5 AI Tools that make you feel like you’re in 2050

Shaan Puri

Alright—forget ChatGPT. That's old news. You're talking about ChatGPT? My mom knows ChatGPT. Today we're playing a little game called **"Blow My Mind with AI."** Sam and I set each other a challenge: I said I'd bring some AI tools that I think are pretty cool, you do the same. Let's have a little show-and-tell and get smarter together about cool things you can do with AI. The premise for the episode is: what are the AI tools that everybody should know about—tools that feel like they're on the *bleeding edge* or that are going to be big in the future? That's what I want to see today. Okay, so you've seen—everybody talks about agents. Oh, agents, agents, agents—agents are all the rage. Is anybody using an agent? What are they using agents for? I kind of got into a rage about agents and I was like, let me see if I can get one useful agent working. I tried a bunch of different tools, but I want to show you a fun one that this guy Garrett made. Garrett is the founder of **Pipe Dream**, which is a very cool company. He's in Austin—you might know him. They're basically building underground delivery tunnels.

Sam Parr

Is this company working?

Shaan Puri

Yeah. I mean, it's a really hard company, but they built this thing out in Austin. They have a first product called **"Goods,"** which is an autonomous system—it's kind of like **DoorDash** or **Instacart**, but it's the **fastest grocery pickup in the world**. You order, and in two minutes a crate of picked items gets to you. You then pick it up from your local **Goods** locker.

Sam Parr

*That's insane.*

Shaan Puri

Okay, so check this out. He— I guess as a side project— built this thing called **doanything.com**, and I want to give it a shout because a lot of people know about Claude and these other things, and I thought this was a pretty cool site. If you go to **doanything.com**, you'll see it, but I'm going to show you what this can do. So check this out: I go to this thing, it says *Do Anything*, and it basically promises, "Just tell me what to do—I'll go do it." It's a to-do list that does itself. I said: **"Analyze my YouTube channel performance like you're a world-class content strategist."** The channel's called *My First Million* on YouTube. It's a podcast. I couldn't even bother to give it the link; I was like, "I don't know, see if you can figure this out." It thinks about it, and at first it asked me, "Hey, can you log in with YouTube?" I was like, "No, I'm not going to do that right now. I don't know the password. I'm just going to say no to that—tell me what you could do." So it said, "Okay, got it," and then it does the reflection. Here's what it does: it gives you the "state of the union." A couple minutes later, it said: > State of the union: You built a powerhouse with 869,000 subscribers and 300,000,000 total views. That's a massive footprint in the business niche. However, a world-class strategist looks past the vanity metrics to see the real story. Then it noted: "Your average videos get 20,000 to 45,000 views for a channel with nearly 900,000 subscribers. That's a big lag, where your core audience—only 3% to 5%—are clicking on new uploads immediately." It then looked at what the winner episodes were, why that title was good, what's working, what's not, and provided strategic recommendations. "That's a great title."

Sam Parr

Wait—scroll up. It says, *"How to make $1,000,000 so fast your accountant gets worried."*

Shaan Puri

"Gets nervous." "Yeah, exactly." I'm like, "Okay — that's already better than most of the titles we come up with when we try." They did it right away, and then it basically says, "Okay, here's what I think you should be doing." Right: **double down on contrarian wisdom**. So instead of just "Guest X tells their story," make it "why everything you know about X topic is wrong." Get your thumbnails right. Get your shorts to lead to longs. Then I was like, "Can you help more?" So I said, "Okay — create a one-month content plan designed to **10x** our monthly views while staying on-brand [original: 'brown on brand'] and in our niche of **Business Junkies, Money Wisdom,** and **cool business trends**."

Sam Parr

Let's just make a comment here. You did this **19 minutes** before we were supposed to record.

Shaan Puri

Yeah, yes I did — and look what it did. Okay, so then it starts doing it and it's like, "Alright, here's a one-month viral outlier content plan." > "We're gonna go all in on **video essays** because they're going to have a higher conversion rate and a higher AOV. We're going to make our thumbnails feel like a business documentary rather than two guys talking on a podcast or a Zoom call. For topic selection, we're going to look for things that are a big promise with low friction — unsexy businesses, money traps to avoid, 'one-person, million-dollar portfolio' — how to do it." Then it gave me this sheet that was like, "Here's your plan," so it starts giving me actual video ideas: - How to Build a One-Person, Million-Dollar Portfolio in 2026 - Brutal Truth Series: Why Your 9-to-5 Is Actually Your Biggest Financial Risk - The Seven Money Traps That Are Keeping You Poor - Future Hype: The Next NVIDIA - The Death of SaaS — What's Going to Replace SaaS? - Inside the Brain of a Billionaire - How X Built a $100 Billion Empire - 17 Lessons from a Thousand Episodes of My First Million Then it's like, "Would you like me to script these out? Which one of these do you like? Would you like me to come up with thumbnail ideas?" And you just keep saying, "Yeah, yeah, go ahead, keep going. Yeah, have at it, knock yourself out." You just do that. I mean, I just feel like the idea of having background workers that are constantly working for you... And by the way, this is getting better and better. So, have you seen **ChatGPT's Pulse** feature?

Sam Parr

Yeah, I pay for it. It's **expensive**, but it's *pretty great*.

Shaan Puri

But it's cool, right? It knows what you're already trying to do. Every day it'll proactively search for stuff and write to you about the things you care about. You don't have to prompt it — it'll come up with prompts and serve them to you. I think this is becoming more and more common: tools are automatically triggered by reading what's going on. I'm downloading other tools that will read all my Slack messages and, based on what's being said, just try to do the work. Furcon is working on a tool like this called **Nebula**. It's an app where you connect your Slack, Gmail, Google Calendar, or whatever you want to connect, and it just knows what to do. For example: if you give it your calendar, it knows what meetings are coming up and will create prep docs for every meeting. I haven't gotten this to work well enough for me yet. It's an early prototype and I'm consulting with good friends on it, so fair warning — it's not the most functional thing yet. But I'm playing with it because one version has an even more interesting premise: you shouldn't need to tell it what to do. It should figure out what to do based on what's going on. For example, if a bunch of people on your team upload new code to **GitHub**, it will automatically read it, summarize the changes, and tell you about it tomorrow morning. You don't need to tell it, "Go look at GitHub." If it's connected to GitHub and something changed, it's going to tell you about it.

Sam Parr

So, is this worth... is this worth trying now?

Shaan Puri

Yeah, so for me it's not that useful because I don't have a huge core team. We don't have a daily stand-up, we don't push code to GitHub—it's just the three of us and we mostly just talk to each other. I already saw the messages. But if you have a 10-person, 20-person, 30-person team, this can be **extremely useful**. You can set up different workflows, like telling it, "Hey, keep checking my dashboards and give me summaries, and tell me when anything pops off." You can just tell it to do stuff like that in a chat interface, and it figures out how to do it. That's the idea behind something like this.

Sam Parr

With everything—I mean, this...

Shaan Puri

Can do.

Sam Parr

It does your job for you. I'm already using **ChatGPT**. Sometimes, if I get in an argument with one of my employees—or if what they did angers me—but I need to be productive about it and reply in a way that solves the problem, I'll upload the email and say: "You know, my instinct is to write this angry thing, but obviously I know that's not right. Can you help me write this email in a way that gets me what I want?" So it's already doing a lot of the talking anyway.

Shaan Puri

Hey — quick break. If you're liking this episode about **AI tools**, I have something else you might be into. A while back our friend **Greg** came on the pod and did a "show-and-tell" of his own. He walked through five tools you can use to build a business that doesn't require a bunch of employees or a lot of capital. They're not the obvious AI tools everyone's talking about; they're more the underrated, under-the-radar tools that power nerds like us—and Greg—try to find and test. So, if you're looking for a list of AI tools to check out and kick around for your business, you can get it right now. It's in the description below — totally free. Hope you enjoy that. And now, back to this episode: are you using **WhisperFlow**?

Sam Parr

Oh my God, dude. I don't—I barely... I barely type anymore. Do you use it on your cellphone? It's *so good* on the phone, too.

Shaan Puri

**WhisperFlow** is pretty amazing. It's an app that basically does—well, "text to speech" is actually the wrong way of saying it; that's the nerdy term. What's the real way of saying it? Typing is a pretty slow thing to do. It actually requires a lot of energy, and you don't really realize that until you're able to stop doing it. What WhisperFlow is, is it's on your phone. If I go to our iMessage, there's a thing that just says **"Start Flow."** All I do is click this button. Any app I'm in—this isn't just in ChatGPT or Gemini—I can do this in Slack, Facebook, Instagram. I could write a caption, an email... I can do this anywhere. I just hit Start Flow and it starts listening. For example, I'll say: > "Hey Sam, for today's episode do you wanna do the AI episode? > I'm thinking that we could each show two or three different cool apps. Not sure if you've already prepped it, but I have two for sure that I can do. Let me know what you're thinking." Boom—now it's captured that perfectly. What's cool about this is that because this is all they specialize in, the app starts to notice my preferences. For example: I like to break up lines. I tend to remove periods because they look too formal in text messages—they can make you come across like an asshole. I separate lines by pressing Enter; I don't like big blocks of paragraph text. So it learns as I do that once or twice, and the next time it's just going to do it automatically. It's going to format everything to look more like me. [Note: "WhisperFlow" here refers to a mobile voice-typing / speech-to-text app that converts spoken words into formatted text.]

Sam Parr

Mine will add bullet points, so I'll be like, "Hey, I've been thinking about — I think I've been thinking about these **three things**: - The first thing is this. - The second thing is this other thing. - The third thing was this thing. Also, I was thinking... actually I didn't mean to say that. Scratch that.

Shaan Puri

That, yeah.

Sam Parr

Yeah, so it does one and two. Then—yeah—when I said "actually," I didn't mean that. I meant the *fourth* thing. Like, it actually adds a fourth, and it doesn't say what I said previously.

Shaan Puri

I will go for a walk, walk my dog, pop an AirPod in, and just come up with stuff — whether it's an email, a blog post, or a book chapter. I'll ramble aloud for about 30 minutes. Sometimes what comes out is a line like, *"creativity is the new productivity."* The last 50 years were about who could be most productive, but now with AI everybody's instantly productive. It's actually about who has the best ideas, so creativity is the new productivity. I'll talk like that and ramble, then at the end I'll give the recording to ChatGPT or another service and say, "Hey, can you remove the rambly parts, structure this, and make it into an actual book chapter?" It usually gives me something that's 85–90% there. I didn't even have my phone out or have to look down at my screen — I'm out in nature, just walking around. It's a very different way of working. My dad, if he saw it, would be like, "What are you doing? This is not work. I don't even understand what's happening right now."

Sam Parr

I think that's the difference between the **winners** and the **losers**. Basically, the people are going to use it and they're going to 10x or 100x their output. The losers are going to say, "This is too new, this is too fast — we should keep it old school."

Shaan Puri

Yeah, exactly. I wrote this blog post six months ago—maybe a year ago—and I basically called it the *K-shaped economy*. People say, "AI's gonna take your job," and of course it will if you don't use it. It's like computers taking jobs too. What I argued with the *K-shaped economy* is that it's not that your whole job disappears. A job is basically a bundle of tasks. Even if you work at a grocery store, your job is a bundle of tasks: restocking shelves, working checkout, bagging, pushing carts… whatever you're doing. It's a bundle of tasks. So first you have to think: what is the bundle of tasks in my job? Then ask how many of those can I get AI to do instead of me, or how many can AI make me much better at doing than I am currently? I said that roughly **80%** of those tasks will be replaced by AI, and **20%** of the tasks will become *enhanced* by AI—things you'll be able to do much better. For example, take the YouTube content strategist [role]. That's a job. Now imagine you set up a little system that constantly scans YouTube to find what adjacent channels are doing, what's working, and which videos are outliers. Break those outlier videos down, then have a second system analyze them and ask, "Okay, that was an outlier—why do I think it worked so well?" Then it references that against the "My First Million"-like brand corpus and says, "What should you be doing? How could you take this outlier video and make it your own?" Use that to pitch content ideas to us. Kind of like that.

Sam Parr

It's crazy.

Shaan Puri

Here's what I think you guys should be doing now. All of a sudden you're **ten times** better at your job. Google keeps releasing these really powerful things, but they suck at marketing them. If you just poke around what Google has, they actually have better tools for a lot of things. There's this thing called **NotebookLM**. If you go to the homepage, you just create a notebook. If you say, "Alright, I want you to make something for me or teach me something," it's a great way to learn about anything. What I did was drop in a podcast. Dwarkash does great AI podcasts, and Karpathy came on. Karpathy was the former head of AI at Tesla; he's one of the self-driving gurus and one of the best thinkers about AI in the world. The episode is an hour and a half. I can either listen to it, or I can first just get a quick summary. I dropped in the YouTube link and then asked it to turn this into a slide deck. It created this: **"Summoning Ghosts: The Decade of AI Agents"** — an entire presentation auto-generated from a single podcast link. Here's a quote from him: > "I've had fifteen years of experience with people making predictions. If I just average it out, it feels like a decade for me." And then he's talking about things like "animals wait versus wait" [unclear].

Sam Parr

So... did those? Can you go back? Let me see exactly what he did. So, all of these images are from it [unclear: "it made it"].

Shaan Puri

"It made it. He didn't show anything. It just looks at the transcript and turns it into a **PowerPoint presentation**, right? God — so this is *amazing*."

Sam Parr

Oh my God. What's this called on **Google**?

Shaan Puri

No, I...

Sam Parr

I know, but I use that. "What's this feature called?"

Shaan Puri

This is no big deal. I just dropped it in. I clicked **"See here"** on the right — I clicked **"Slide Deck."** So all I did was give it the link, right? I gave it this YouTube link — see this YouTube video right here. I just dropped in the link; I didn't watch the video. Then I said, "Make me a slide deck about it." That took like ten minutes, and then it made it. So, like, okay — that's cool. Then you think about what's next. So I saw this guy...

Sam Parr

Was all of those graphics *legible*? Did it all make sense? Because...

Shaan Puri

Was—yeah. They're not... okay. They're all *legible*, as in they're not just gibberish. But, you know, this wouldn't be—I'm sure there are some things in here that are a little bit off or a little bit dumb.

Sam Parr

But something like *overly complicated*.

Shaan Puri

I don't know what this is — this chart looks kind of crazy, so I'm not sure how useful it is. But I saw somebody do this with Ray Dalio. Ray Dalio gave a talk somewhere and he turned it into *this*, and I was like, "Oh wow — this is so much better than going and listening to a Ray Dalio talk for an hour." I can get this. Now I might actually be interested in listening to the full talk, because this gave me the *gist* of it. Okay — so then, check this out.

Sam Parr

So, man, staying on top of all this stuff is *really hard*.

Shaan Puri

So now this guy built a tool that basically makes presentations look *sick*. The thing I gave you before was cool because it made the presentation, but it looked kind of granular, academic—a little whatever. Check out this guy's thing. He's like, "Do all of my presentation decks like this." Okay, watch this animation. He gave it his slides—he gave it an initial slide and then the outline of what he wanted to talk about. Watch what it makes. Watch this animation, exactly in the style. So when you just scroll, check out what happens if I start playing my presentation. I get not only perfect slides, I also get—check out what happens: it just gave it this, and then it transitioned perfectly into the next slide. All the next slides looked like this static, right? The first slide looked like this static. AI made it so that it will automatically make it look like you animated one into the other, even though they weren't related at all the way I wanted them to be. But isn't that *sick*?

Sam Parr

"What was this called? What's this called?"

Shaan Puri

That I’ve made, you know what I mean? *Like, how sick is this?* These are just two independent slides. Remember how Prezi broke everyone’s mind because it would, like, “whoosh” over from one slide to the next? This was like Prezi, but the AI thing — for me I was like, “Oh, that’s amazing. I want… I want my… let’s take…”

Sam Parr

A shot up.

Shaan Puri

Look like that.

Sam Parr

**GlyphSlides.pages**

Shaan Puri

Yeah, so this guy **Fabian** posted this, and his app is called **Glyph** (glyph.app). I haven't actually tried to use it to make one of these, but I thought the demo was unbelievable. It's kind of like— you know, in any animation you have a start and an end. Usually you have to create every step in between, like the 30 steps between the keyframes: where it starts and where it's supposed to end. But with AI you just show it the start and the end, and it automatically animates it to make it look like A transformed into B. How cool was that? By the way, he didn't even make that. He just used this off-the-shelf thing called **Cling** that does video transitions, and he's like, "What if I just plug video transitions into a slide deck — will that work?" And then—boom—it worked.

Sam Parr

Alright, so let me share my screen. Okay — here I'm showing off stuff that I made with **AI**. I've been using **Claude Code** all week... or **Claude Cowork**. I don't even know the difference, to be honest. I think Claude Code is one thing and Cowork is an easier way to use Claude Code. Is that right?

Shaan Puri

Yeah. **"Co-code"** is like *"you're gonna create a program."* **"Cowork"** is like a junior worker that you could ask, "Hey, can you analyze this data? Hey, can you create this plan? Hey, can you make a presentation out of this?" You hand off work to a coworker.

Sam Parr

Alright, so I made a thing. Look — here's the problem: I read a ton of business biographies. You say you want to read a ton of business biographies. I don't know if you actually do, but you say you want to. Someone in the comments on YouTube said, "Sean's misogyny this year is to finish a book," which was pretty fun.

Shaan Puri

It's going to be a *tough* one.

Sam Parr

So, I like reading business biographies because they're a great way to learn about business and they also make me feel better about my life. You see the problems other entrepreneurs had, so it makes me not feel alone. But when you read a 400- or 500-page biography: a) you lose track of the timeline; b) it's hard to know exactly what applies to you; and c) if the biography is from the 1930s or 1940s, they'll talk about dollar amounts that are hard to apply to modern day. They'll be like, "John Rockefeller bought this thing for $50,000," and I'm like, "Well, that's not a lot of money." Then it's like, "Oh, well, that was actually about $8,000,000." Okay—now I have the context. So I built an app. Here's what it does: **I can download and upload 1–5 business biographies under the same person.** The way I tend to read is: if I want to learn about Ted Turner, I'll read three or four biographies on him to get different context. I made this for my pre-work / pre-reading because I'm kind of weird — I like to say, "I'm going to learn all about Ted Turner; I'm going to read these three, four, or five business biographies." I made a website where I can upload three to five biographies in PDF, EPUB, MOBI, or text formats.

Shaan Puri

So, just to slow that down: do you go to **Amazon** and buy the **Kindle** version, or do you go somewhere else to get the **PDF** or the **EPUB**, or whatever?

Sam Parr

I do — I do read the hardcovers. But in order to make this work, the easiest part is to do some *bootleg*... Okay? You do need to download them illegally, but I don't feel bad because I'm paying for the book anyway. So, I upload it to a website and I've been able to basically make a business Wikipedia. Let me explain what I mean. Here's what the site looks like. It's called **"Bio to Notion"** because the output is going to be a Notion page. Here's an example: I had this book about **Ted Turner**. I uploaded it and the output (which I'll show you in a second — it takes like five minutes) — I'm just going to show you the final output. It's basically this: My output is a Ted Turner business-empire timeline and financial history. The first part tells me three to five bullet points that I could read from his journey. Now here's my favorite part: it gives me a *financial summary*. Wikipedia is great, but it doesn't dive deep on the numbers, so it makes it challenging to actually understand...

Shaan Puri

Not now. I'm *nosy* enough.

Sam Parr

It's not nosy enough. Yeah, I want to know about *numbers*, so I create a *financial summary* of his whole entire life. You can see I adjusted the money to the 2025 era, and it will go through each major...

Shaan Puri

**At age 24**, he's worth about **$10 million** in today's money. He had a **$1 million** company back then. By the time he's 31, he's now worth maybe **$16 million** net worth. And then he crosses a 100 when he's 37 years old [meaning he exceeds **$100 million**].

Sam Parr

Cool — yep, yes. And then what it does is it goes through each section, just like *Wikipedia* does with the milestones of someone's life. But I have it.

Shaan Puri

"Oh, by the same... I gotta ask about that table. So it's taking raw info from the book, because the book is not telling you 'at age 31 he was worth an estimated this many dollars.' It's guessing. It's kind of like it takes the raw data and then it basically does AI analysis to try to figure out, like, 'okay, if it said his company was worth X and he probably owned—let's call it 75% at this [point in time]—then he would be worth this much money.' Is that what it's doing? Did you tell it to do that?"

Sam Parr

I told it to do that. For example, **Ted Turner** — his father owned a billboard company. His father died when Ted was only 22 or 24 years old, and Ted inherited the business. But in order to inherit it, he actually had to raise money (debt) to buy back a bunch of the assets. Long story: his dad was mentally ill and sold parts of the company in bad deals. Ted had to raise money to buy back the business. That's why it says he had a million-dollar company and it was just adjusted to $10,000,000. As Ted's story progresses, he goes from a billboard company to buying radio stations, then TV stations, ultimately buying the Braves and then starting CNN in his forties. This paints the picture of his major transactions.

Shaan Puri

Very cool. Okay — love it already. I learned something: the thing he's known for — let's call it **CNN** — didn't start until he was 40. I'm not even 40 yet, you know. It's like your *best work could be ahead of you*. Just one big takeaway right away.

Sam Parr

It's crazy, right? And just like Wikipedia — because I love Wikipedia; I read it like crazy. I love how they break it into a bio and just give you the stuff you need to know. For example: Ted Turner — his father committed suicide, so Ted, at the age of 24, inherited the billboard business, which was saddled with debt from his father's recent purchase of all these other companies. Notice I convert it to *2025 money* [adjusted to 2025 dollars]. What I did was add a section where it could apply to me. I actually uploaded to **Claude** a ton of information about me and my business, and I later deleted it because I didn't want to show it here. Where it says **"Founder's Playbook"**, it's a section where it can apply that portion of his life to my life so I can learn from certain tips. When you read a biography, sometimes you skip over the interesting learning because it's often just facts — not what he did that was really rare, and you have to pay attention to this particular... I did that with each section of his life. Then what I tried to do was: *"Founder's Playbook — what you can learn."* So when he created **CNN**, that could go in a category called **"Survive the Ridicule."** Ted was made fun of a whole bunch when he launched CNN. It was not easy for him. Here's what people mocked him for.

Shaan Puri

"The Chicken Noodle Network" (CNN)

Sam Parr

Yeah, they made fun of him *all the time*.

Shaan Puri

I feel like Trump could use that now as a great diss.

Sam Parr

It's a pretty good one, yep. What I tried to do — and this was *actually quite challenging* — was, at the end, try to get a whole bunch of photos that were relevant to the era so I could... like, make—*like this*.

Shaan Puri

Visual timeline, yeah.

Sam Parr

I created a digital timeline where I could look at them. This is actually a lot more challenging to figure out how to do, but it's been really effective. I've done this now with three or four different people who I love learning and reading about, and it's made it so much easier. I used to do this in *ChatGPT*, but I had to prompt it in a certain way and it was always really challenging. So that's project number... and

Shaan Puri

This is now publishable. You can have this on—you could create your own little *Samapedia* online, right, where you're studying great business; you know, the business biographies, basically.

Sam Parr

"It's pretty awesome, man, because I think when people read— I don't know about you— when I read a book, I always read the **Wikipedia** page of the person or the *summary* of the book. That definitely gives the story away, but it makes it so much easier to know what to look forward to. If I know, like, that he's been divorced a bunch of times, I start reading about his first marriage, and I'm like, "I know this is gonna end." I can see the cracks already. I can see why it's...

Shaan Puri

Yeah, totally. It's very helpful to have the *blueprint* of what the overall *scaffolding* of the story is. Then, when you're reading, you're putting things in buckets basically, and you know what else is to come. Versus you go in blind and then... it's sort of just like, "Did they hook me right away?" You might bounce before you get to any of the good stuff, right?

Sam Parr

And what I love doing about David—David Senra has a podcast called *Founders*—is he's really good. He'll be like, "Check this out; check this part of the biography. It talks about hiring, and here's what he says." Then David will say, "You know, I've read all these biographies. I've noticed that's a trend among a lot of the people." As they say things like that, that's challenging too... Those insights are actually hard, and so I put those insights into this project.

Shaan Puri

That's great. Okay — really like that. You could actually make this even more powerful. Right now you're doing the step where you're saying, "I wanna do Ted Turner," and "I gotta go find the biographies, I gotta get the EPUB, download it, and then do this." You can actually just have an *agent* do that. You could tell an agent, "Hey, every week — every Monday — I want you to deliver this to me." Think through who you think I might like. Find the right biographies, get the PDFs for them, put them into this tool, and deliver this. Then I want you to send me an email or turn it into a podcast — read it out into a podcast that I will subscribe to, my own personal podcast.

Sam Parr

I did not know how to do that, but that's a great idea. Then the really cool part — I didn't want to show this here, but my **ChatGPT**, which I use as my business coach, has all this context on my strengths and weaknesses and the shit I complain about. The section where it says "Founder's Playbook" is crazy tailored to what I'm experiencing on a day-to-day basis, and that was really magical.

Shaan Puri

One thing I like about biographies—or about reading in general—is I have this theory that everybody today is anxious and upset about things that our ancestors would just laugh at. None of us grew up in a war-torn state or in the Great Depression. We haven't faced real adversity, so we magnify small problems and make them bigger. My little quote on this is: "We don't need therapy; we need history." If you were actually a student of history, you would need less therapy. I think it's cool what you're doing because, in a way, it's actually *founder therapy* for you.

Sam Parr

There's this crazy story where Ted is in his forties. He's worth hundreds of millions of dollars because of a bunch of stuff he's done. He starts CNN and invests almost all of his money into it, and no one would watch it. So he does just ridiculous stuff. For example, late night they allow one of their newscasters to be a dog, and they interview the dog. Another time it's actually Ted being interviewed, but he puts a bag over his head so you don't know who he is—he plays this *secret character*. What's interesting is guys like you or me, or some people who have some amount of success, think, "I can't stoop that low" or "I can't do that." Then I read this stuff and I'm like, look what this guy did—someone who's supposed to already be this successful person. There was another time where he had to get stitches on his forehead because when he first bought the Braves, no one would go to the games. He did this seventh-inning-stretch event where it was, "who can push the ball to first base the fastest with their nose," and he was like, *"screw it — I'll do it."*

Shaan Puri

"He's like a *drunk uncle*."

Sam Parr

And I just think, if he's willing to *literally* get on his hands and knees, like "there's nothing beneath me."

Shaan Puri

Yeah, that's great. And, by the way, you could even feed something like that back into your system. You could say: the things that I remembered and took away were this — for this reason. Remember this about me: I like these types of anecdotes for these types of reasons. *Surface those in the future.* Create a **read‑me file** of the things that **Sam** really resonates with, and always check against that the next time you do another biography. Right? So the system gets smarter every time.

Sam Parr

I have to figure out how to set it up so it can email me. But I did all this on **Claude** [CoWork?], and it took 45 minutes. It was *pretty awesome*.

Shaan Puri

Alright, let me show you something. I have two big themes that I think are the themes of the year for **AI**. One of them is *mass personalization* — everything personal for you. What do I mean by this? In the last ten to fifteen years... if you just teleported from fifteen years ago to today — say from 2010, maybe even earlier, maybe twenty years ago — and I snap my fingers and you're now in 2026, what stands out to you? One big thing is: not only is everybody on their phone, which is already like a tiny computer, but what are they looking at on their phone? I would say one of the biggest inventions was the "news feed." Twenty years ago, all content kind of needed to have mass appeal. You had the newspaper that gets printed and delivered to everybody, so it had to be a little something for everyone. Now, with the TikTok algorithm and the Instagram algorithm and the Facebook algorithm and all that, it's not a little something for everybody — it's a hell of a lot just for you.

Sam Parr

Yeah... I still remember coming home from school, and we would all watch "TRL." Do you remember "TRL"?

Shaan Puri

Totally. If you look at some of the numbers for famous TV shows, the last episode of Friends had 55,000,000 viewers. More people watched that than will watch anything now. Some people are nostalgic and say, "I'm gonna create something that has that," or "I want to get back to that." But no — we just didn't have better options back then. People have voted with their eyes: they don't want that any more. They want to look at something that's hyper-relevant to them. My wife and I have very different tastes. I want to see the highlights from today's NBA games; she couldn't care less. She wants to see oddly satisfying people doing crochet on Instagram because it helps her de-stress after a day with the kids. The algorithms now give us that. This idea of mass personalization — something that would never have been possible — became possible because first everybody started creating. Instead of just journalists, you had bloggers. Instead of just bloggers, you had tweets. And then, instead of just text tweets, you had people making images, people making videos, people making text. Everybody became a creator; supply went up massively. On the other side, the algorithm was able to editorially curate for you based on your signals automatically. There wasn't a human editor choosing what goes on the front page for you. That's what news feeds did. I think that's what's going to happen now for software: personal software. You just showed a piece of personal software — something you really want to exist that I didn't really need today. Historically, software was built centrally, like the newspaper industry: some really talented people made something that would appeal a little bit to everybody. But what you showed is that you can make highly personalized software — not just "Sam who likes business biographies," but "Sam who likes business biographies and is going through these challenges in life right now," and then tell him what he needs right now. That's what that piece of software does. One category of personal software is not just apps — it's anything. What you did is called **vibe coding**. This trend — vibe coding — is basically: you roughly describe what you want, the thing makes it, and then you refine it. You're at the table ordering; the chef keeps making things for you over and over until you feel the vibes and get to what you wanted. That's happening for everything, not just code but other skills. I saw this on Twitter: a guy showed that if you ever wanted to make music but don't know how to use Ableton or Fruity Loops, don't know what a MIDI is, and you're not super musically trained, you can just describe what you want. I thought we'd try one together from scratch. Maybe it'll be a bust; maybe it'll be cool. This is not a song — it's a beat. For example, I coach a high school basketball team and we have warm-ups before the game. We want it to be hype. So I typed: "hype basketball intro music for a high school team, high energy build-up, make the crowd go wild." I didn't talk in music terms — I didn't specify a key, BPM, or instruments. I went straight to the end, as our buddy Elon says, and just said, "This is what I want—make the crowd go wild." You can see the system is thinking. It says it's coming up with drums and will create a track. It listed things like stadium brass, 808 bass, drum kit. On the side it showed the drum buildup and the 808 bass. I played the basics. For each element it asked, "I made a hook — do you want to keep that or change it?" So I said, "Add a big drop at the beginning. Make it lower in tone, more masculine." Then I sent that.

Sam Parr

Have you seen that famous clip of **Justin Bieber** where he's humming to an orchestra and they're trying to follow along?

Shaan Puri

"This is like the [derogatory slur against disabled people] version of that."

Sam Parr

Go look this up on *YouTube*. It's like *Justin Bieber* — who, I guess, I don't know if he knows how to play an instrument — humming to an orchestra. I think it's a 12- or 15-piece set. He's humming what he wants, and they eventually get it. This is sort of like that.

Shaan Puri

"**Keyword** *sorta* is doing some heavy lifting in that sentence."

Sam Parr

"It's—and it is—it's cool that you're able to do this."

Shaan Puri

"So, I told it this thing, and now it's like, *'alright, me think — let me try to do this while this does that.'* I'm gonna show you the one I made last night. I made one for the podcast. I tried to create a *podcast intro song* as my first thing. Check this out — I went back and forth a bunch, but let me just play this. [Plays audio] "Okay, alright, matrigue." So that was a little intro loop you could do, and then you could put vocals on top of something like that. You can add vocals with a different tool, or they'll add vocals eventually to this.

Sam Parr

And it gives— is it called a **MIDI**? So you can... it's like you can change the...

Shaan Puri

Yeah. So all of this is super exportable. If you're—let's say you're a DJ and you actually do have skill—you can export any of these individual pieces or the whole thing and alter it however you see fit. The important thing here is basically **skill and taste**. To do something you always needed both. You need skill and you need taste. The skill is the actual knowledge of how to use the tools: musical knowledge, technical knowledge, things like that. *Taste* is: what do I think is cool? What do I think is interesting? What do I like? Now, with AI, those got decoupled. You don't need the skill; you just need the taste. If you have both, obviously you're going to make better things, but you could still be in the game if you just have taste. So let's just see what it did real quick. This is what it thought it was made: "Make it more masculine."

Sam Parr

That's pretty funny.

Shaan Puri

So, imagine the lights go out. Now this is on **volume 100**, and our team is running out of the tunnel. Lights are flashing, and it's — that's our little... that's our little "intro sound" when we come out.

Sam Parr

Dude, are we gonna...?

Shaan Puri

Things like this make me say, "wow." I literally... this was, you know, like—I felt segregated. I couldn't participate in this field at all; I was on the outside. Now I can make things, and because it's so easy to make, I can create them for use cases that otherwise wouldn't have made sense. I wouldn't hire somebody to make this for me, but if it takes me two seconds, sure—then we'll do it.

Sam Parr

Are guys like **Rick Rubin**, who can't play music, creative geniuses? Are they, like, "I'm so happy this stuff doesn't exist when I was trying to convince everyone I was a guru, because now otherwise I'd actually have to go and make..."?

Shaan Puri

No, I think he's actually pretty big into this, and he's like, "Yeah — told you all: *taste is what matters*, right?" This has been his thing from the beginning. There's actually a very popular... I don't know. Have you listened to any *AI music*? Like... like, yeah, like...

Sam Parr

Well, I listened to something where there's a *cool* John Lennon song that wasn't finished when he died, and they had him sing it.

Shaan Puri

"That's cool. Yeah. There's a YouTube channel called *'Impossible Rap Songs'* and it's basically collaborations—Tupac and a rapper today, right? Tupac's dead, but they'd collab on a song. It's music that couldn't exist, but we make it exist. There are the first AI musicians that have gotten record deals; that just recently happened. I think one of the top country songs in the world right now is an AI song [I think it's called "Walk the Line" or something like that]. So **AI music** is a thing. I've been using *Suno* a lot. I won't bore you with my Suno track—I'll send it to you after this. I make a lot of music on Suno now and it's incredible fun. If you've ever had any level of a music itch, you gotta get on Suno. Forget that little beat-maker thing I showed you in Muse—Suno is the real deal where you can actually make good songs. If I'm working out, we no longer listen to mainstream music. It's all music we've made for ourselves—me and my training. We have an hour-long playlist every day and we can each just add to it, so I'll just add, play the...

Sam Parr

Top one: "Play the most" — list your most-listened **Suno** song. I mean, I used **Suno** the week it came out. You can't tell it, for example, "sing Frank Sinatra like this," but you can describe a Frank Sinatra song or a Frank Sinatra voice, and it will create something very similar, right?

Shaan Puri

Alright, I'm gonna give you some choices. Do you want to listen to something that kind of makes you feel like you're on drugs—maybe like you're in that one show, what's it called... *Euphoria*? Alright, do you want the "Euphoria" song, the indie singer‑songwriter song, or a fun kind of rap song? Which one do you want?

Sam Parr

I have to go with *Euphoria*, okay.

Shaan Puri

Alright, this is the *Euphoria* track. It's called...

Sam Parr

**Deshaun James**

Shaan Puri

"Close your eyes. You're in the club. Here we go."

Sam Parr

This is awesome.

Shaan Puri

Alright. Now my sister is moving to Spain, so you could just make songs for friends about whatever is going on — a birthday party or whatever the situation is. I set up this song called "America's Finished."

Sam Parr

Alright, let me show you something. Can I show you something? Yeah — so, Sean, let me ask you a question: what is the reason why you don't shop online more?

Shaan Puri

I don't know what's going to look good on me.

Sam Parr

Okay, well—I'm half solving for that. I love buying stuff online. The problem is that your boy's got big old thighs, and finding clothes that fit is quite challenging. I don't know—did you know this? A lot of people don't know this (I actually don't know how common it is), but when you buy clothes there's this thing called the **size guide**, and you can see the measurements of the garment.

Shaan Puri

Yeah, well... every site has this. But the problem is they're like, "Oh, is your clavicle four inches?" I'm like, "I don't know—how would I know those measurements?" This is not helpful to me.

Sam Parr

So here's what I did: I went to a bunch of websites and found the most common measurements that they ask. *Oh my...*

Shaan Puri

Ask you is incredible.

Sam Parr

I created a photo where I measured all the common parts of my body—**thigh, knee, calf,** and **sleeve length**. For a long time, when I was going to buy clothing from a high-end store, I would send this to them. Occasionally they'd say, "Just send me the right stuff."

Shaan Puri

You sent them this, yes?

Sam Parr

To a.

Shaan Puri

"High-end store, yes."

Sam Parr

And it was *incredibly effective*.

Shaan Puri

"What did they say?"

Sam Parr

They're like, "I suggest that you buy. We've been given..."

Shaan Puri

"This, to security: you will not be allowed to use... [ending unclear]"

Sam Parr

Are now on a list. I'm telling you—it's weird, but it's been *very effective*. Very strange, but it's been effective. Now, here's what I did: I used Cloud Code first of...

Shaan Puri

"All—*no way*. You measured all these things yourself? You had a friend, or Sarah, who measured these for you. You can't physically measure all these things with one hand like this."

Sam Parr

No. I had to have my wife just sit there and do it, and it was a pain in the butt. For a long time I kept this in my **Google Drive** and used it as a reference when I would go buy certain clothing. The problem was that it was cumbersome — kind of a pain in the butt. So what I did was I made an app using **Claude Code** where I would enter all of those measurements. Then I added two more things. The first thing was a link to the product. For example: let's say I want to get a dress shirt. I have to go somewhere and wear a tie, and I need a dress shirt. Right now I'm not sure if I'm a Large or an Extra Large because I'm right in between sizes, but the product has a size chart. What I built was an app where you take a photo of the size chart and you also get the link to the product, so you know it's for this style of shirt which should fit a very particular way. You enter it in here. When I posted the link, I just took a screenshot of the size chart, and it tells me — based off of these measurements — I should get a 16?

Sam Parr

Five-inch shirt — here's my reasoning. In order to train it, I actually used a bunch of books on how clothing should fit. I uploaded those to this [tool] to be reasonably good at picking which sizes to buy. I've been using this, and it's been really effective. In fact, I buy a lot of stuff off eBay because I think it's fun. I like finding vintage clothes, and many eBay sellers include measurements in the listing. The problem with vintage clothing is that labels like *Large* or *Extra Large* often don't mean much, so you have to go by the measurements. The same issue happens with European sizing: *Large* and *Extra Large* don't always work. Japanese sizing is different too — I actually wear **3X** in Japanese sizes — so you have to follow the size guide. I made this [tool], and it's been incredibly effective. I've been using it for a couple of weeks, and I've ordered three items based on the size it recommended.

Shaan Puri

"Wow — **incredible**. It is incredible what you've made. So, you're actually using this to buy stuff?"

Sam Parr

Yeah. Would you think a guy wearing overalls and a Midwest trash trucker hat would be trying to find the right-fitting shirt so there's no collar gap when I wear a tie? You probably wouldn't. But that's called being *eclectic*, my friend. It's kinda... you're...

Shaan Puri

"Like, what are those foods—turduckens—or, you know, foods that look like one thing on the outside but are wrapped and, on the inside, are a whole different animal that you didn't even know was there? That's what you are, my friend. You're a *turducken*."

Sam Parr

It's pretty awesome. I think if you are a *clothing nerd*, you guys will understand what I'm talking about. Or if you're in between different sizes, it's kind of a pain in the butt to figure out.

Shaan Puri

So, are you going to **publish these**, or are you just like, "I'm just gonna **use these myself** — I don't need to make this available in [unclear: 'dude main']"?

Sam Parr

**Anthropic**—the API stuff is really expensive. I used about $40 this weekend just on using it. It was really expensive. But I'll use it on *eBay*: if you go to eBay and you want to buy something old or vintage, it tells you the measurements. One night of doing that costs about $3. So no, I'm not going to publish it. But maybe... I don't know. Maybe I could charge for it. I guess that'd be a cool project—just try to build an interface that you could charge for.

Shaan Puri

Wow — okay, this is cool. Alright, my turn. I have a more serious business use case, and this will kind of blow your mind for people actually in business. High level: I have a company that’s growing really fast. It scaled from zero to tens of millions in revenue in about two years. Normally, to do that you’d have to hire a lot of people, right? It’s a services business: you have a lot of clients, and those clients need to be happy or they’ll leave at the end of their contract. So how do you actually manage all of this? Our team built an internal tool that does the following. It basically reads into all your systems. For example: “I’ve got a connection to **HubSpot**, I have access to **Fathom** video recordings from all your sales calls, I have access to **Slack**, and I have access to your accounting software.” That means the tool knows when bills get paid — it can read anything inside. That’s the first thing it does. Now check this out: this is our dashboard. Any member of our sales team or customer-success team can go in to look at a customer. When you open a customer record you immediately see everything — the overall status, when they started, who’s working on it, the link to their project plan. It auto-generates all of these things; nobody has to manually enter anything. The tool automatically goes and grabs all that information.

Sam Parr

Was this like a *CRM* that you built, or is this like a...?

Shaan Puri

We built this tool. Now, if I click the **renewal** (which is important), it will show: - Here's the current **ACV**. - Here's the target **ACV**. - Based on all the conversations we've had, it reads the last status update and the last client call, analyzes the sentiment, and says—based on the sentiment—the client is really happy. We have high confidence they're likely to renew. - Then it shows the adjusted **NRR** we expect: it'll be a **7% expansion** based on what we see right now. That's the first thing you see, which is *sick*. There’s also an **agent** we created so you can just talk to the contract, the deal, or the customer record. For example, you can ask: "What are the top three expansion opportunities that we should be pitching them? What other services might they be interested in based on what you know about them?" The agent will look at public information, call transcripts, and whatever internal data we have. It will cross-check our internal database of services we offer and say, "We think it could offer X, Y, Z." Then you can say, "Cool—can you make a to-do list or create an action/task for me?" It will update automatically in our task manager with who’s going to follow up and when. It creates the task, and now that task is sitting there so anyone in the company who looks at this customer can see the conversation and benefit from it. If someone goes to the customer record, they’ll see that somebody has already brainstormed expansion opportunities and that a to-do has already been created—someone's already on that.

Sam Parr

Has that been—have your customers' employees found that to be *productive*, or is it *buggy*? Does it work well enough that it adds a lot of value?

Shaan Puri

Yeah, this is our **command center** for the business. There's one guy who built this whole thing in, like, three months.

Sam Parr

There's so much information and so much stuff happening — it's so fast. I'm *very* thankful I'm not 22 right now. I'm *very thankful* I'm in the position I'm in now to be able to use this stuff.

Shaan Puri

But it would be easier if you were 22. It'd be easier than ever to be successful.

Sam Parr

The competition's *way* harder.

Shaan Puri

But the opportunity is kind of like *musical chairs*. If you're playing musical chairs and suddenly five new players get added, they double the number of players in the game. Yeah — it gets way harder. But if you 50x the number of chairs and you add 50 people, it doesn't matter. There are still so many chairs to sit in. That's **AI**. AI is like every single product in every single category, plus a whole slew of new categories, just came up for grabs because AI changes what can be done. The opportunity set is so much bigger that, yes, there's more competition, but it doesn't matter. You would rather have this huge multiplier on the number of possible ways to win and how easy it is to build than fewer people.

Sam Parr

And every company just needs one of these — one of these **AI** guys. You need many; you need many of them in the company.

Shaan Puri

I talked to a guy yesterday who runs a $1 billion company. I asked him, "Are you interested in—what are you doing with AI?" because I know he's smart and I figured he'd be doing something. He said, "Yeah, I'm playing. I'm constantly messing with cloud code and figuring out what I can do." He basically just appointed an AI general manager. He said, "I actually just—like—I hired, not a human. I created an AI that could be a manager for these projects. I'm trying to see how good I can make that." He told me that just by improving customer support with AI, making the developers more productive, and getting rid of his junior programmers who are not very good at using AI, they doubled their profit margin. They're already a billion-dollar company, so imagine doubling your profit margin off something like that—it's insane. Here's my last perspective that I think is helpful, because I feel like you came in excited and left stressed out. I think if you put on yourself that you need to become a *top 1%, 5%, or even 10%* AI builder/user, that's the wrong frame. All you need to do is say, "Alright, I'm already top 10% in business and in content. I've already done all that in my life. Now, if I just get to about the 50th percentile in AI—which is not that hard to do—it multiplies against all my other skills." The wrong game to play is: "Let me go try to beat all the AI geniuses who spend all day doing this and are more technically oriented than me." The right game to play is: "I already know a lot. For other people who have my skill stack or my domain knowledge in an industry I've been in for ten or twenty years, how do I use just a little bit of AI? How do I get *just good enough*, *just dangerous enough*, where it multiplies against what I've already got?" If you compare yourself to people who are technically brilliant or who are 22 and can learn anything super fast, you'll feel behind—but that's not the accurate way to think about things.

Sam Parr

"Alright, what do you think? Did you have fun?"

Shaan Puri

"That was fun. I think we should do this more often. I think we should use this as the *forcing function* to play with different **AI** tools and see what we come up with."

Sam Parr

I'm down — it was great. It's just crazy how there are **levels**: from buying the right-size overalls to adding 10 points to your $1 billion margin.

Shaan Puri

Let us know in the comments on **Spotify** and **YouTube**. We read all the comments. Tell us if you want more episodes in this style, or if you have a tweak that would make these more fun or more interesting for you. We try not to bring on an **AI expert** every time because sometimes the *"blind leading the blind"* can help. We don't know too much—we could both be kind of dummies figuring this out together. I suspect that's actually a more fun and better way to learn about this stuff than when somebody who knows everything comes on and just spouts a bunch of jargon you don't really understand, leaving you feeling stupid. But maybe we're wrong. Maybe we'll mix in some experts as we go.

Sam Parr

Alright, that's it. *That's the pod.*