TINA HE (Pace Capital) - Vibe coding as creative practice

=== TINA HE (Pace Capital) - Vibe coding as creative practice ===

(03:02:49): So excited to have you.

(03:02:51): Tina, you're awesome.

(03:02:52): We've been friends for a while.

(03:02:55): You've been a writer on Every.

(03:02:56): You're doing a lot of vibe coding.

(03:02:58): Thank you so much for joining.

(03:02:59): I am excited to be here.

(03:03:02): For people who don't know you,

(03:03:04): well,

(03:03:04): first of all,

(03:03:04): they should they should follow you on Twitter or on X at fake pixels with no

(03:03:08): vowels.

(03:03:09): But for people who don't know you give us a little bit of a background like what

(03:03:12): are you doing now?

(03:03:12): What have you done in the past and and what are you vibe coding on?

(03:03:17): Totally.

(03:03:18): So I've always been a tinker, I would say, since I was very young.

(03:03:23): I like for the very niche, I would say indie game lovers in the audience.

(03:03:27): I was in love with this tool called RPG Game Maker.

(03:03:32): It's like a very niche Japanese make tool.

(03:03:34): And that's actually how I got into programming, like when I was really young.

(03:03:38): And I've always loved gaming, I think both as a storytelling medium, but also like over time.

(03:03:46): kind of fell in love with just how you could design like a world where you can like

(03:03:52): honestly hang out with your friends and the social aspects of it.

(03:03:57): And the other thing I do, I think outside of potentially being a gamer and like kind of a

(03:04:04): the designer developer of games is that I've been a blogger, like since I was pretty young.

(03:04:09): And so I will say that the main thread of my life is that I'm a,

(03:04:13): like a tinker,

(03:04:14): like perpetual tinker.

(03:04:16): I'm a writer.

(03:04:17): And, and I also fell in love with investing.

(03:04:19): So yeah,

(03:04:21): Basically, my career has been kind of a combination of those things.

(03:04:25): Did product before.

(03:04:28): Started a product studio in like even back when I was in school and then actually

(03:04:32): started my career in venture in 2019 and joined this firm called Pace Capital,

(03:04:38): which is a New York based venture capital firm.

(03:04:41): And then started my own company in the crypto space because crypto is kind of this

(03:04:46): consilience of

(03:04:48): programmability of literally one of the most important concepts like inhumanity,

(03:04:53): which is money.

(03:04:55): And then you have,

(03:04:56): you know,

(03:04:56): this very crazy kind of developer community in crypto for kind of both the,

(03:05:04): you know,

(03:05:04): incentive around the ecosystem,

(03:05:06): but also I think that it's just like a really kind of intellectually stimulating

(03:05:10): concepts at a time.

(03:05:11): So I built basically a developer platform that helps people create smart contracts,

(03:05:15): ended up selling the company to Coinbase and was basically leading developer tools

(03:05:21): at Coinbase's layer two blockchain called Base and basically built a bunch of open

(03:05:27): source tooling.

(03:05:28): for them as well to make building on chain easy.

(03:05:30): And I think I've like kind of taken a lot of those like how do you make kind of

(03:05:36): money building blocks like really fun and easy to use into this kind of new chapter

(03:05:42): of

(03:05:43): obviously agentic capabilities.

(03:05:44): And I feel like as a writer myself and same with you,

(03:05:47): Dan,

(03:05:47): like I think the technology of LLM is just like deeply philosophically interesting

(03:05:53): where there's this kind of interesting dichotomy of like the worst cell and the

(03:05:58): shape rotator discourse on Twitter specifically.

(03:06:01): And we kind of joke that,

(03:06:03): wow,

(03:06:03): this is like,

(03:06:05): you know,

(03:06:05): actually a renaissance of worst cells power.

(03:06:08): Like if you're able to describe something, you can really well into existence.

(03:06:11): Um,

(03:06:12): and I feel like for a lot of writers in my life,

(03:06:14): they,

(03:06:14): they ended up actually becoming pretty good at vibe coding.

(03:06:17): I think for that reason, um, especially if you have some like technical knowledge.

(03:06:21): Uh,

(03:06:22): so I think that like,

(03:06:23): there's a lot of audience and,

(03:06:24): and lineup and guests that you have here that,

(03:06:27): uh,

(03:06:28): you know,

(03:06:28): have probably much better vibe coding skills than I do.

(03:06:31): Um,

(03:06:32): but I really do think that the,

(03:06:34): the kind of concept of like,

(03:06:35): okay,

(03:06:36): vibe coding might still be a form of coding and programming can be somewhat

(03:06:39): intimidating to people.

(03:06:41): So I think that kind of perspectives maybe I can bring today is like,

(03:06:45): it's actually more about the vibe than the coding.

(03:06:49): It's like,

(03:06:50): if you are having fun building what you want to build and,

(03:06:54): you know,

(03:06:54): that's actually the most important motivation.

(03:06:56): And I think a lot of my best ideas literally just come from like a joke or,

(03:07:01): or it comes from,

(03:07:02): you know,

(03:07:04): kind of using a piece of software that was like,

(03:07:06): oh,

(03:07:06): I don't want to pay for that.

(03:07:07): So I just like build it myself.

(03:07:10): I love it.

(03:07:11): You heard it here first, folks.

(03:07:12): It's all about the vibe, not about the coding.

(03:07:16): If you want to be a good vibe coder, it's all about the vibe.

(03:07:18): And I also love that.

(03:07:19): I do agree.

(03:07:20): I think if you're a good writer,

(03:07:22): it translates kind of well to being a good vibe coder or prompter because you're

(03:07:27): just good at casting spells that specify the specific thing you're thinking about,

(03:07:32): which is helpful for writing.

(03:07:34): It's also really helpful for coding.

(03:07:37): On that note, I'd love to see what you're working on.

(03:07:39): What are you building?

(03:07:39): What's your workflow?

(03:07:40): What do you want to share?

(03:07:42): Yeah,

(03:07:43): so I definitely have a lot of workflow that's a little bit more,

(03:07:45): I think on the boring side,

(03:07:47): given my current full-time job is like,

(03:07:48): I focus a lot on developer tools and infra in my investment practice.

(03:07:52): So,

(03:07:52): you know,

(03:07:53): I built kind of these very utilitarian tools,

(03:07:56): like,

(03:07:56): you know,

(03:07:56): kind of GitHub trackers,

(03:07:58): like,

(03:07:58): you know,

(03:07:59): analyzing on like the latest developer trends with LLM,

(03:08:02): with research and happy to share some of that as well.

(03:08:05): But I think my main goal is really to emphasize the fun part of coding.

(03:08:10): i'm going to share my screen okay i will just share the entire screen can you see

(03:08:22): my screen i can all right so um sorry about the

(03:08:32): infinite maze here all right so i basically vibe coded this app um that i've always

(03:08:39): wanted to have in my professional life where uh you know sometimes you get kind of

(03:08:45): impatient with your you know coworkers or people like not merging your prs or you

(03:08:51): know you kind of want to scream you know into the slack

(03:08:54): Slack channel.

(03:08:56): And sometimes you're just really angry.

(03:08:58): And,

(03:09:00): you know,

(03:09:00): and I think I find myself sometimes it's like,

(03:09:03): you build this like auto generated,

(03:09:05): like,

(03:09:05): hey,

(03:09:05): just call me,

(03:09:06): or you mark my email as low priority.

(03:09:09): But then like, why don't you respond?

(03:09:11): You know, or things like, you know, I put it put in PTO three months ago.

(03:09:14): And you're asking if I can reschedule.

(03:09:18): And all of these prompts are actually generated by cloud code.

(03:09:23): My coworkers has been about to retire for six years and does nothing while I covered their work.

(03:09:28): So you can basically be as frustrated as you want and then basically can translate

(03:09:33): what you're about to say that's probably not work safe into a corporate language.

(03:09:39): Okay, there's an issue here.

(03:09:41): Let's see.

(03:09:44): Deep live demos, folks.

(03:09:46): Live demos, baby.

(03:09:47): I love it.

(03:09:48): All right.

(03:09:49): So it was working.

(03:09:50): Let me try to redeploy it.

(03:09:54): How would you fix it?

(03:09:55): Yeah.

(03:09:55): Let's see if we can fix it live.

(03:09:58): Yeah.

(03:10:00): Let's see it.

(03:10:06): Actually, I'm going to go to the deployed version versus local because I know this one works.

(03:10:10): Oh, this is so cool.

(03:10:11): Wait, like, look at this.

(03:10:12): I love this UI.

(03:10:14): Tell us about the UI.

(03:10:15): Oh, my God.

(03:10:16): Look at it.

(03:10:16): It looks like an operating system.

(03:10:17): Very intricate.

(03:10:19): And it looks and there's more to come.

(03:10:21): So let's say that,

(03:10:22): OK,

(03:10:23): you're wanting to do the work of a senior developer by paying me as a junior.

(03:10:26): I'm going to translate this.

(03:10:34): I think it might be my API key, to be honest.

(03:10:36): OK, let me try again.

(03:10:39): Up your limits.

(03:10:42): You're maybe using too much.

(03:10:43): All right, so let's see.

(03:10:47): I've been preparing for this session, so.

(03:10:49): I can give you.

(03:10:53): Yeah, and let me see.

(03:10:53): OK, so.

(03:11:00): I,

(03:11:01): so one thing about vibe coding that's a little annoying is when things break,

(03:11:05): sometimes it's like a little mysterious what happened.

(03:11:07): You have to kind of dive into the code base.

(03:11:09): That's true.

(03:11:10): Yeah.

(03:11:11): I'm in the code base.

(03:11:13): Can you ask Claude to fix it or like ask Claude what's going on?

(03:11:17): So in terms of like workflow, I usually just use Claude code.

(03:11:21): So I'm in this repo and let me see the error message.

(03:11:29): There's one issue here.

(03:11:31): Okay, call stack.

(03:11:41): There you go.

(03:11:42): As easy as that.

(03:11:43): Pasting them into cloud code is next.

(03:11:45): Is basically what, yeah.

(03:11:46): Let's see.

(03:11:55): Yeah, it's yeah, the API call failed.

(03:12:00): So I think it probably is.

(03:12:03): Yeah, my guess was API code.

(03:12:05): So do you think you ran out of anthropocode?

(03:12:07): We can run out of credit.

(03:12:09): Have you stopped anyone in the audience?

(03:12:10): Yeah, why don't you go do that?

(03:12:12): Let me I can I can take over for a bit and give you some space to fix that.

(03:12:16): Does that work?

(03:12:17): Thank you.

(03:12:18): Okay.

(03:12:19): So while Tina's doing that,

(03:12:21): I've like during the,

(03:12:23): during the stream,

(03:12:23): at the beginning of the stream,

(03:12:24): I started demoing an app that I built.

(03:12:26): And so if you just got here, I think it's worth me demoing this for you.

(03:12:30): I'm going to just share my screen.

(03:12:34): So the app is called proof and it's super cool.

(03:12:39): So proof is an agent native Markdown editor.

(03:12:42): And what that means is it's a really simple Markdown editor that I use with cloud

(03:12:45): code and codex that I built.

(03:12:48): From scratch, totally vibe coded over the last two weeks using cloud code and codex.

(03:12:53): It's pretty cool.

(03:12:54): So basically on the left,

(03:12:55): you can see we have a sidebar which tells me what in a document,

(03:12:59): what has been made by human versus what has been made by AI.

(03:13:02): So this is a cloud code generated plan.

(03:13:05): But if I go type in something and say this is from Dan,

(03:13:08): you'll see a little green pops up here and if I type more,

(03:13:12): it'll give me more human percent.

(03:13:15): So it's a really nice way at the base to keep track of what I've written versus

(03:13:21): what Claude or any other agent has written,

(03:13:24): which I think is really helpful in this era.

(03:13:26): It also has all the things you might expect, so I can leave a comment.

(03:13:30): I can do track changes.

(03:13:32): So I can do like this and it'll track my changes, all that kind of stuff.

(03:13:37): But the really cool thing is it has a built-in, it's agent native.

(03:13:43): So anything that I can do in the app,

(03:13:45): agents can do in either an agent that is built in here or any other agent.

(03:13:50): So for example, if I open up,

(03:13:52): My cloud code and I have cloud code open over here and I have a I have a little

(03:13:57): demo that I had to do.

(03:13:58): So this is my agent presence demo slash command.

(03:14:01): So it's just a it's just a slash command that says connect to proof and do some stuff in there.

(03:14:06): And what will happen is which which is really, really cool.

(03:14:10): If we just wait a moment.

(03:14:17): It's it's currently boondoggling.

(03:14:19): It's going to take a second.

(03:14:24): So you'll see there's an there's an agent here.

(03:14:26): My cloud code is in here.

(03:14:27): It's starting to it's starting to make changes and as it as it makes changes,

(03:14:32): it's scrolling me around so I can see the changes that it makes.

(03:14:37): This is all I literally built all this in like in between meetings for the last

(03:14:42): couple of last last two weeks and it's actually quite useful.

(03:14:47): And I think it's a good sign of what's possible right now that you might not really expect.

(03:14:53): Anyway, the following features is a little bit buggy right now, but that's how proof works.

(03:14:59): We will make this available to every readers.

(03:15:02): I'm actually currently working on this today while we're on the stream,

(03:15:06): and I'm trying to get a version out that we'll be able to launch internally

(03:15:09): tomorrow.

(03:15:09): So we'll see if I hit my goal.

(03:15:12): Tina, give me a thumbs up if you're ready to come back on.

(03:15:19): Let me see.

(03:15:20): Oh, she says she can come back.

(03:15:21): Okay.

(03:15:22): Hello.

(03:15:22): Tina, welcome back.

(03:15:24): Thank you.

(03:15:25): So I have to say that I ran out of credits.

(03:15:28): That's exactly what happened because I've been testing this.

(03:15:31): So I just refilled another $20 to Anthropic.

(03:15:36): And I think it might take some time for the credits to refilled.

(03:15:42): But I can demo the rest of my app.

(03:15:44): Please.

(03:15:44): Because that doesn't require an API key.

(03:15:48): Let's see.

(03:15:50): or it or it uses like another part of the part of the sack so yes you can you can

(03:15:56): do that um and then obviously like if you want to have music going on here you can

(03:16:03): play music um while you're not we're not seeing your screen give me a sec okay here

(03:16:07): we go there we go um wait just what is this app what is what is it

(03:16:14): This app is my vision for the future of work.

(03:16:17): Like in the future,

(03:16:18): when you're working,

(03:16:20): you can basically create your own customized workspace for you to do all kinds of

(03:16:25): different random stuff.

(03:16:26): And I was going to basically,

(03:16:28): after my credit get filled,

(03:16:30): buy a couple of new apps to this app store that basically belongs to my own little

(03:16:36): workspace.

(03:16:37): And is this like a website or is it native app?

(03:16:39): This is a website.

(03:16:40): This is a web app.

(03:16:42): Oh, can we go on it?

(03:16:43): Is it live right now?

(03:16:44): Yeah, it's live.

(03:16:45): What's the URL?

(03:16:46): The URL is here, but some of the features will be updated and you can use it already.

(03:16:54): And then some of them need my clock code.

(03:16:58): And how is this built?

(03:16:59): Like what is the underlying technology that does this?

(03:17:02): It looks so good.

(03:17:04): Yeah,

(03:17:04): so I strongly recommend,

(03:17:07): I've been a Vercel and Next.js fan for a long time,

(03:17:10): and I think they definitely are,

(03:17:12): you know,

(03:17:13): a huge inspiration in kind of pioneering the movement of Vibe Coding,

(03:17:18): you know,

(03:17:19): alongside,

(03:17:19): I think that their product v0 is doing a good job.

(03:17:22): But generally, because there's this website actually called Context 7, I will share as well.

(03:17:29): Let me share.

(03:17:43): Can you see my screen?

(03:17:43): We can.

(03:17:46): Okay, perfect.

(03:17:46): Can you see the Contact 7 page?

(03:17:48): Page N. Yes.

(03:17:49): So basically this is a page.

(03:17:51): So Contact 7 is a MCP server.

(03:17:55): What is an MCP is basically, you can think of it as like an API for agents.

(03:18:01): So it basically create an interface for agents to know, to call any external data or tools.

(03:18:08): And basically there's this very cool idea created by an engineer and a team at

(03:18:13): Upstash called Contact 7.

(03:18:15): And basically what it does is that it basically updates,

(03:18:18): you know,

(03:18:18): the tool library for your agents.

(03:18:21): So their agents are basically always able to access the kind of most popular and

(03:18:27): best in class tools.

(03:18:29): So this is actually a good way to see like what stack people are using,

(03:18:33): because it's like I kind of joke internally that tokens don't lie.

(03:18:37): It's like basically I love that.

(03:18:39): Oh, put it up on a banner.

(03:18:41): Let's go.

(03:18:41): Tokens don't lie.

(03:18:44): Like how tokens are being allocated.

(03:18:46): And obviously another great source that I think many of the audience know would be

(03:18:49): like open router.

(03:18:51): Although,

(03:18:51): you know,

(03:18:52): the kind of legitimacy of like the,

(03:18:54): or the actual accuracy of the data is to be debated because you can opt in and out

(03:19:00): as like a app or models.

(03:19:02): So it's still TBD, but you can also see like the amount of tokens that are used.

(03:19:07): So this is like also a great resource for you to basically understand,

(03:19:11): you know,

(03:19:11): the latest,

(03:19:13): the most used models,

(03:19:15): you know,

(03:19:15): week,

(03:19:16): month,

(03:19:17): but also

(03:19:18): I love this.

(03:19:19): It's like you can basically select different use cases like for finance or for role play.

(03:19:25): There are basically different models that are good for that use case.

(03:19:28): And for Contact 7 here,

(03:19:30): like as you can see,

(03:19:30): like the most popular one is Next.js,

(03:19:33): which is what I used to build my app.

(03:19:36): Got it.

(03:19:37): Yes.

(03:19:38): Uh,

(03:19:38): so when in doubt was like,

(03:19:40): oh,

(03:19:40): what,

(03:19:40): like,

(03:19:40): for example,

(03:19:41): you want to do authentication,

(03:19:43): like the most boring,

(03:19:44): but important part of the building an app is like,

(03:19:46): you want users to log into your app and be able to do stuff.

(03:19:50): And traditionally they're really popular libraries like,

(03:19:53): you know,

(03:19:53): clerk is a very popular and like a,

(03:19:56): they're like a very well funded venture backed company.

(03:19:58): You know, they're off zero.

(03:20:00): There, there are many ways to do off now.

(03:20:02): And obviously you can also just use,

(03:20:03): you know,

(03:20:03): the traditional super base or,

(03:20:05): or,

(03:20:06): or fire like Google Firebase to do off.

(03:20:08): But if you want something that's like a little bit nicer,

(03:20:10): it comes with like a prebuilt UI,

(03:20:12): you can use better off.

(03:20:13): And basically like this stack is,

(03:20:16): like shows you or like tailwind for like, you know, making your app look nice.

(03:20:19): Um, it's basically the number one styling, uh, CSS framework to make the website look pretty.

(03:20:26): And,

(03:20:26): uh,

(03:20:27): so you basically like using sites like this,

(03:20:30): you're able to get like a view into the most popular frameworks.

(03:20:34): Can you tell me like, okay, but just looking at this, like going back to your site, um,

(03:20:39): Did you use cloud code for this?

(03:20:41): Did you use codex?

(03:20:42): And what was the, how did you begin?

(03:20:43): Or what did you, how do you even start?

(03:20:45): Like, Hey, I want like a full operating system as a website.

(03:20:47): Like what was your problem?

(03:20:48): What was your process?

(03:20:49): My process was basically like I want a vintage Mac OS.

(03:20:55): I can share like the full prompt with the with if you want to do a follow up like transcript.

(03:21:00): I really do.

(03:21:00): I absolutely do.

(03:21:02): Yeah, I can do it.

(03:21:03): It's basically like a vintage Mac OS operating system and then I want the

(03:21:08): background to look a certain way.

(03:21:09): Like I want the icons to look a certain way and I basically just also give it an

(03:21:14): image of like what that like that era of UI looks like.

(03:21:20): And I think that's also like a really good way,

(03:21:22): like if you want the UI of clock code to look similar to whatever you give them,

(03:21:28): sometimes it's like,

(03:21:29): oh,

(03:21:29): why does it look so different?

(03:21:30): I think a lot of that is actually prompting and also the new feature skills you

(03:21:35): might have seen on,

(03:21:37): you know,

(03:21:37): from the Anthropic release that there's this file called SKILL.md that basically

(03:21:42): describes how to like precisely do a task.

(03:21:48): And it basically creates some level of...

(03:21:50): It's not hard guardrails,

(03:21:51): but it's like a soft kind of in-context guardrail for cloud code or just your cloud

(03:21:59): to understand what you're looking for.

(03:22:01): So if you're a small startup, it's like...

(03:22:05): it will be very efficient to store your ui system like into like a like a startup

(03:22:11): skill uh and then every time you build a site like they would just like refer to

(03:22:16): the style guide of that skill file before they kind of generate any internal apps

(03:22:21): or or any of the products that you want to build and when you're doing this are you

(03:22:26): like uh maybe to start this are you doing a plan mode uh like

(03:22:31): In Claude, how are you iterating until you're satisfied and letting it build?

(03:22:37): Yeah.

(03:22:39): I would say that the biggest thing in...

(03:22:44): kind of creating really good outcome.

(03:22:47): And usually with one-shot success, it's just really well-written PRDs.

(03:22:53): And I would say that even a year ago,

(03:22:54): you need to put a lot more time in drafting the product spec.

(03:23:00): And I think that's probably music to the ears of the product managers.

(03:23:05): in in our world where um they're very used to writing these type of products back

(03:23:09): um but the more specific you are like obviously the better performing um the

(03:23:14): outcome is going to be uh so generally i would say like the the more detailed you

(03:23:19): are the better uh but at the same time like if you want to do something like for

(03:23:23): example i think a lot of apps here like i have a couple funny ones like for example

(03:23:27): you can like build a trading app

(03:23:30): and you can do like 2x leverage and then put all my network.

(03:23:34): Is this hooked up to your Charles Schwab, right?

(03:23:36): This is not.

(03:23:38): For demo purposes,

(03:23:39): but if you do want to hook it up to either your crypto wallet or your real trading

(03:23:46): account,

(03:23:47): definitely you need to use an API for that.

(03:23:50): But this is just for demo purpose of what a personalized UI looks like if you just

(03:23:55): want to control,

(03:23:56): you want to monitor the situation while you're at work in the future.

(03:24:00): And I wanted to demo some like image related features.

(03:24:03): So you can like take a photo and then you can enhance it.

(03:24:07): So here I'm using the replicate AI and I wanted to showcase beyond text,

(03:24:12): like,

(03:24:13): and then you see it generates like a corporate background for me and I can download

(03:24:17): this image.

(03:24:19): And...

(03:24:22): I love it.

(03:24:23): It's very silly.

(03:24:24): And then there you can, I can also do a side hustle where I can do like data labeling task.

(03:24:31): Like this is a dog, this is a tree.

(03:24:34): Just imagine, you know, I'm like also.

(03:24:36): This is the new economy.

(03:24:37): This is the AI economy.

(03:24:39): This is the AI economy, right?

(03:24:41): And then I was planning to also vibe code live, potentially an app for Valentine's Day.

(03:24:48): We're like, wow, I'm doing all of that at work.

(03:24:50): I can also just vibe code a bunch of Valentine's Day cards to my friends and my loved ones.

(03:24:57): But I know that we're a little running out of time.

(03:24:58): I mean, we've got four minutes.

(03:25:00): What can you do with four minutes?

(03:25:02): All right, let's try it.

(03:25:02): Let's try it.

(03:25:05): I'm going to...

(03:25:09): Oh, I'm not sharing all of my screen now.

(03:25:11): Let me try that.

(03:25:20): All right, so let's say create another app on the desktop.

(03:25:28): Also, I know that I saw Nat's session where he was using a voice to text.

(03:25:33): He was using Monologue by Every, which you should definitely be using.

(03:25:36): I don't know if you're using Whisperflow, boo, but it's okay.

(03:25:40): I'm using SuperWhisper.

(03:25:42): SuperWhisper, also boo, but it's fine.

(03:25:43): Also boo.

(03:25:45): First to market.

(03:25:46): I just like it kind of works.

(03:25:48): So I just have been sticking with it.

(03:25:50): But for the purpose of time,

(03:25:52): I'm just going to do create another app on the desktop that creates beautiful

(03:25:57): Valentine Day cards to send to your loved ones.

(03:26:02): All right, let's see how it does.

(03:26:04): I love it.

(03:26:06): I love it.

(03:26:06): While it's while it's cooking,

(03:26:09): I'm I just like love how creative and like whimsical and thoughtful all of this is.

(03:26:16): And it feels like you're playing in this amazing way.

(03:26:21): Tell me about that,

(03:26:22): because I think a lot of people approach AI from like,

(03:26:24): I need to get something done and it needs to be valuable.

(03:26:28): And this is obviously quite valuable,

(03:26:31): but it also just has this playful part of it that I think is actually so important

(03:26:37): as a way to approach new technologies,

(03:26:39): in particular AI.

(03:26:40): Tell me about that.

(03:26:41): Yeah, I would say this is less definitely not useful.

(03:26:46): And I'm keenly aware of that.

(03:26:48): And I think I'm trying to take a slightly alternative approach with this session

(03:26:54): where I obviously use AI.

(03:26:56): I think 80% of it is actually for work, like maybe 90% of it.

(03:27:01): I use a lot of the features that I think other guests on the show probably also have shared.

(03:27:08): everything from research to operating or automating my own day-to-day life.

(03:27:14): But I feel like those are going to get even easier to do.

(03:27:18): I feel like with the progress that Anthropic and ChatGPT is making,

(03:27:23): there's going to be less and less

(03:27:25): So I was like a new better lesson.

(03:27:26): There's going to be less and less like handholding on the agent and on the AI to do

(03:27:31): specific tasks.

(03:27:33): In the future,

(03:27:33): we can really just describe what we want to do in natural language,

(03:27:37): especially for personal use case.

(03:27:39): So we need to get pretty good results if we don't have to think about scalability.

(03:27:44): And then the 10% of it,

(03:27:46): you know,

(03:27:46): is actually the meaningfully lowered cost to create joy and fun.

(03:27:50): And I think that that's why in the intro, I kind of like emphasize that gaming aspect.

(03:27:54): I feel like this is my childhood dream to be able to just like,

(03:27:58): create fun experiences that make people laugh and make people have fun and enjoy a

(03:28:04): kind of life a little bit more.

(03:28:05): And given the cost is literally lower than probably buying someone a cup of coffee,

(03:28:11): like definitely lower than buying someone a cup of coffee.

(03:28:14): Like it can just bring a smile on their face if you build an app for someone.

(03:28:18): So I think it's just like a new way of self-expression and a new way of also,

(03:28:23): you know,

(03:28:24): socializing that I think will become more and more common.

(03:28:27): And there are like tools that kind of,

(03:28:29): you know,

(03:28:29): clock code,

(03:28:30): it still feels a little intimidating.

(03:28:32): You need a terminal,

(03:28:33): although there are like tools like,

(03:28:34): you know,

(03:28:34): conductor and whatnot,

(03:28:35): that's like creating a better UI,

(03:28:37): but it's still a little intimidating.

(03:28:38): And there are apps out there that are like, you know, lovable or Wabi.

(03:28:42): There are more and more companies out there that are basically removing even like

(03:28:47): seeing any line of code at all.

(03:28:50): And in the future,

(03:28:51): it's like,

(03:28:51): you know,

(03:28:52): you can kind of imagine a future where people are,

(03:28:53): you know,

(03:28:54): sharing with loved ones this funny app that they make.

(03:28:58): Like, what does that content look like?

(03:28:59): It's probably going to look even like way more creative than this little experiment that I did.

(03:29:04): I totally, totally agree.

(03:29:06): I mean,

(03:29:07): the reason I have the same thing,

(03:29:08): like I got into programming because I wanted to make games and then games were like

(03:29:13): too hard to make.

(03:29:13): So I ended up making B2B SaaS.

(03:29:16): And I feel like now it's possible to just do the games and do the like really fun,

(03:29:21): creative projects.

(03:29:22): So thank you for re-inspiring me, re-inspiring all of us.

(03:29:26): If people want to find you, where can they find you?

(03:29:30): Um, they can find me on, uh, yes.

(03:29:33): So my Twitter handle is in below like fake pixels,

(03:29:36): but my,

(03:29:36): I also have a newsletter called fake pixels.substack.com.

(03:29:41): And,

(03:29:42): uh,

(03:29:42): yeah,

(03:29:42): that's where I share more of my long form writing about usually the philosophy of

(03:29:46): tech and investing and AI and all of the above.

(03:29:50): Amazing.

(03:29:51): And on every, you've got a couple of great essays on every, we'll, we'll put them in the chat.

(03:29:55): Um, thank you so much, Tina.

(03:29:57): Um, yeah.

(03:29:58): See you soon.

(03:30:02): All right, so we've got a stacked lineup coming up.

(03:30:05): We've got a couple of really cool people just off stage, just getting ready to get started.

(03:30:11): If you just got here, this is Vibe Code Camp.

(03:30:14): We are live all day.

(03:30:15): We've been live for almost four hours now with some of the top coders in the world

(03:30:20): showing us their workflows,

(03:30:21): how they build stuff with AI.

(03:30:22): We just saw Tina He showing her personal operating system website.

(03:30:26): It's like one of the coolest things I've ever seen.

(03:30:29): And if you're here and you want to see the agenda, go to every.to slash agenda.

(03:30:35): You'll see everybody that we've got coming up.

(03:30:39): And you should check out Every.

(03:30:41): Every is the only subscription you need to stay at the edge of AI.

(03:30:44): If you subscribe, you get access to ideas, apps, and training.

(03:30:47): It's a bundle of all the things you're going to need to stay at the edge.

(03:30:51): And with that...

(03:30:54): I'm very excited to welcome to the stage, Paula Doja.