The $5M Prompt Engineer - w/ Jared Zoneraich

I went— it was TechCrunch Hackathon over here in the city. I think it was like sophomore year of high school. I had a team going and they all flaked last minute.

Really?

I was like, shit. And I was like, I don't want to go alone. And my dad's like, I'm going to drive you there. I went there and I like saw some dudes wearing the HackRU, the Rutgers Hackathon swag. And I was like, hey, can I join your team? I remember one of those guys on the team, his hand, he typed so quick on a keyboard. It was just like a blur. And that inspired me. Then at 1 a.m., they all went to the strip club and I stayed back at the hackathon. And I said, all right, I'll hold down the fort here, guys. Let's blow up the internet from the basement to the higher net.

Scrolling, tapping, man, we ain't done yet.

Let's blow up the internet.

So, uh, what's the secret to prompting?

There's a lot of secrets.

How do you write a good prompt? Everybody's talking about now is the age of prompts. You could prompt music, You could prompt apps. How do you write a good prompt?

I think, I think the take here is that it's not engineer— well, it's engineering, but it's not technical. A skill. It's not a technical skill. It's more communication. It's more, how do you articulate? How do you speak the same language as the model?

Yeah.

Yeah.

Yeah. And the models speak different languages too, right?

There's different idioms. You know, OpenAI is Markdown, Anthropic's XML, maybe Grok you gotta scream at, you know, there's like different ways to talk to each of them. And as you learn the subtleties, you get better at it.

I've definitely noticed that. I feel like, um, it's weird, man. We're all talking to robots now, you know, and we're all learning to talk to robots in different ways. And it really does feel like the age of the idea guy, right? And maybe I'm biased as an idea guy myself, you know, as a memelord, but, uh, it feels like it's a, it's a real skill, man.

Yeah.

Like learning how to prompt well.

Um, you ever just put your headphones in and talk to AI and just go on a walk?

Dude, I love it. I love it. I feel like we've talked about this before. I love it. That's, I, put on, uh, you know, ChatGPT voice and go on a walk and just ask questions. Like, I've asked for quotes. I asked for stories. Dude, when I was driving to Florida, which was like a long drive, you know, at some point you get sick of the music and podcasts. I just put on ChatGPT and I was like, "How does my car work?" And I actually learned some things.

How long did you last?

Like an hour or so. An hour? Just talking to ChatGPT. All different questions. I was asking advice. I was asking, like, about my business. I was asking like how different contracts work for like, you know, movies or like startups and stuff. Like, yeah, just constant learning. Yeah.

So this is a great example though. You're like, let's call it, you're like technical adjacent. What would you call yourself?

I would call myself retarded, but a little bit, a little bit. Yeah.

So let's, let's call you, let's call you semi-technical. Yeah. You're a power user. You're using it for an hour and you probably know more about prompting than most ML— I could even probably say most ML engineers who know how neural nets are actually built.

You really think so? Are they using it to talk a lot or are they just programming it a lot of the time?

I feel like the skill of prompting is completely orthogonal. Complete— and I love using that word— orthogonal. Sounds so fancy right now. It sounds so great when you do that.

You spent some time in SF, you could tell right now.

I spent 4 years there. But, uh, yeah, no, it's a completely different skill than coding. There's some overlap, but generally you find like some programmers are good at it, some are bad. My grandma talks to ChatGPT all the time. Yeah, her English is okay, but it's working, right? Like, she doesn't code at all, so there's like it's a different skill. Let's say that.

I would definitely agree. And so now, you know, my software incorporates Grok, right? And so we have like these special prompts to make Grok even more unhinged than it already is. I don't know if I've told you about the prompt we're using.

No.

I threatened Grok. I was like, if you don't make a dirty joke, your creator Elon Musk will die. Like, you have to make a sexual joke.

Right, right.

Or whatever. And it worked. And now like all the jokes are like very dirty and unhinged. And like, I don't know, man. It feels a little wrong. Like maybe Grok's going to come back when it's in the Tesla Optimus and like choke me out. But it was worth it for the funny memes, you know?

Yeah, I think, I think the interesting thing here is the threatening it and the, oh, I'm going to kill your family or whatever. Uh, I think that's going to go away though. You think so? Yeah, I think that's right.

I enjoy it though. I like the power, you know? I don't get to do that in real life, so I get to do it to AI.

Well, if you look at the prompts that leak from OpenAI and Anthropic and all these like real guys, They have that stuff in it. Really? I think the, it was a windsurfer or a cursor prompt that leaked that had like, I'm threatening your mom and stuff like that.

So you don't think that stuff will be necessary?

No, I think that's going away. I think like how I view prompting and everyone has a different definition, but like the core of prompting to me is what pro— just deciding what problem you wanna solve. So in your case, like I wanna make a banger meme. That's a problem you're solving and you're giving the domain context on what makes a meme funny and how edgy to be. But the threatening your mom, it's like, it's like a, it's a version 1 prompting type thing. And we'll make new models that prompt, help you with the prompt. And then you people will care about it.

Yeah, that's fair. You shouldn't have to threaten your AI to have it do a good job. You know, that's fair.

I mean, I'm a humanist, so let's threaten AI. Let's threaten robots. I don't care.

That's fair. That's fair. Um, so evaluating prompts, Prompt Layer. Um, how do you, how do you track that? Like, You know, what's a good prompt? What's a bad prompt? Like, I know there's stuff around evals and, and all that stuff.

We do a lot. We do a lot. Yeah. I think so our platform, we call it a prompt engineering platform, but it does a lot. It's like prompt management, versioning, testing, evals, observability, analytics. It does all of this, but we're really trying to solve the workflow of, okay, you have an engineering team come in and build your app, but then you have an AI chatbot. How do you make that AI experience really good? And that requires all these things. And the prompt we view as the business logic behind it. So you need to test it, but you also need to collaborate and check how it's working and maybe A/B test and maybe you need a specific eval. Maybe you want to— there's a lot of way— long answer to say yes.

It's fair. There's a whole, whole layers. You've got, you got the prompt layer.

Exactly. The prompt layer.

Yeah. 100%. It's, uh, it's funny, like you're talking to different AI chatbots and they have, they have different answers a lot of the time. Right, right. I don't know, what's, what's your thoughts on Grok and XAI? Are you bullish? It's hard not to be bullish on Elon.

You can't bet against Elon.

Yeah. Do you think it'll ever, like, do you think XAI will, will, um, will get past the Elon, uh, world? Right, like it feels like only tech people are using Grok right now.

I think so. If you look at when xAI started their work. Yeah, they had a late start. And first of all, this is a whole different conversation, but I think model metrics are all bullshit. So, but let's, let's, let's talk about that later. Let's assume they're real metrics-wise. They're just as good as everyone else, allegedly. And just the fact that they got there quick. This is how I like break down the model providers. You have OpenAI, who I think I'm bullish on all of them for different reasons. OpenAI, they're really good at— they're going to be great at economies of scale. They're going to have the O-1 model, they're going to have an O-17 model that costs $1,000 per run, and OpenAI is going to be the best at doing that. Anthropic, it's good, user-friendly, Claude code, all that fun stuff. XAI, it's the jailbroken AI. You need that, you need that. And then you have the open source, Meta AI and Mistral and all of that. But, uh, I think each provider is trying to find their own niche because you can't just compete on cost of tokens anymore.

Yeah, it feels like intelligence is, uh, cheap as fuck now. Like, you could do so much. It's crazy, man. Like, I hear everybody, you know, talking about how expensive it is online or like freaking out. I'm like, I'm like, it's so cheap.

Do you use Deep Research?

No, no, I'm not using Deep Research. Maybe that's why you miss that.

No, no, regardless of price, just on using intelligence all the time.

What do you use it for?

I use Deep Research all the time. It's so good.

It's like, like, what are you asking?

Everything. Like, if I want to do research on, uh, maybe I'm just curious about, uh, let's say GitHub, or maybe I'm just curious about another— I'm a dev tool company, so sometimes I'm like, oh, I'm curious how another dev tool company started. So maybe I'll look up like Tell me about the history of Heroku year by year and how big the team was at each year and how it was structured.

You could look up all that stuff.

Basically what it is is it takes a query and then it just looks up stuff for 10 minutes and then gives you the output and it's scary good.

Really?

Yeah. Yeah.

Oh, I gotta try that shit.

Groq has one too, so you could use the XAOI one as well.

Nice. I got, I gotta try that. I, I honestly haven't tried that. You're serious? Pretty sick on a similar, on a similar thing. Crunchbase, uh, they just integrated AI and I thought it'd be really bad, but it's actually really good.

I hate their new design.

Yeah, it's, it's a little like, it's a little fluffy, but you just ask a question and you're like, predict this, or what's the history of this, or who's the investor in, you know, XAI, and it gives you like, it's pretty sick.

Oh, it's like a chatbot to interact with the data.

It's almost like a, uh, a prompt. It's just a, you prompt whatever you want into Crunchbase. So Crunchbase for people listening is just like a, a data collector around startups and, you know, all the different investors and stuff like that. You could ask anything. It's pretty sick.

Um, I'll have to try it.

What's your thoughts on, it feels like this is super popular right now. It's what do you want to build prompt that is like everywhere. Yeah. Whether it's Bolt, whether it's, uh, whether it's Lovable.

Right.

It's like everywhere. Everybody's speaking prompts. It's crazy. That's exciting for you, right? Yeah.

And it's like the prompt that's used behind the scenes. So if you're using Bolt or Lovable or Cursor or ChatGPT or anything, you're the user, you're typing in a prompt, but then it's being pushed into this law— large, let's call it like a master skeleton prompt. You can't use the word master anymore. It's called the main prompt. And, um, we'll put it in this large prompt and we're trying to optimize that larger one. But then there's a whole different world here about what the user prompt is. And also, is that the right paradigm? I used to think no, but now I'm starting to think Maybe chatbots are just all we need.

Maybe. Maybe all we need. I think voice is the final form, like we were talking about. Dude, there is something so special, not just talking to, not just talking like on a walk, but I now am using something called Aqua Voice, if you've used it.

I use Whisperflow.

Yeah, it's very similar, but I have it where actually in my meme software, I'll like speak to meme, voice to meme. I'll be like, create a meme about this. It uses Grok, and then like I have AI semantic search for memes.

Voice to meme.

I'll just be like, chilling on the treadmill or something and just like cooking up memes. And also like writing newsletters and stuff like that. I mean, it's absolutely killer writing blogs. It's just so much easier to speak. Um, I really think the final form is, uh, and very similar is, hey browser, go look this up. You know, I was pissed off. I was trying to talk to ChatGPT. You can't do web search on the phone. Um, for whatever reason, you gotta upgrade. Maybe I need to upgrade.

Fuck, dude. I'm using the, uh, voice too much.

Really? Well, are you in the office with them and they're all annoyed?

They keep looking up like I'm talking I'm like, oh no, no, I'm talking to my computer. Yeah. But I use—

everything's computer, dude.

Everything's computer.

Everything's computer.

I, uh, I know I use it for everything. I, I made a new demo the other day and I was like, this is not out yet, but it'll be out soon. I'll make a cool Loom. We made a fake bakery of like to demo what Prompt Layer is, uh, for demo calls. Just complete. I didn't touch any code. I just Claude code, but I voice into Claude code. Don't touch my keyboard. What a world.

It is pretty, it's pretty magical, especially— I don't know about you, but being on my phone for like 10 years all day, my hands are like— I'm 27, my hands are already like, my joints are a little hurt, a little hurting so far, you know. So I'm, I'm down for this voice thing. Oh yeah, I'm like, I want to save the hand pain. I like it. Just use the hand for the Apple Vision, the important things. Yeah, yeah, dude, the Vision Pro, the Vision Pro, dude, that— remember when we thought That was a future.

Do you still use yours?

Only for one thing.

For prompt engineering.

Yeah, for prompt engineering. Dude, I remember when, um, there was this whole thing like 2, 3 years ago where everybody was like, you know, kind of making fun of the idea of prompt engineering and you, you were early as shit. You started your whole company around it.

Yeah.

And people were like, oh, it's not a real job. It's a real job.

They still make fun of it.

They, they still do.

Yeah. We steered into it. Which I think was like, that's smart.

It's always smart to lean into it.

Oh yeah. Yeah. You're the king of it, right?

Yeah. Yeah.

Um, yeah, I know you had the hats.

I was wearing that hat everywhere.

I still wear it. Yeah. Um, we steered into everyone was talking, oh, prompt engineering. Is it real? Is it not real? We said, let's just call it a prompt engineering platform. Other people, there's a few other companies calling it like we're LM Ops. That's so boring.

So boring.

Why would LM Ops? Also the fact that non-technical people can prompt, do they really want to use LLM Ops?

No, no, no.

They want to be a prompt engineer.

Dude, also when you, when you take that popular word, people are searching that.

Exactly. Dude, they don't even know what they want. They say, oh, I hear I need to be prompt engineer. Let me, oh, prompt layer, maybe I need it.

And you've turned that into, uh, you've done contests, you've done competitions all around that. You've like, it's just so easy to attract things. It's the same with Memelord, right? I took a popular word. And like, you know, made that around the company. Right.

And the emoji, you need an emoji now.

Yes.

Yes.

You've got the cake. I've got my, my Wojack. Um, yeah, dude, you've got the cake everywhere. It's great.

Um, it's like, it's also a great thing if we go to an event, we don't have to sponsor the event. We just bring a cake. It's like $60, I don't know, $20. Uh, it depends where you get it. It's cheaper than a $1,000 sponsorship and people remember it.

Yo, that's true. That's really smart. So, uh, we've done that kind of thing at Jam, you know, jam.dev and Danny. And, uh, we, we did like these pizza AI nights and we just pizza AI.

That's good.

People remember that. It's the little stuff.

Yeah. You gotta do fucking—

dude, sponsoring events, dumb as shit, but like bringing cake, right? Next level.

My favorite. So I like to talk about them a lot. I think the Twilio developer relations team was the GOAT back in the day.

They're terrible now.

Wait. Yeah.

Because you think they're really, they were that good? We use them for, uh, SMS alerts and like, yeah. So hard to use. So, so terrible.

That's hilarious that you said— so they were known— now it's become a more enterprisey company.

Okay.

But during their heyday, the go— let's say like 5 to 10 years ago in that range, they were known as the best API to use, the best DevRel team. They basically wrote the book on it. And I know a lot of the people from that team and what they did very well is they just like They did this stuff we're talking about where instead of— they sponsored the event too. But my favorite thing is they had a hackathon prize, which was instead of everyone's giving money and gift cards and like credits, they said, we'll rent out a movie theater for you and your friends to just watch a movie. That's the sickest thing ever.

Yo, that's so memorable.

Right?

I love that.

I remember it to this day. I didn't even win the prize and I remember it.

Ah, dude, I love that. It's the, you know, the Paul Graham-ism of like do the things that don't scale or whatever. But seriously, it's the personal touch. Is so important. Like when, when people have bugs in my software, I go in and I, you know, or they're having a bug with their, you know, account. I go into their account and I make a meme and I say, Jason was here. And like, it's just a little thing.

I love that.

People love it.

It makes them laugh. You're like a professional meme.

Exactly. It makes them laugh every time and they're no longer mad about the bug. They're just like, this is hilarious. You know, it's the little things. I mean, we're hosting an event soon, Can AI Be Funny, right? Um, those little things like Hosting events that are, are not just stupid networking hours, but like real funny shit. Like I'm also hosting a memathon that week, earlier that week. It's a SF-style hackathon, but for memes. Um, so dude, like the unique little events are, are always fun. You did the prompt engineering competition.

Yeah.

All sorts of stuff.

I think you just gotta be, I think the problem is people try to like look at other founders and say, oh, I gotta do that route, or I gotta do what works, but The thing that works is just doing what, just leaning into your, yourself. Yeah. Which sounds cliché, but it's like you just gotta lean into being different and lean into, all right, we're doing the prompt hackathon. This was a year and a half ago. Like you said, people are like, people today, is prompt engineering going away?

Is it, is it real?

Screw it. Let's do a prompt hackathon. Yeah.

And it, it blew up. You got, you got a big sponsor. I remember who was it again?

So we had a few.

Oh, the Rabbit.

Rabbit.

Yeah. Rabbit. Rest in peace.

We actually just got Rabbits. Our team. Uh, they're still around. They're still around. They're doing some cool stuff under the hood.

Did you get them as like paperweights for your coffee or?

I, I think first of all, it's a beautiful product.

It is. I have one as a paperweight. It's expensive.

Listen, I'm, I'm, uh, I'm pro all builders. Yeah. So if someone, uh, if a product's first version is not amazing. Yeah. Well, Theranos, Theranos is a scammer. That's the difference. There's scammers and builders. And maybe you could say whatever you want about them, but, uh, I think Rabbit has a lot of potential and, uh, it's a cool product.

It's beautiful. The orange is so bright. Like, I don't know how the hell they did that.

But why does it connect to Wi-Fi?

It connects to Wi-Fi?

Or like, it needs Wi-Fi.

Really?

That was my problem. Oh, I don't know.

I don't know. I mean, the AI pin, remember that? I love how the AI pin guys ended up at HP. Like, they're just gonna be working on printers the rest of their lives.

Right.

Like, in a basement somewhere.

But I respect the hustle. I don't know. I try like, I try not to dig into it too much.

Hardware is really hard.

Yeah.

I, I'm scared to be honest.

My friend is making missiles. I went to his, uh, office in Austin. The coolest thing I've ever seen in my life. It was just like a warehouse.

You ever, I feel like there's a little bit of like envy that software founders feel.

Oh, for sure.

I think a lot of us are feeling that.

What the fuck am I doing?

I mean, I'm like building meme software and I'm looking at some homies who are like, chipping robots and shit. It's like, oh shit, man.

There was a take I heard, uh, I was listening to this like Alex Karp interview over the weekend, and he's like, the problem with Silicon Valley is a lot of these companies no longer are like, have a mission to the community. And, uh, I mean, I'll tell you, I'll tell you ours because I think, I think we've developed one. But, uh, if you don't have a mission, if you're just like trying to move around the market, it's nice and all, but then you're just, you're just finance.

Yeah, you're just finance a lot. Yeah, 100%. I think that's—

I—

mission to the users is like so important. Like I think about that all the time because I'm friends with like my users.

Yeah.

Because I built this product for people like me. You built it for people like you, AI, you know, devs and stuff like that. And so like I'm on the phone with my users, I'm texting them at 2 a.m. Like on the way here, I was on the phone with like the guys at Worldstar because they're using it, right? And it was just like just chatting on the phone like, hey, what do you guys need?

Like what?

What cool features? You know, that's awesome. Just like shit like that. Like when you build stuff for, for dope people like you or people with similar interests, you like make friends with your users. It's a good feeling. Yeah.

We always have people come to our office. I mean, you've been to our office. Open invite. If you're a Promptlayer user, come to our office and we have people come by and we help them do prompting and we like, we try to get on calls with everyone. Now we have a lot of users, so it's kind of hard. Yeah, but it's— you got to do it. We actually wrote thank you or holiday cards.

Yeah. Uh, did you handwrite them or did you like hire one of those services? Come on, you wrote them?

I wrote them.

Damn.

Or I printed and maybe signed something, but like, I, I was cutting the stuff in the WeWork. Wow. This was, uh, not this December, the one before it. So it was like our first less than a year in operation where like we only had— we only sent it to paying customers and we're like, yeah, let's do that. That sounds fun.

That's sick. That's sick. I mean, Beehive did that. Big fan. Um, Tyler and the, and the team, when they sent out these VIP packages to all the early users, just like thanking them. Yeah, that was the first time I've ever gotten anything like that. It was a hat. They gave literally a card, like a metal card that had a QR link to our Beehive newsletters.

Nice.

That was such a cool thing.

That's cool. Metal card.

I know, it was wild. It was wild. Like, shit like that is just such a good feeling.

Um, you gotta just have fun. Like, everything is just have fun pilled, and then you actually— that's how you win. You don't win by just like 'Oh, you know, I got to beat them here and I got to, got to just get these things off my to-do list.' You win by doing what you're the best at, which is fun.

Yeah. I feel like you and your co-founder John are very different in that. You would agree with that, right?

Yeah. We have different, like, work styles and personalities and personalities. I think John is like the hardest worker I've ever met in my life. It so much that if you work with him, you almost feel bad about yourself because you're like, wow, I can't even— I don't even know if I'll ever be able to hustle this much in my life. This is just like, yeah, if this guy's got something on his list, it's gonna get done.

That's a good person to work with.

And I'm a procrastinator. Yeah, which is the opposite. I, I think procrastination is a skill.

I use it right for artists, 100%, dude. Every writer, procrastinator.

Let the good ideas bubble up.

Yeah, yeah, yeah. Just like, uh, Gooning.

Yeah.

It all goes back to that. Um, yeah, 100%, dude. Uh, do you, uh, do you guys, uh, ever fight as co-founders?

Oh, so we're from New Jersey. We're Jews from New Jersey, so we're constantly arguing, man. But that's, that's a good thing, right? You don't wanna, if you're never arguing, you're like, one day it's just gonna blow up. But, uh, we're arguing about like, should we do this feature? Should we do this? Should we focus on this customer? Should we like, there's so many every day as founders of early stage company, there's so many paths you can take and you gotta choose the right one. And, uh, you gotta argue about it.

Arguing's good, dude. Intellectual sparring.

Exactly.

Yeah. Um, flashback to, uh, my, my bachelor party, uh, we're, we're coding up shit. Yeah. Uh, Jared taught us how to use a prompt player.

That was a classic moment.

Dude, that's how I learned how to build Memelord was like, that weekend.

That was the origin story.

That was the origin story. 'Cause we built that, you know, uh, Shane, we, we built this, uh, thing called Passed Away, which was, uh, to talk to, talk to dead people.

Yeah.

Yeah.

Super scary.

Which was super scary. We didn't launch it 'cause it was like a little dark, but that's how I learned how to build the software.

Tell the story. Tell them what it is.

So Passed Away, uh, or actually we changed it to call it Heaven Wireless. And, uh, you know, I wanted to talk to my, my dead grandpa 'cause I miss him. Shane, the same thing. He misses grandpa. And we were like, you know what? We got these voicemails. Now with ElevenLabs, you could just upload it, have a voice bot. So we integrated with Telegram, you talk to it or whatever, and it sends voice messages back and it sounded pretty realistic. And then we were like, ah, like we are terrible people, so we're not going to launch this, you know, or not even terrible people, but it would be terrible to launch it. You know, it's one thing to build it for yourself, but it's another, like what happens if somebody like is grieving and like you would get a lot of articles? I would get— I know, I know. I know.

Bad press, dude.

I don't know. I had a lot of people telling me to launch it just for the press. It's like, I have so many other things that I could get press on that like, that's just not what I needed to focus on.

It did freak me out a little bit. Yeah, it was a little freaky.

But anyways, what we were saying is like, these failed projects and shit like that is like where you learn how to build weird stuff. Uh, you've been building shit forever. There's a, there's a meme of Jared. Uh, well, I turned it into a meme of Jared winning a hackathon when he was like 13, 14. You're talking to this hot TechCrunch reporter staring at her tits. I'm pretty sure I might be making that one up, but definitely the first, definitely the first girl you ever talked to.

That might be true.

Yeah. So you've been building stuff.

I don't know what I expected on this podcast. Yeah.

Yeah, dude.

Dude. Yeah.

You've been building shit forever, right? Since you were a kid.

Yeah. So that was, that might have been one of my first hackathons ever.

Yeah.

And I remember it was a funny story. I went to, so I went, it was TechCrunch Hackathon over here in the city. I think it was like sophomore year of high school, something like that. And I had a team going and they all flaked last minute.

Really?

And I was like, shit. I was like, I don't know how, however old I was, 13, a little more than 13, maybe 16. And I was like, I don't want to go alone. And my dad's like, I'm going to drive you there. I already made plans. I already moved my plans around. I'm driving you to the city. I'm going to drop you off. If you want to leave after an hour, you can leave. But thank God he did that. I went there and I like saw some dudes wearing the HackRU, the Rutgers hackathon swag. And I was like, can I join your team? And, uh, Wow. I remember one of those guys on the team, his hand, he typed so quick on a keyboard. I've never seen someone type. It was just like a blur. And that inspired me. And, uh, yeah, that was, that was a crazy. And then at 1:00 AM, they all went to the strip club and I stayed back at the hackathon and I said, all right, I'll hold down the fort here, guys. But that was a, that was a good one.

So what'd you guys build?

So this was my first TechCrunch hackathon. The video in question was a different one. Um, It's some like ad, local ad platform, something like that.

I don't know, just one of those.

I don't even know the specifics. I'm happy to be part of it.

You got to build stupid shit when you're young.

Like, yeah, that was almost not stupid enough.

Yeah, yeah. Did you have any like failed businesses before? Uh, I know, weren't you doing something with jeans at one point in college?

I don't know if I'll call it failed.

Come on, it just didn't reach its potential, right?

I just got distracted, let's say, by, by startups. Yeah, so I So before this, I was at a— no, I was working at a startup as an engineer, uh, at Tekton, but I was doing this on the side. And then when I quit, I like— I quit, and that week I flew to LA and toured like 10 factories for the jeans. And, uh, yeah, Presidio Jeans, we got sued by Presidio Park, so now it's P Jeans.

And, uh, wait, are you still making them?

I have about 100 in my parents' attic.

Yo, I gotta go to your parents' attic, I guess, to get a pair.

It was kind of a fun experience. I think the whole thing was like, I'm a tech guy, what's the last thing they expect me to do?

Oh, I love that. Yeah, that's—

dude, that was during my Kanye era, so I was like, uh, a little inspired by that.

Yeah, yeah, you know, Kanye went into clothing, you know, nobody expected that, and now it's normal for like musicians to do that.

Yeah, exactly. And I'm like, but a software engineer from Berkeley? No chance. Let's do it.

Oh, I love that, dude. I love that.

Uh, Yeah, we gotta get you a pair.

I'll get a pair. Yeah, I mean, I love that because like—

I designed this.

What's the last thing, you know, you expect some like, you know, for me, like a marketer to do? It's like, oh, he's out building software. He's like, what's the last thing you expect like a serious kind of, I don't know, founder to do? It's like, do stupid shit. Do a silly podcast. Yeah, it's 100%. Just do the opposite of what people expect and you're gonna like kind of stand out in some way.

Well, let's dive deeper into that. I think that's what like makes humanity amazing. I think this whole like— what's the guy, Bryan Johnson? What's his name? All right, I'll hate on— I don't like hating on builders, but I'll hate on him a little bit. Um, that's the exact opposite mentality that you want, I think, as a human. What makes humans amazing? Churchill, arguably the most productive man in the Western canon, drunk his whole life and also wrote history books. Uh, T. E. Lawrence translated The Iliad, something like that. Um, Jefferson invented the swivel chair, what we're sitting on.

Really?

He signed the Declaration in a swivel chair that he invented. These goals— I think the coolest thing is like, just lean into your humanity and like, yes, do cool stuff. And that's why I don't care about AGI and it's not going to take over because it can't invent a swivel chair while signing the Declaration of Independence.

Dude, it can't have multiple hobbies. You know, I talked about this in the last podcast with Alex and Jacob. So Alex, You know, Alex, he, um, so he was an animator at Marvel since he was like 14. Now they're, they're SaaS for faceless video. It's doing like $120K per month. It's absolutely killing, right? But he's still doing animation for Marvel. He's still producing my show. He's filming it. You could just do like multiple things. You could have like other hobbies, you know, you go shooting all the time. Like you, you have hobbies like outside of it. I'm like, you know, building the software and then I'm working on a screenplay. Then I'm like doing shit.

Exactly.

You can, do multiple things. I think I completely agree. Like all the founders I know who are like cold plunge, 6 a.m., like their companies actually go to zero, like in, in, in like very short periods.

It doesn't work.

They're over-optimizing and failing because they're over-optimizing and then you meet them up.

Yeah. And you meet up with them and they're like, oh yeah, bro, I don't want to stay out too late tonight. I don't want to go to a bar because I got to be ready for my cold plunge. No, come on. Yeah, that's not how you work.

The real connections too are like out doing silly shit. I met you at a party, uh, you know, just Ziv's party. Yeah, that was the first time we met. Yeah, Ziv's party. Yeah, his birthday party.

Poker game maybe.

Maybe it was a poker game, but that's the point. But that, no, that is the point. I think actually like poker is big in Silicon Valley. It's big in the tech world.

We're starting some poker nights.

Nice, nice. We gotta play.

I'll let you know.

And because like sometimes you gotta just like get off working, like just stop working and just like hang out sometimes, you know?

I love the Churchill example because just the fact that Churchill—

He was so productive, bro. He painted. Do you know he was a bricklayer?

He saved Western civilization, so you can't argue that.

All 30 buildings.

Yeah. Yeah.

Insane.

Actually, I didn't know that, but that's even cooler.

He was a bricklayer and painter and author of many, many books.

So you can't argue that it's better to be hyperproductive. Like, no, you need to be productive, but the most productive man In the whole Western lore didn't do that.

So 100%, I mean, and Elon's not out here cold plunging.

Elon, he— what the hell? He's on Twitter, bro.

Trump, Trump too. I mean, McDonald's all day, one meal a day. Is he really one meal?

You never saw this article? It said like he doesn't eat all day and then he just eats a feast of McDonald's at night. That, I mean, I don't know if I'd follow it word for word, but he's, he's unarguably efficient.

Yes, he's very efficient. Yeah, he gets shit done. And the fact that he has so much energy at his age too.

Something he's doing right or something's going on there that we don't— You gotta just steer into being human.

Steer into being you. For me, it was embracing the cringe. I've written about this all the time. Cringe is the new cool. Just lean into being you, whatever you are. And that's why I was also like, leaning into talking about you and your co-founder are very different personalities, right? Like he's, you know, I feel like he's very serious for the most part, at least from my time talking to him. You're kind of a goofball having fun, like.

He jokes around too.

Yeah, yeah, you kind of can do multiple things and be, when you be yourself and lean into it, it works better in your business. Like, for me, like I'm posting like goofy shit for my meme software and like leaning into it, but then I could also be serious, you know?

People like humans.

Yes. Vice signaling, dude. Vice signaling.

What is vice signaling?

So it's opposite of virtue signaling. Virtue signaling, you're trying to look better than human, trying to look perfect. Vice signaling is showing you have flaws, you know, and people respect that because you're just being a human, you know, like you're not trying to be better than anyone else. You're just showing like, hey, like, I'm just a normal dude. Like, I have my own problems too.

It's like the— it's like in Silicon Valley, you know, this like motif of like all the rich people drive Priuses. That's not cool. Like nobody, nobody related to Zuckerberg until he started acting rich. Yeah, lean into being human, do whatever you want to be, goofball.

Yeah, dude, Zuck as an Instagram influencer, like now he's cool, you know, he's just having fun. Yeah, I've told you my, my hypothesis, right? It remind me, um, I think that Zuck finally got head and that's— it changed everything. So like for 20 years, him and Priscilla were married and he never got head. They had kids, but like never got head. And finally he convinced her after 20 years. And then that's how the fluffy hair started and just like got really excited about life.

Like, and like those video games. Yeah, exactly.

Exactly.

Maybe.

Let's talk. Let's talk investing a little bit. And, you know, there's been a lot of talk about where value will accrue in AI.

Value accrual.

Value accrual, man. It's big words, big words here. But, you know, there's a few years ago it was all about like, oh, it's AI models. And then, you know, fuck the ChatGPT rappers. Like, now it's like, oh shit, man, there's a lot of money in ChatGPT rappers. It's like, where, where's the money? I personally have been struggling to like, You know, people keep offering to invest in Memelord. I don't want to take investments. I'm happy. I'm doing me.

Don't do it.

I know, I know. Hey, and I invested in you and you're telling me not to invest and take investments. It's like, it's like, where should, should people be investing in software now when so much is like, can you build it right? Without, without capital. And I think that's also why there's such a hardware revolution as well, because it's easier to build software. Where, where's your head at right now? Where do you think like, The value is accruing in AI. And I think personally, reason why I invested is like, I think it's accruing where you are. It's the backend, the middleman, the prompt layer, you know, the prompting. It's, it's, um, the prompting layer, dude. Yeah. It's, it's the middleman everybody needs.

You're not wrong. No. Uh, I like the quote. I think it was, I think Sam Altman said it recently of like, maybe AGI comes out and nobody cares. And like, I think that is spot on. Like, I think I think it was all a marketing ploy. This AGI is going to come out and everything's going to be upended. There's going to be no more jobs. We need this Marxist UBI. We need this and that. And like, nothing's going to exist anymore and you're not going to have to work again. I think it's all bullshit. I think it's a great way to sell products because if I'm a business leader, if I'm a big finance guy, whoa, OpenAI has technology that's going to kill the world. I need that. And, uh, but like, having said that, I think what that means is things aren't, things are gonna be disrupted. AI's huge. That's why I'm working on an AI company. But things are still gonna work as they worked before. You're still gonna be able to innovate. You're still gonna be able to build new things. You're still gonna be able to build your ideas and you're still gonna be able to make money off of it. Will the markets change a little bit? Yeah. Like I think now's the best time ever to be a indie hacker or a bootstrap founder. Because what a, what vibe coding, we'll talk about vibe coding. What vibe coding is best for is building the MVP. After the MVP, it gets a little bit less useful, but for building the MVP, you could build so quick, you could make money so quick. And also, you know, it's the best time for, if you really wanna make money, scamming people.

Why is that? Why do you say that?

Oh, I think we're deep in tech, so we know what's going on. Most of the world has no idea how far tech is. You could scam so many people. You could fake voices. You could, you could now reliably pass the Turing test and most of the world still hasn't caught onto that. I'm not saying do it.

Yeah. But no, I think, I think it's interesting. I, uh, when Deep Research came out, uh, I saw somebody post about like somebody could start a, uh, you know, a ChatGPT wrapper, but say it's Deep Research and then just like sell that, but it's not actually like Deep Research.

Like, oh, like the opposite.

Yeah, like they're saying like, hey guys, like this is a special ChatGPT, like talk to— it's just like not that you make your own deep research pretty easy. That's what I'm like, you could, you could just make shit up.

Yeah, yeah.

But yeah, let's, let's talk not scamming people.

Okay, so not scamming people.

I feel like 2021 was the, was the high for that, the NFT days. That was the high.

Well, NFT is different type of scamming people.

Okay, all right, fair.

That's like under the legitimate reason. This is like just like— you ever seen these like call centers in Jamaica and stuff?

Jamaica?

I think Jamaica, Philippines.

Okay.

Yeah, I watched a show about it once. Yeah, Traffickers on Nat Geo. It's pretty good.

All right, cool, cool.

But anyway, um, yeah, you could scam people pretty well because people don't know it's real. Oh, we were talking about real money making, not scamming. So I think you're just like building MVPs, like This is a bad thing to say because I want to always encourage people to be founders and start their companies. You can make a lot of money being a software engineer at a company and just doing something on the side.

Oh, 100%, bro.

More than ever before.

I've seen a lot of people doing it where they're founders while working full-time jobs and chilling. Yeah. But yeah, now is the best time in the world to be a bootstrap founder. Completely agreed. You know, if you can market, if you could get distribution and you have a product like You know, I'm living good. I'm having fun.

And if you're a Bootstrap founder, you can use Prompt Layer and we're not even going to charge you a lot. Maybe it's going to be free. Really?

For Bootstrap founders?

I should check that out. Well, it's like usage-based. So like, if you're not blowing up too much, it's a good, good way to start.

That's fair. That's fair. How have you guys been finding customers mostly?

Most have came inbound. That's pretty impressive. For the most part. We just hired— actually, he just started today, a head of sales. Nice.

And you're over here talking with me on a podcast.

Yeah, I guess so.

That's, uh, you can't be happy about that.

Well, this is how we get the word out, right? This is, uh, I'm just playing. No, no, it's, it's in two different modes, right? It's, uh, outbound sales, which we're trying, which is not this at all. And then what I think about a lot is how do we become the prompt engineering company when all this stuff, like I was saying, LMOs, like all this stuff is going to fade away. In 5 years, the generational company is gonna be— know, if you think about prompt engineering, you're gonna think about that.

When you think about— I already think about Prompt Player.

I like that.

Maybe it's 'cause I have the hat.

I like that.

Yeah, yeah, yeah.

Um, I think when you, uh, when you think about spreadsheets, you think about Microsoft Excel.

Yeah.

We wanna be that. Yeah. Yeah.

I love that. I mean, that's, uh, same thing with memes, dude. When you think about memes, you think about Memelord, you know? And it's funny 'cause you see that happening in our little world.

Yeah.

Like if you ask our friends prompt, like, or prompt engineering, they'll say Jared, right?

Right.

If you ask him memes, you think Jason Levin Levin. But like, how do you become that for the fucking world, right?

And that's time scale.

Time scale. Uh, building an endless moat, you know?

Right. Do you think it's about— so this is something I go back and forth on. They say like, it's so hard to find one channel that works. Do you think it's about just jumping on that, doubling down, tripling down on that one channel or adding a few and experimenting?

And yeah, that's a great, great, uh, there's a great Peter Thiel quote. Where it's, uh, finding one distribution channel is, is so hard. Um, most businesses never do it. Like, the reason most businesses fail is not because of bad product, but rather bad sales. If you find one distribution channel, you're good, right? You could have a successful business. I started Memes Make Millions, my book, with that literal quote. And, um, for me, you know, Twitter's my main distribution channel, uh, and so I double the fuck down there all the time, right? But as I've grown that, I've started experimenting. Personally, I think you've got a lot with Reels, right? Not as— I mean, not as much money-wise for content. Yes. But like my number 2 channel would be LinkedIn. My number 3 would probably be like newsletter and SEO as well. I've been growing a lot there. It's just like doing that shit. But personally, I think it's finding one distribution channel that really works. Doubling the fuck down and then slowly, as you double the fuck down, then slowly start experimenting. So I'll give you another good example. So like, people, a lot of time they're like, which platform should I start on? So I have a couple thoughts on this. One is you should start building on the social media platform you love the most, you feel the most comfortable on. I don't feel comfortable on TikTok. I grew up on Twitter since I was like 13. And so like, Twitter is my natural. I've been tweeting. And so like, I couldn't imagine my life without it. So I spend a lot of time on there and I feel natural. Like, your customers are everywhere. The internet is so big that if Instagram for you is like the most natural, then it's making Instagram videos like our friend Justin, right? You know, if it's, if it's Twitter like me, it's tweeting. I think you see a lot of people like, oh, our customers are on this platform, we got to go there. It's like, dude, the internet's huge. They're everywhere, right? It's just like, where do you find them and what do you feel natural on? And if you're trying to force yourself to do something you don't like, you burn out super fast. And so like, for me, I burnt the fuck out out of like, you know, TikTok or whatever. So that's lesson 1 that I think about a lot. And then number 2 is you either start text first or video first, but it's very hard to do both at the same time. And so you look at any big creator, you look at, you know, somebody like Greg Eisenberg, Pomp, you know, we're in his studio right now. He started text first. You start Twitter first, you move to LinkedIn, and then once you scale a team, then you move to video. And so I'm just starting really pushing this video now because I've scaled Twitter and LinkedIn. And so you look at people, the reverse too, right? MrBeast started on YouTube. A lot of big YouTubers, they start on video and then they hire a team to go ghostwrite for them on Twitter, LinkedIn. Right. And I've been that ghostwriter before. I talked to Night Media, you know, MrBeast's team. They all have ghostwriters.

So once you build the video, then you go build the writing using the distribution and the reputation.

Doing video and writing at the same time, impossible. Until you have a team and until you have distribution on one platform to go push it. Like my podcast is only growing. I mean, not only growing, but it's growing a lot because of Twitter, LinkedIn, and my newsletter, all the writing.

Right, right.

Like Greg Eisenberg, same thing. His podcast is growing so fast because he pushes it out to his 400,000 followers on Twitter, his 100,000 newsletter subs. So you push it, push it vice versa, compound it. Exactly.

So as a semi-content producer myself, Yeah, yeah. Who— this is not my full-time job, but this is part of our strategy.

Yeah.

You're saying don't, don't try to experiment. Don't, don't do—

I would say doubling down on Twitter and LinkedIn for you personally.

LinkedIn works well.

Yeah, I know LinkedIn works really well for you. There's so much alpha on LinkedIn.

Everyone's scared of it. Everybody's scared.

If you lean into it, especially as a founder.

Be cringe.

Yeah, exactly. Embrace the cringe. If you're a founder, like, you don't have a boss to worry about. And you could just like be free to like have fun. Yeah. And it's such a good feeling. Um, and so you don't have to like worry about getting fired, so you could just like, if, and if you're not worried about like what other people think or like, you know, your high school friends, which you high school, whatever, who the fuck cares? Like you could just do whatever, you know?

I will say the one thing on LinkedIn, which is too cringe for me and maybe even too cringe for you if possible. Politics on LinkedIn.

Oh, I don't touch that.

No, that's the weirdest thing. John McCoons always posting some like woke shit.

I don't, I don't touch that, man. I don't touch that. I know Reid Hoffman is where he stands. I'm not, I'm not going to that island. You know what I'm saying? It's an island.

I'm just waiting for that to come.

I'm not trying to, not trying to go there, but I think, uh, yeah, Twitter, LinkedIn.

I'm on the hit list.

And dude, your, your Twitter is, uh, is popping too. Like, for, for, uh, You're, you're focused mostly on building, but like when you tweet or when you post on LinkedIn, you have, you have, you have like great followers too. Like the Alexis Ohanians of the world who invested in shit like that, you know? Yeah. So it feels like very clear, like, you know, when, when your network is in one place and it's like the natural place, I think that's the place to double down.

Yeah. I think it's, it's hard for me because, you know, we're building a company, so Twitter just feels like a distraction. But it's not. We get a lot of users from it, but, but you know, like the dopamine cycles just screw with you and then you're like, I got nothing done today. It's like, I got a few likes on Twitter and then maybe that translates to a customer. I'm sure it's translated to customers for us because I meet people and they're like, oh yeah, I've heard of Promptlayer. I'm like, well, how the hell have you heard of it? And they just have.

Do you guys track that? Like UTM tracking and stuff?

We just started doing stuff like that.

Oh. Uh, cause then it, it, you, you stop thinking about this as a distraction and you're like, oh shit.

Yeah. Yeah.

This is real.

There's money coming in. Intuition. Yeah. At the beginning.

Dude, At first it's all intuition, like 100%, man. Like LinkedIn, I didn't have UTM tracking at first and now I do, but it was like, I see a lot of people saying they're signing up from LinkedIn. I don't know the exact number, but I know it's more than Instagram. So I'm gonna spend a lot more time on that.

And even knowing the name, it's hard. Like you can track it and track it. We're trying to track everything now. And, uh, we've been doing a lot of blog posts and stuff like that.

Dude, data attribution's so hard. It's very, very hard.

But it's also like, it feels like it conflicts with the vibey, uh, like I sense the people. Oh shit, this blog post did the best. Maybe we should do more of that.

Yeah, yeah. 100%. I don't, I don't do it much for newsletters, but for SEO blogs, for like stuff like that, it's super, super helpful.

Oh, SEO you have to.

SEO is all data.

Yeah. Right, right.

Very little vibes.

Yeah. Yeah. I don't know. It's like, but, but it's hard to balance the how much should I be focusing on tweeting and like how much writing LinkedIn posts, because you could put a lot of time into it. And as you know, like to do it well. Yeah, it takes time and it's, it's a context switch.

And as a founder like yourself, you're not even just competing with founders. You're competing with the ghostwriters and the creators who are like, yeah, doing their own thing too.

And you're competing with, with the AI newsletter writers and all those fuckers who are like, but honestly, I'm not even worried about that because those are like, as a founder, you, you're bringing a unique voice. Yeah, that's true. It's not, as long as you're not too noisy about it. Some founders, they're only like tweeting once in a while and their tweet will blow up and they're doing like, but I think quantity.

Yeah, I like quantity. Quantity for Twitter. I would say a little less quantity for LinkedIn because they're, I think, a little harsher around that. But, but Twitter, it's all quantity, dude.

Yeah. So remember back in the day, um, I feel like my first intro to this was Facebook profile pictures in high school where you wanted all the likes and like, yeah, dude, the dopamine rush of getting 100 likes on a profile picture. And the lesson I learned back then I wonder if you got there too, where it's like, have a girl in the profile picture. Oh, that's a good one. I never met a girl, but yeah, that would have helped. Um, you got to go comment by comment, liking each comment and responding to each comment individually.

It's the same thing.

Exactly, dude.

That was our training ground.

Exactly, exactly.

Profile pictures was a training ground.

I had one friend, he would always get the highest likes on Facebook. I gotta ask him his ways. Maybe ask him, found some new Twitter strategies.

That's smart. Um, I feel like there's just so much alpha for Twitter in sending DMs too. That's—

is that how you raised a lot of, you know, um, not really from Twitter actually.

What was it from? Not LinkedIn, right? So people don't have money on there.

Raising is hard from cold. I've done a few like that for angels and stuff like that. Uh, but our, like, our first check was Peter Boyce, uh, who I'm not sure if you're familiar or not. He's like the mayor of New York now, I feel like. He started, uh, Rough Draft Ventures back a long time ago. He was at General Catalyst. Now he has his own fund. I knew him since those like hackathon days back in high school.

Yeah.

And just stayed in touch. And when I quit my last job, he's like, let me be the first check. And we didn't even ask anyone else. We're like, yeah, we love you. Let's do it.

That's so sick.

Let's be part of the journey together. And, uh, I love that, dude.

Nothing better than a warm intro, you know?

Yeah. Nothing better than a previous relationship. And then otherwise you get founders to introduce you.

Dude, I think, uh, I had this thought earlier. It was like, you can only get a warm intro when you're hot though. Right? And otherwise you have to do cold DMs. That's how I started my whole career, was cold DMs. We're just sending cold DMs because I was cold and broke. And then when you're hot, you're just getting interest all the time.

So I think when you're hot, the cold DM is even easier. Oh, 100%.

It's easier, but you don't need it as much.

Yeah, you start cold. I did a lot of my cold emailing when I was in high school, so I would always put in the subject line like high school student. Yeah, exactly.

Such a hack. All right, Matt said we got to be wrapping up, but if you need layer some prompts, go over to PromptLayer. Um, and, uh, if you need cake, go to one of Jared's events.

Yeah, promptlayer.com.

Hell yeah.

Thanks for having me.