. [29:06] It's like, it feels like, I don't know, but the fact that we have all of these new, I don't know what this one cost. The other one was like 20 grand, dude, for a golf cart. [29:16] But the fact that [29:17] Like people obviously believe in this, like they believe in this as being an important [29:21] method of transportation in the future.

[29:25] which I think is odd. It's not something I saw coming. [29:27] Yeah. [29:28] Kind of interesting. I saw... [29:31] This was interesting, the Tinder rebrand. [29:34] What do you think it is? I sent it to my co-founder, Kyle, who is one of the first designers at Tinder. Are you allowed to share his thoughts? Ah, man, I should have pulled it up. I think he just laughed at it, basically. [29:50] I, you know, like I can see the mood. [29:54] They're going for it here. [29:55] Big mood board fan, by the way.

[29:58] But on an island, I don't... [30:01] You know, it's yeah, it's not bad. It just doesn't do anything for me. It's fine. [30:08] This is like the thing that we were talking with Fonz about. [30:12] Fonz is an absolute pro. We've both worked with him on other projects. Like he's one of the top three people I recommend every single time someone asks me for his skill set. And he gave us a bunch of like quick mood variations in like six hours. [30:25] and they're beautiful. [30:27] Right. And I think some of the feedback was like, [30:31] These could fit in a lot of different contexts, different startups.

[30:35] And I think your feedback was, [30:37] I want something that if it was put on any other startup's website, it would look absolutely absurd. [30:43] Like what is I use the word foolish. I said I wanted to be foolish. If any sass like ingested our design MD file and used it, it should feel foolish. [30:52] They should get angry letters from their investors. [30:55] You got some angry letters recently. [30:59] When I look at the Tinder rebrand, I feel like, yeah, I could probably interplay this. [31:05] across a lot of different contexts.

That's the next hot HR startup. [31:10] Yeah, the next like data docs. [31:14] Yeah. Let's go. [31:15] Speaking of hot... [31:17] startups. I don't even know. Can we call Cursor a startup anymore? [31:21] Dude. [31:22] Yeah, how do you define startup? It certainly can't be by valuation or size anymore. [31:29] - You know Mamuso here? [31:31] - Manuel? - Of him. I don't know. I haven't met him, no. - I met him when I went and filmed with Vercel. - Oh yeah, you did on site. - Beautiful human being. Just a wonderful person.

[31:42] I was kind of shattered. I was like, there's no way. He left Vercel at some point in the recent past. [31:49] And just found out he's at Cursor. [31:51] And then we found out in like almost the same 24 hour period that Jenny Wen left Anthropic. [31:58] Like six hour period, dude. Went to Cursor. She went to Cursor as a head of design, by the way. And it sounded like I immediately bopped over to Riolu's... [32:07] Twitter and I was like, what's going on here? And it sounds like he's like, now I get to just build.

[32:13] And so it sounds like he really wanted to lean into... [32:16] Sounds perfect. [32:18] Do I get it? Imagine, head of design and you're scaling that quickly, that's a set of design challenges that are operational and organizational. They sound horrible, to be honest. They sound horrible. To wake up every day and have that on my plate sounds tough. I'm happy for them. It sounds perfect for both. [32:34] Rio and Jenny. Like what a huge, huge deal to get [32:38] brought in like that as Jenny. It's amazing. [32:41] They're a talent magnet, man.

And can we just talk about the fact that [32:46] I can't, timelines are insane. We're all aging in dog years, but like, it felt maybe, I don't know, six to eight months ago, that Cursor was just like, eh, just life support from enterprise contracts. And now it is as cool as it gets, dude. Everybody I'm talking to is switching back to Cursor. If they're not already on it, they're attracting the most elite talent in the industry. [33:06] Crazy, crazy comeback arc. [33:08] Now, you're not a cursor user, right? [33:10] I'm not.

I'm still a team conductor. [33:13] - Cursor's a big tool in our household. My teens love cursor. It's really like, it's a really accessible product for them. - Yeah. [33:23] And I think there were claims, and I'm not validating this, but [33:27] That like at least at one point you were getting... [33:30] more tokens out of like your Fable 5 usage at one point. I don't know if that's true or not. [33:36] I didn't know that. [33:37] But they certainly were my my teens just like grinding on it. [33:40] I like Cursor.

I like Cursor a lot as a product. It's clear they're attracting a lot of talent. I think the SpaceX acquisition is really compelling. [33:48] I'm curious to see kind of where this is. It sounds like they're taking incredibly big swings too, like coming at GitHub. They said that that's going to be like a fall thing. [33:57] If [33:58] I don't know, it feels like every major player right now is building the same thing, but the same thing is just everything, end to end, full stack, every possible thing that you would need.

And so in a world where Cursor... [34:09] is an actual alternative [34:12] to GitHub, it's agent orchestration, it's everything that you would need in an IDE. They're probably bringing in more review functionality. They've all these tools have started showing the PR natively inside of the tool like, [34:23] You could just do every single step of the process in cursor and in that world, [34:29] Man, they're going to have a lot. I mean, it's going to be a magnet. It's going to be hard to leave. [34:34] Yeah, I'm very interested. I think the days of focused products are...

[34:41] I don't want to say they're over, but it is interesting and noteworthy to see [34:45] companies like Mid Journey is now also a health company. Okay. And like Cursor is going to be, you know, psychiatry. I don't know. Like what else is Cursor going to do? Because I think I [35:01] with cursor there was an initial reaction in the earlier days when they what was their name it was any sphere or something yeah and people were like [35:09] What are you trying to replace VS Code for? Like, there's no way you're going to do that.

VS Code is the tried and true thing. [35:15] And now it's like, [35:17] Who in here is using... I was a VS Code user. [35:21] Is there anyone here using VS Code? You can install AI plugins in Cloud, and they've got a great ecosystem. [35:27] Ridd, I know you're not using VS Code. I think the design community is going to shame anybody who says that they're using VS Code right now in the chat. I'm just going to throw that out there. [35:37] You know, and it's kind of, it's interesting to see...

[35:41] It's interesting to see the big swings becoming the defaults. [35:47] But it's also it's got to be hard, man. I do not envy you. [35:51] I don't envy anybody building software tools right now. I get a lot of FOMO. [35:58] talking to the people who are doing it. And then like, I talked to people offline and they're like, it is hard, harder than ever. Yeah. [36:06] Dude, it's freaking hard, man. It's freaking hard. Because it just feels like everybody's, like I said, like everybody's building... [36:11] Everything everything's so copyable right like tech is always bundled and unbundled and [36:16] It's since the dawn of time, it's been the cycle that we've been.

And this is just the greatest bundling in history because [36:23] the marginal cost of every new feature is like nothing. It's nothing. [36:28] - Yeah. - You know? I've prototyped [36:30] So many versions of all of these features [36:34] that all of these tools are building, it's easy, man. Anybody in this list can one-shot all of the visual inspect or agentation-style tools that are now table stakes in all of these tools. [36:46] So it's like, of course, all the best features and all the best UX approaches get copied in like hours, not weeks.

[36:52] Yeah, Liz, good to see you. Yeah, yeah, that's my understanding early on to a fork of VS code. And and so like open source, you know, how do you what has been your take on open source? Have you ever dabbled in open source? Do you have any like [37:08] inclination to go in that direction with where software is going? [37:12] I was thinking about it a little bit more recently. [37:16] Because have you ever heard of reflect notes? Maybe some people here use that. [37:20] That's what I've used for notes for a while, but then...

[37:23] there was this shift where I wanted everything to be local. [37:27] And I wanted to be able to change things, and so I actually built my own note-taking environment, because I didn't like Obsidian, so I was just reskinning it my way. [37:36] I left reflect even though I liked the product because it wasn't open source. [37:42] And then sure enough, like six months later, I think actually two days ago, they just released [37:47] the reflect, [37:48] open. [37:50] that I've been already starting on. [37:51] And I don't know, I feel like it's...

[37:55] There are certain types of tools, but that is a requirement, actually. [37:59] It has to be extensible. [38:03] What do you think makes it like, okay, so for example, [38:06] Grok build just open sourced. [38:09] the harness for their coating. Did you see this? That was yesterday. I didn't know. [38:13] And I thought that was so much news. It's like it's like, how do we get through it? And that's why we're going to talk a little bit about it before we get into the build fund stuff. [38:20] And so, okay.

[38:22] So Elon, so SpaceX, they're just like, okay, Grok, we're going to open source the coding harness around Grok. [38:28] so that anyone here can just build on top of it. [38:31] What do you think, what's the decision around that? What pushes a product to say, [38:36] it is beneficial for us. [38:39] to open source. [38:40] What would make the math make sense? [38:43] where you go, actually, in-flight should be open source. [38:48] Man, I don't know. [38:52] I'm going to ask all the hard questions. [38:57] Just throw it in the bucket of insane strategic decisions.

[39:02] that I didn't anticipate having to [39:04] be faced with. [39:07] I don't know. I really don't know. [39:09] Well, not problems I have to solve because I'm in the middle of trying to make money making media, which is also its own thing. Very, very hard. Can't open source media. [39:20] I don't know. Could you? [39:21] Did you open source media? Open source my voice, create an 11 labs. I was about to say, you could totally just license your voice, man. It's [39:30] Could I, though? Like licensing a voice sounds so...

[39:34] It's starting to feel antiquated. Like, do you need to license a voice? I'll just go to 11 labs. [39:40] There's better voices than mine, man. I was listening. I follow a lot of like voice over artists and voice actors. And there's like, there's just such differences to people who are really pros like they act it out and they make faces and they're just very good at delivery. [39:54] Yeah, I don't know. But yeah, I do want to talk a little bit about something that's cooled off a bit since last week.

[40:03] And I gave a take on this. I think we talked briefly about it. [40:07] But... [40:08] you know, I haven't had a chance to like vent about this and I do want to. And so I [40:15] I want to hear your take and I want to do a little bit of moaning [40:20] about Gal Shear, who, by the way, the most latest and greatest from Gal Shear, he's painting. [40:27] Gal is painting. It looks good, too. It looks good. Yeah. [40:30] I love this. I'm wondering how he films this, by the way.

He always has these overhead... [40:34] Looks like he's just holding the phone, maybe. [40:37] I think you're showing the cursor tab. [40:39] Oh, I sure am. Yeah. [40:41] There we go. [40:42] That's not the right one. [40:44] Mm-hmm. [40:45] There we go. [40:46] I love this. [40:47] So this is what Gal's up to. [40:50] The other shoe dropped, and it turns out [40:54] He was announcing. [40:56] that he is making art. [40:57] And I love that because everyone was so cynical about this take. I want to say something about this.

We're going to get into this. We're going to touch on it briefly. We'll move on. But I do want to say something because it really ticked me off. [41:11] This is why Gal [41:13] Earned. [41:14] hundreds of thousands of followers on X. It wasn't to engagement farm people. [41:18] I haven't followed him through his whole journey, but I'm fairly certain he saw some explosive growth. [41:25] After he shared that visual of headroom. [41:29] Right? [41:30] Let's find that real quick. One of the best brand release videos I've ever seen. [41:35] Like genuinely this was in so in film, we call this a match cut where you keep a central object in the focus.

[41:43] and the composition around it is changing. And he goes even further and extends that concept [41:49] And he's great at it. [41:50] Um, it's like, so he comes out and he's announcing, Hey, I'm retiring. Uh, I had a client that, [41:58] who basically used AI [42:00] And AI beat me. [42:02] Where is that post? Is it sticky? It must be. [42:05] you know? [42:07] And I just want to call something out. I do understand that there were people who read this and immediately felt like this was an affront to all the things that they've invested into and believe in.

But Gal never said design is dead. [42:22] He might have insinuated parts of that, but didn't say that. And he certainly never said, you, designer, should quit because AI is going to beat you. [42:31] Those things were never said. He said, I'm retiring because this experience with AI. [42:37] And he got dogpiled by the design community for this take. I mean, and even adjacent, you had VCs, you had the all-in crew talking about this. You had Dylan Field comments about this. [42:51] It made the rounds. I know everyone in this room has already seen this talked about.

[42:56] And I just thought that that was really interesting that [42:59] This is such a hot topic that AI can do some work. [43:03] that people are going to come at someone like Gal, who I think is – [43:07] His logo marks are good, but it's his packaging and the motion graphics that I think he's extremely spiky in. And if anyone... [43:14] challenges that I think that that's a I don't agree with that. What are your thoughts on that? I [43:20] I think there's, you kind of got to look at it from the output.

[43:23] that was produced by whatever the situation was, and then also the inputs that were given to the model. [43:30] Because... [43:32] I [43:34] I don't know if you saw, there was this article that went viral. It was career advice in the age of AI. [43:40] If you didn't read it, I'll just spoil it in two lines, because I actually thought it was really great. And he talked about the fact that the models are getting... [43:48] really good at executing against well-defined problems. [43:52] And for sure that's in code. And honestly, in design, [43:56] It's kind of the case, too.

I haven't told anybody this. Hopefully, the people from my team are not here, but I had Fable [44:05] ideate and generate some mockups about this feature set that I'm working on and [44:12] like visual editing of a PR kind of thing. And they're pretty good. You know, like, I made like, maybe a couple tweaks, they're pretty good. And I was showing the concept on zoom screen sharing, this was on Monday. And [44:24] my co-founder was just like, [44:25] Damn, you crushed it on that. That looks crisp. That's some of your better work.

And I'm gutted because, dude, it was mostly fable. It's mostly fable. That feels uncomfortable. And I think that that is probably a lot of what Gallant was talking about where it's like, [44:40] Given the thing that you want to accomplish, what is the output? [44:43] like AI is getting good at it, right? It's getting good at it. - Yeah. - Now, [44:47] The article, what I liked was basically the TLDR was don't be a problem solver. [44:53] which is what we've defined ourselves as, as designers for so long. [44:56] You know, execute against this problem, solve the problem, you know, be a problem finder instead.

That's where the alpha is. If you can find problems and set AI up for success to help you nail it down and create something worth sharing. [45:14] That's where the opportunity is, which is why I then go back to the gal thing. I'm like, what was like, is there a chance that maybe the founder themselves was [45:24] actually had pretty good ideas, you know, had some kind of novel insight. And like, given that, yeah, I mean, [45:32] when the output that you're paying for is [45:35] pixels and motion and colors.

[45:38] All of that can be trained. [45:40] That's tough, you know? [45:42] So I had this thought and the thinking was that, [45:47] Um, one of two things. It was either Gal was in fact farming for engagement was going to release a tool. He's released tools before has a color tool in the past. [45:57] Um, [45:58] Not really his style. So I was like, that'd be interesting. And I don't know that it would be very effective because he was in the comments like, [46:05] assuring people that wasn't going to be the case.

[46:08] And then the other thought was like maybe a client did go and farm all of his match cut motion graphics – [46:16] fed it to Fable 5, [46:19] And, you know, maybe maybe still had someone create a logo mark and like showed him like, look what you can do. I can do your I can do your your spiky thing. [46:28] Without you. [46:29] And maybe he was like, okay, I'm flipping the table, I'm done. So I tried. So I went, and I'm not releasing this, but I went and pulled his work.

And I was like, Fable, go look at this. And would you be able to create a Galshier skill [46:44] that would take my UX tools logo mark, [46:47] and build it into this brand presentation. And I'm gonna show you [46:51] what that looks like. I have to pull it up. Give me just a second here. [46:56] I'm so glad that you did this. I want to ask you. [47:01] How do you think it did? [47:06] Ah, man. [47:08] Let me share this. We're using like, like, re-motion. Is that... [47:12] What you're using?

Remotion. So Claude Fable 5 on Max and Remotion were the two things. [47:20] that I made use of there. [47:24] I think it's going to be cool. I think it's going to look good. [47:26] I don't think it's going to be that close to your bar for actually putting your name behind something that you would share out in the world. I think it'll mostly just be like – [47:35] Interesting as a hobby is my I don't think it will be bad, though. [47:39] So we're not going to look at... There is some sound to these.

We were having some sound issues outside of... [47:45] So you're not gonna see that. But I basically said, hey, [47:49] We've got some motion graphics. I got a brand here. A bunch of posts of gals from his videos. I want you to analyze these videos. And then I attached it. My SVG of the UX tools brand mark. Again, the golden here is not to release any sort of skill. I'm not going to do that here. But I do want to see if it is able to do this. And so it went in and.

[48:12] That's not the one that I want to see. Let's pull up. [48:15] Thank you. [48:17] Oh, that's interesting that the... [48:19] Sheet. [48:22] Okay. [48:23] There we go. [48:24] Yeah, I want you to actually see here... [48:28] "How it parsed the parts of Gal's projects?" [48:34] Um... [48:38] So this is showing. [48:40] a lot of [48:41] Mine. [48:43] Let's see. Let's see. Let's see. So first it built the skill, did the markdown. And then I said, OK, we also need to feed you some brand guidelines so you know how to incorporate these into the ideas that you come up with.

I did not give it any sort of composition direction. This was this was not a one shot, but pretty close to that idea. [49:02] I really want you to see [49:05] Not my examples. I want you to see how it broke down his videos, but we may have to forego that part of it if I can't figure that out. [49:12] Um, [49:14] Artifacts? [49:15] No. [49:16] Okay. Well, let's look at this, all right? [49:19] Here was version one. [49:21] All right, let's see how this did. [49:23] Thank you. [49:24] Thank you.

[49:25] Thank you. [49:29] Thank you. [49:31] Thank you. [49:32] Bye. [49:33] Thank you. [49:34] Thank you. [49:37] All right. You know, so, uh, uh, so that was interesting. That last part was interesting, actually. Mostly ass, right? Mostly ass.