YouTubeFeed

How this “non-coder” used Cursor to add AI to retro hardware

Summary

Maddie Reese, a self-described “non-coder” turned vibe coder and hardware tinkerer, joins Claire Vo to show three physical projects that bridge software and the analog world. Her dad introduced her to Lovable about a year ago, and after locking herself in her room building things, she started wondering what else — beyond software — had suddenly become accessible. The answer was hardware, and it snowballed from there.

The episode walks through three builds: a tiny thermal receipt printer wired to a Raspberry Pi and a Convex database that lets anyone in the world send Maddie a physical message via a website (maddiereese.com/message); a hacked 1990s pager sourced through an “authorized reseller” of the last surviving paging network (Spok) and rigged to buzz whenever her Twitter account gets an interaction, via a chain of Cloudflare Workers, Resend, and Gmail; and a “personal API” that exposes her coffee order, pet names, favorite snacks, time zone, and favorite restaurants so friends — or eventually agents — can do something thoughtful without having to ask.

Maddie’s workflow for each project starts the same way: dump the idea into Cursor’s clean agent view, let it interview her with questions until a plan and shopping list emerge, then triple-check every hardware purchase before buying because a wrong part means a useless project. She’s candid that the code behind these hacks is often “not practical” — routing a tweet notification through four services just to light up a pager — but insists that’s fine, because the goal is joy and physical connection, not elegant engineering. Claire draws out the through-line: AI is lowering the barrier to hardware just as it did for software, letting people who could never solder or read a datasheet build tangible, delightful things anyway.

Highlights

”I’ve always felt jealous of people who could code”

Maddie's AI pill moment

“Kind of my journey was that I always felt jealous of people who could code. And then when I finally was able to kind of join that world, I just started thinking, oh, well, what was inaccessible to me that’s completely accessible now? And hardware was one of those things. So I just got really, really into it, starting with my printer, and it kind of snowballed from there into a bunch of weird things.” — Maddie Reese, 2:02

Clip command
yt-dlp --download-sections "*2:02-3:00" "https://www.youtube.com/watch?v=KCGKb3huDsY" --force-keyframes-at-cuts --merge-output-format mp4 -o "jealous-of-coders.mp4"

The “Print Me a Message” thermal printer demo

Print Me a Message live demo

“It was something I knew couldn’t be too hard, right? Like people use printers all the time. That’s how printers work is you send something from your computer to print out, right? So I knew it couldn’t be too difficult.” — Maddie Reese, 4:23

Clip command
yt-dlp --download-sections "*4:23-5:22" "https://www.youtube.com/watch?v=KCGKb3huDsY" --force-keyframes-at-cuts --merge-output-format mp4 -o "print-me-a-message-demo.mp4"

The pager mafia — buying from pagers123@gmail

The pager reseller

“Their website looks like it hasn’t been updated since like the late 90s, honestly. And that was the most modern looking one I could find, that’s why I went with them. And it truly was… it truly was something to behold. Their email address, I believe, was pagers123@gmail.” — Maddie Reese, 12:43

Clip command
yt-dlp --download-sections "*12:43-13:34" "https://www.youtube.com/watch?v=KCGKb3huDsY" --force-keyframes-at-cuts --merge-output-format mp4 -o "pager-mafia-reseller.mp4"

”It’s fun” — throwing out practicality

Not practical, but fun

“I don’t care if it’s not like—of course it’s not practical to like send it to Cloudflare and then resend and then proxy it through a Gmail and then send it to the official—like, of course that’s not practical. But guess what? It’s fun. It’s fun, and so I think your like idea of just throw out the idea that this has to be like practical or even like reasonable code and just build the thing that you want, especially if it’s for fun, I think is really exciting.” — Maddie Reese, 16:05

Clip command
yt-dlp --download-sections "*16:05-16:54" "https://www.youtube.com/watch?v=KCGKb3huDsY" --force-keyframes-at-cuts --merge-output-format mp4 -o "its-fun-not-practical.mp4"

The personal API — reducing the friction of being thoughtful

The personal API pitch

“This is, this is a way to kind of reduce that friction of like, oh I want to do something nice but I don’t want to bother them so I just won’t do anything at all. In this case, you can just be like oh I want to do something nice, I’ll hit the API, I’ll get their coffee order and then we can meet up. Right?” — Maddie Reese, 19:06

Clip command
yt-dlp --download-sections "*19:06-20:20" "https://www.youtube.com/watch?v=KCGKb3huDsY" --force-keyframes-at-cuts --merge-output-format mp4 -o "personal-api-pitch.mp4"

Key Points

  • Origin story: jealous of coders (2:02) - Maddie always felt locked out of coding; once it became accessible she immediately asked what else — like hardware — was now within reach
  • Dad introduces Lovable (3:10) - Her father, already deep into AI, showed her Lovable about a year ago; she spent days locked in her room building things
  • The thermal receipt printer (4:06) - A tiny thermal printer that fits in her hand, wired up to accept messages from strangers online
  • Raspberry Pi + Bluetooth + Convex architecture (6:33) - The printer connects via Bluetooth to a Raspberry Pi on her desk; every submission also logs to a Convex database even if the physical print fails
  • The Cursor brainstorming workflow (6:33) - She dumps the raw idea into Cursor and has it interview her with questions until she has a full implementation plan and shopping list
  • Buying hardware on impulse vs. need (6:33) - Sometimes she buys a retro device first (like her iMac G3 off Facebook Marketplace) and figures out the project later; other times the project comes first and dictates the purchase
  • Triple-checking hardware purchases (7:51) - Unlike software mistakes, a wrong physical part means an unusable object sitting around, so she always double- or triple-checks Cursor’s shopping suggestions
  • Cursor has been wrong before (9:30) - Mostly minor wiring mistakes, always caught during the triple-check step, never a major failure
  • The printer’s second life via OpenClaw (9:46) - Claire’s household uses OpenClaw-driven printing for a daily kids’ newspaper and one-off explanation sheets — “the revenge of the printer”
  • 90s-childhood pager project (11:08) - Built to give her (and Claire’s kids) a more analog alternative to constant device notifications
  • Spok: the last paging network (11:54) - Pagers can only be bought through authorized resellers, typically ones that sell to hospitals — not consumers like Maddie
  • The Twitter-to-pager pipeline (11:54) - Twitter API event → Cloudflare Worker → Resend email from a custom domain → Gmail → pager’s restrictive USA-something.net email address → the page finally displays
  • “Not practical, but fun” philosophy (16:05) - Maddie explicitly rejects the idea that hobby hardware code needs to be clean or reasonable — the goal is joy, not elegance
  • Zen “agent view” vs. Claire’s “aggressive Cursor” (17:41) - Maddie starts every project in Cursor’s clean, distraction-free agent view for brainstorming, only opening terminals/browsers later, unlike Claire’s everything-open approach
  • “Enough to do some damage” (18:19) - Maddie compares her code literacy to picking up conversational Spanish growing up in San Diego — she can follow along without being fluent
  • The personal API concept (19:06) - Born from the friction of wanting to do something thoughtful for a friend (their coffee order, their favorite restaurant) without having to ask and feel like a bother
  • What’s actually in the API (20:41) - Coffee order, upcoming San Francisco travel dates, pet names (Mabel, Lola, and Moon), favorite snack, time zone, and favorite restaurants
  • Agents, not just humans, will call the API (21:52) - Claire imagines telling her own OpenClaw to hit Maddie’s API and book a reservation at her favorite restaurant automatically
  • Adding POST endpoints (22:47) - Claire suggests the API shouldn’t be read-only — it should let people post directly to Maddie’s pager or printer
  • Lightning round: the dream build (23:59) - A dot-matrix printer for banners, though Maddie hasn’t settled on exactly what she’d use it for yet
  • Lightning round: prompting when AI misbehaves (25:34) - She troubleshoots first (is it her or the AI?), then asks it to double-check and confirm intent — staying calm because “it’s not conscious yet”

Mentions

Companies

  • Cursor (2:02) - The AI coding tool Maddie uses for every project’s brainstorming, planning, and shopping-list phase
  • Lovable (3:10) - The tool Maddie’s dad introduced her to about a year ago that kicked off her whole AI journey
  • Spok (11:54) - The one surviving commercial paging network in the US; pagers are only available through its authorized resellers
  • Cloudflare (11:54) - Workers receive Twitter API events as the first hop in the pager notification chain
  • Resend (11:54) - Email service that relays the Cloudflare Worker’s event to a Gmail address in the pager pipeline
  • Google (Gmail) (11:54) - Intermediate hop that finally forwards to the pager’s restrictive email address
  • Twitter / X (11:54) - Source of the interaction events that trigger pages
  • Convex (6:33) - Database logging every “Print Me a Message” submission, even if the physical print fails
  • Anthropic (24:19) - Team behind a Monet exhibition in San Francisco that hooked a typewriter up to AI, cited as a kindred retro-tech mashup
  • OpenClaw (9:46, 21:52) - Claire’s household automation tool, used for printing a daily kids’ newspaper and imagined as the future caller of Maddie’s personal API
  • Facebook Marketplace (6:33) - Where Maddie bought her iMac G3 on impulse before deciding what to build with it
  • Amazon (7:26) - Referenced as the fast-shipping source for hardware shopping-list parts
  • Firecrawl (1:14) - Sponsor; web data API for AI agents
  • Customer.io (14:29) - Sponsor; AI-generated customer engagement campaigns

Products & Technologies

  • Thermal receipt printer (4:06) - Tiny hand-sized printer at the center of the “Print Me a Message” project
  • Raspberry Pi (6:33) - Connects to the thermal printer via Bluetooth and runs the project’s website/backend
  • iMac G3 (6:33) - Retro computer Maddie bought impulsively, to be repurposed for a future project
  • Twitter/X API (11:54) - Fires an event for every interaction on Maddie’s account, kicking off the pager pipeline
  • Personal API (19:06) - Maddie’s own endpoint exposing coffee order, pets, snacks, time zone, and favorite restaurants
  • Card-puter (Code with Claude event) (5:35) - Small device Claire’s nine-year-old claimed, now routed through a Mac Mini to Telegram as a messaging gadget
  • Dot-matrix printer (23:59) - Maddie’s stated dream hardware build, envisioned for printing banners

People

  • Claire Vo - (all episode) — Host of How I AI; founder of ChatPRD
  • Maddie Reese - (all episode) — Guest; vibe coder and hardware tinkerer behind the thermal printer, Twitter pager, and personal API projects; also an advisor to a startup she previously ran

Surprising Quotes

“A problem I never anticipated us girls would have in 2026 is where to clip my pager.” — Claire Vo, 17:14

“It’s kind of like how growing up in San Diego, you kind of, you pick up a little bit of Spanish and can kind of understand it, but you can’t like really speak it? That’s kind of where I’m at with code right now.” — Maddie Reese, 18:19

“If I’m playing like futurist, I think the reality is is agents are going to be doing a lot of this like kind of work for us. Like I can imagine my OpenClaw where I’m like, ‘Hey, I had such a great podcast with Maddie. Go hit her API, figure out when she’s coming to San Francisco next and let’s make a reservation at her favorite restaurant.’” — Claire Vo, 21:52

“I don’t normally get too upset because I recognize that it’s not like conscious yet.” — Maddie Reese, 25:34

“You seem very chill. You’re gonna be fine in the inevitable AI uprising. They are not gonna come for you.” — Claire Vo, 26:38

Transcript

Maddie Reese: 0:00 I’ve always felt jealous of people who could code. And then when I finally was able to kind of join that world, I just started thinking, what else was inaccessible to me that’s completely accessible now? And hardware was one of those things.

Claire Vo: 0:14 Can you pull up your Print Me a Message?

Maddie Reese: 0:17 Here is the webpage where people can send me a message. You just go to maddiedrees.com/message, and you fill in your name, you fill in whatever you want to send, and then print the ticket, and it prints out on my desk.

Claire Vo: 0:30 I love it. So we’re going to test it. I have your Message Maddie website up. I’m going to click print ticket. There you go. Oh my god, it’s so fast! So walk us through how we might build something like this. Welcome back to How I AI. I’m Claire Vo, product leader and AI obsessed, here on a mission to help you build better with these new tools. Today, I have a really fun episode with Maddie Reese, who’s building the hardware projects of my dreams using some retro gear that you might not expect. If you’ve ever wondered, what’s my friend’s coffee order? And then thought, it would be amazing if they had an API that told me, this is the episode for you. Let’s get to it. Quick word from today’s sponsor, Firecrawl. If you’re building with AI agents, you’ve probably hit the same wall. Your agent needs data from the web, but the right pages are difficult to find, buried in JavaScript, or blocked behind logins. Firecrawl is a web data API that lets agents search, scrape, and interact with the world at scale and get back clean, structured data they can actually use. Over a million developers, including myself, build on it. It’s open source and it’s free to start. Stop fighting the web for data and start powering your AI agents and apps with Firecrawl at firecrawl.dev. Use code HOWIAI to get 10,000 free credits today. Maddie, welcome to How I AI. I was so excited when you agreed to hop on the podcast because I think you do some of the funnest projects in AI that I’ve seen lately. And what I love about what you’re doing is you’re intersecting software builds with hardware builds in a way I personally have felt has been really inaccessible until about the last six months. So what brought you to all these fun hardware projects?

Maddie Reese: 2:02 Kind of my journey was that I always felt jealous of people who could code. And then when I finally was able to kind of join that world, I just started thinking, oh, well, what was inaccessible to me that’s completely accessible now? And hardware was one of those things. So I just got really, really into it, starting with my printer, and it kind of snowballed from there into a bunch of weird things.

Claire Vo: 2:28 Well, I want to pause about what you said, which is you were jealous of people who could code. The code. And then all of a sudden that became accessible. What was your journey there? Just tell us, like, how you got from being jealous to hacking hardware projects, personally.

Maddie Reese: 3:10 Honestly, through my dad. My dad introduced me to Lovable about a year ago. He’s been really into AI for a while and he goes, ‘Hey, did you know this is possible now?’ And I was like, ‘No, I had no idea. Oh my god, what do you mean?’ I spent days just on Lovable, like, just locked in my room, just completely building things. And then I kind of started exploring, like, okay, what’s next? What else can I do with this? What, how can I expand my knowledge in this? And it’s been my whole life ever since.

Claire Vo: 3:48 I love it. Ever since the magical last year, you have completely changed your whole life. Well, show us the printer because I want one of these things. And we’ve got to show it on camera for people that can’t see. Can you describe it?

Maddie Reese: 4:06 It’s just a little mini thermal printer, a thermal receipt printer. It’s really tiny, it fits in my hand.

Claire Vo: 4:14 And can you pull up your ‘print me a message’ flow and just walk us through like what was the idea and how did you end up building this?

Maddie Reese: 4:23 Here is kind of the webpage where people can send me a message. You just go to maddiereese.com/message and you fill in your name, you fill in whatever you want to say, do a CAPTCHA here, and then print the ticket and it prints out on my desk. Honestly, I just wanted some way to interact with people that was kind of more physical because everyone that I would have been talking to about these kind of things I knew online, like everyone I’d been talking to about Vibe coding and AI, apart from my dad, I’d been talking to them online and I wanted some way to kind of bring that into the physical world. It was something I knew couldn’t be too hard, right? Like people use printers all the time. That’s how printers work is you send something from your computer to print out, right? So I knew it couldn’t be too difficult.

Claire Vo: 5:11 I love it. So we’re going to test it. I have your message Maddie website up. I’m going to click ‘print ticket’. Oh my god, it’s so fast!

Maddie Reese: 5:22 It’s so fast, I know.

Claire Vo: 5:25 Right from the front door! There we go. It says ‘Hey girl, hey’.

Maddie Reese: 5:30 Thank you very much. And you’re welcome for joining How AI, thanks for having me.

Claire Vo: 5:35 That is so adorable! I went to the Code with Claude event and they gave these little card-puters. And my nine-year-old immediately snagged it. And we’ve been trying to figure out a way for it to be like a messaging device and we hooked it up, it like runs through a Mac Mini in our house and then it federates out to Telegram for us. But how cute would it be for him to be able to like type little notes to me down in my office? I just think that’s so cute. And so I

Maddie Reese: 6:00 I have to ask, are you getting a lot of messages?

Claire Vo: 6:02 I’m getting a ton of messages. I’ve gotten messages from all over the world. People on the side of mountains, like then just telling me like, ‘hey, this is where I am, this is where I’m sending a message from’. People from boats or ships in the middle of the ocean. A ton of countries. Yeah, I’m getting tons of messages and it’s so cool.

Maddie Reese: 6:25 Can you pull up the GitHub and just kind of show us a little bit of how it works or tell us a little bit of the prompt that you gave it so we can understand how how the system kind of flows?

Claire Vo: 6:33 Well, I have a Raspberry Pi on my desk over here and then it’s connected to this to the printer via Bluetooth. And there it has the the website here and then there’s also a Convex database. I have it set up so every time someone presses this print ticket button, even if it doesn’t go through to the printer, it still goes to my database so that I still have a log of like what people tried to send me, which has been helpful because actually the printer, this is my second version, because the printer just broke a few days ago because it had been on continuously since the end of October. How I how I kind of started with the with the entire project is again, I just dumped everything out from my brain into Cursor. I’ll actually kind of demo it for you now. But really what I just kind of did was I went like, okay, I want to create a printer where anyone in the world can send me messages from a website. How best would I implement this? Let’s brainstorm. So from there, I typically go on like a brainstorming session. Typically I have it ask me questions and I just have it keep asking me questions until all of my questions are answered about what how this works, how I can do this. Then we kind of zero in on like what I should buy if I’m starting from something where I don’t already have the thing. Sometimes I’ll like make a purchase, like I have my iMac G3. I just bought that off of Facebook marketplace and decided I’ll figure out something to do with it later because I just wanted it so badly. But in this case, I had the project first and then I bought the printer. So it just kind of depends on the workload.

Maddie Reese: 7:26 I I love it. And so you’re really using this process of just ‘this is my grand idea, interview me, help me understand.’ And then because it’s hardware you’re like, ‘I don’t know what I need, like just give me a shopping list. I’ll buy it. It’ll be here in a day or two. Thank you Amazon.’ and then we’ll go ahead and hook it up. And you’re you’re proxying all this through just a Raspberry Pi that’s sitting on your desk.

Claire Vo: 7:51 Yep. When I’m buying pieces of hardware, I make sure to double or triple check. Like I’ll say like, ‘okay, well what if what if this happens? What if this happens?’

Maddie Reese: 9:00 Went like, ‘Are you sure this is still the the correct piece of of thing to buy?’ Sometimes it isn’t, so, but when you’re making making kind of physical purchases, I don’t want to just have something around that like I won’t be able to use. Like what’s the point of that? So I make sure to always triple-check and make sure that I actually will be able to use this thing that I bought.

Claire Vo: 9:21 So have you have you been led astray by your purchasing? Like, has has Cursor ever told you to buy something that doesn’t work?

Maddie Reese: 9:30 A couple times, but I’ve always caught it um through the the triple-checking process, which is which is great. And it’s never been like a big thing. It’s mostly just been like, ‘Oh, you’ll need these kind of wires.’ And then I’ll like look and I’ll be like, ‘I don’t I don’t think that’s correct.’

Claire Vo: 9:46 So you’re a you’re a real trust but verify kind of girl. I love that. Well, so you had this like cute printer, which I think is adorable. I really want one. What I was telling somebody the other day, I was like, ‘This is the revenge of the printer right now.’ because I print so much stuff through my OpenClaw. So I get my OpenClaw, my husband gets my OpenClaw to like print a daily newspaper for our kids. We’re printing like one-off explanation sheets. We are just like printing up a storm. But I love the idea of this like little form factor of this like tiny cute receipt printer. And then how satisfying is it to like rip that piece of paper off?

Maddie Reese: 10:24 So satisfying. So satisfying. Yeah.

Claire Vo: 10:31 And I have I have executives that I work with on a coaching basis and just help them think about AI, and they all want this receipt printer on their desk so they can get their agenda for the day like printed off in the morning, and they can just like rip it off. Or their to-do list like just ripped off. I just think it’s so fun to be able to get the like non-digital form factors so accessible using using AI.

Maddie Reese: 10:54 Exactly. And see now it’s sometimes it takes it takes a little bit of kind of like that tedium of of like ‘Oh, here’s your here’s your daily to-do list.’ and you get to add a little little fun thing to it, it makes it feel not as bad. So.

Claire Vo: 11:08 I agree. Let’s have like a little bit more fun with AI. And so speaking of fun, you have done the thing that I actually really want to do, which is, you know, all of us parents on in my in my age bracket are trying to give their kids what we’re calling like 90s childhoods, um where they’re like not on the laptop and they have just a little little more analog life. And when I was a kid, I had a pager. And you have built a Message Maddie pager. Um and so can you walk us through because you said before we joined the show, you said there was like a little bit of trickiness to getting the right pager and there’s like a pager mafia or something that you have to get through. So walk us through how we might build something like this.

Maddie Reese: 11:54 To set the scene, there is kind of one paging network left. So pagers go over like radio transmission through on specific networks. So there’s this one paging network and it’s called Spok. And that is where there might be some other smaller ones, but they’re the big one. That is where all of your pages will likely get sent through if you decide to buy a pager and get it set up with this network. But they don’t sell like the pagers themselves. You have to go through like authorized resellers, and that’s what I did for this. So I bought it through a authorized reseller who typically sells to hospitals, hospitals and doctors and things like that. So they weren’t quite accustomed to dealing with me. Their website looks like it hasn’t been updated since like the late 90s, honestly. And that was the most modern looking one I could find, that’s why I went with them. And it truly was… it truly was something to behold. Their email address, I believe, was pagers123@gmail. Got the pager finally. It comes set up with a network, it comes set up with an email address and a phone number, which is great. So I have it set up kind of as a Twitter pager. I use the Twitter API to send notifications to my Resend email, and then it sends it to a Gmail account because the email address that comes with the pager doesn’t accept new custom domain emails because it’s… it’s very outdated. USA-something.net is the domain for the email address they gave me. They’re very restrictive email. But the Twitter pager, I’ll kind of read through it here. So every interaction that anyone has with my Twitter account or my tweets gets sent as an API event to a Cloudflare worker. The worker does some things, sends it through Resend as an email. Resend sends the email from a custom domain to a Gmail address, and then the Gmail address sends it to my pager email address. So I have to go through multiple steps to get it, but then after that is when it finally displays on the pager. Or…

Claire Vo: 14:24 Oh my gosh. I have to… so while you’re telling this, I have to pull up and see if I can make this work.

Maddie Reese: 14:29 Yes, we’ll see, we’ll see. Sometimes the service is a little spotty. That’s the only issue I’ve really run into because it’s… the pager network is just not super big. But we’ll see, we’ll see if it pulls up. This episode is brought to you by Customer.io. You’re here because you’d rather use AI than talk about it. With Customer.io, you describe the campaign you want to build and the AI agent creates it for you: the audience, the messages, and the timing. You review it, make any changes you…

Claire Vo: 15:00 And launch. Instead of spending hours stitching together tools and workflows, you can focus on the work that actually drives growth. Every campaign is tied back to results, so you can see what’s working and what to do next. More than 9,000 brands use customer.io to turn the data they already have into messages customers remember. Visit customer.io/howiai to try it today. Customer.io: more impact from every message. And what I love about this story that you told me is that you’re like, this thing is so outdated and so hard to hack, but I know it’s possible to get from Twitter to this pager. I may have to like traverse a bunch of very weird things, I may have to do weird things, but it’s okay to do weird things because AI is coding it and it’s fine. All I want is the ultimate kind of like endpoint. And so what I love about your approach is saying like—

Maddie Reese: 16:05 I don’t care if it’s not like—of course it’s not practical to like send it to Cloudflare and then resend and then proxy it through a Gmail and then send it to the official—like, of course that’s not practical. But guess what? It’s fun. It’s fun, and so I think your like idea of just throw out the idea that this has to be like practical or even like reasonable code and just build the thing that you want, especially if it’s for fun, I think is really exciting. Okay, I sent you an app message. We’ll see if it comes through. I will also send you a little page through your through your website and we will see if it beep—does it beep? Like does it literally go like beep beep?

Claire Vo: 16:41 Oh it does! I have that turned off because it’s annoying, but it does.

Maddie Reese: 16:43 Now what’s nice about the pager is like you can carry it around, right?

Claire Vo: 16:48 Yeah. Uh-huh.

Maddie Reese: 16:49 Do you strap it on like your like your jeans? Walk around?

Claire Vo: 16:54 So it has the you know, the little little thing, the little clip, but um, I have been discovering a lot of my pants don’t have belt loops.

Maddie Reese: 17:07 What’s the deal? So like that was my plan. Um and then I like put on my pants to get dressed for the day. I’m like where do I clip this on?

Claire Vo: 17:14 Where do I clip my pager on? Exactly! A problem I never anticipated us girls would have in 2026 is where to clip my pager. Walk me through why you use this like agent view? Because I am like in opposite cursor. I am in terminals open, codes open, files are open, browser open like, I am in aggressive cursor version. You are in like very Zen cursor version. Explain to me why this appeals to you.

Maddie Reese: 17:41 I want just this one view where I can just kind of brainstorm and get all of that through. Like if I don’t need to be seeing anything else yet, um I don’t want to have it up because it kind of distracts me. But later on in the process, I’ll definitely pull up some terminals, dep—depending on what I’m working on, um or a browser, anything like that. Um, but just for this starting kind of though, I want to just have it kind of clean and be able to just go, okay, let me, let me get through my brainstorming, let me get a solid plan for this, and go from there.

Claire Vo: 18:14 I love it. And then are you reading the code? Or you’re just like, if it pages, I’m happy.

Maddie Reese: 18:19 Yes and no. I mean, I don’t fully understand it. I’ve learned so much, um, more like I have a ton more knowledge than I did at like the beginning of this journey, and I can kind of understand some things. It’s kind of like how growing up in San Diego, you kind of, you pick up a little bit of Spanish and can kind of understand it, but you can’t like really speak it? That’s kind of where I’m at with code right now.

Claire Vo: 18:43 I actually really love that, that comparison. You know what? I think it’s good enough to be able to, for stuff like this, like just grok it a little bit. You know? Just understand a little bit. Enough to do some damage is what we want.

Maddie Reese: 18:52 Exactly.

Claire Vo: 18:53 Well, we’ll wait and see if you can like, you know, wave the pager by the window and see if it gets serviced. No big deal that it didn’t go through, but I want to talk about your third project, which I also love, and you have to tell us about the problem statement of the Maddie API, because I think it is adorable and such a good idea.

Maddie Reese: 19:06 Thank you! Okay, so kind of how I started out with this is I had this idea about personal APIs. And that stemmed from kind of this problem that you have where you, you want to do something nice for someone, buy them a coffee or buy them their, take them to their favorite restaurant and make a reservation, but you don’t really remember like what their coffee order is, or what their favorite restaurant is, and you want to do something nice, you don’t want to bug them about it. Um, so I was like, well, what if there was a way where you could just hit an API and you get all of this information about them? I kind of see it as a way to like bring people closer together. Like my initial thought was like, oh, that’s like too clinical and too, too weird. But then I was like, well no, this is, this is a way to kind of reduce that friction of like, oh I want to do something nice but I don’t want to bother them so I just won’t do anything at all. In this case, you can just be like oh I want to do something nice, I’ll hit the API, I’ll get their coffee order and then we can meet up. Right?

Claire Vo: 20:20 I mean, you’re speaking to the one person in the, on the earth that’s like of course that makes so much sense! Don’t page me and make me call you and tell you your coffee order. I love it. Okay, so, so you have an API endpoint and like tell me what you have decided belongs in your personal API.

Maddie Reese: 20:41 Okay, so we have the basic stuff, we have my coffee order, very important. And then we can kind of go through here. So we have dates I’ll be in San Francisco, currently has a question mark ‘cause I don’t know when I’ll be back. My pet info, this is very important, so it has my pets. We have Mabel, Lola, and Moon. Very important, if you know me. Let’s do favorite snacks. This is very important. My Orville Redenbacher simply salted popcorn coated in specifically Bragg’s nutritional yeast. Delicious. Time zone, this one’s kind of just a nice-to-know kind of thing for when people are scheduling meetings.

Claire Vo: 21:22 What’s your favorite restaurant in San Francisco? I gotta see.

Maddie Reese: 21:25 Oh, yes. Okay, I have two. So it’s called one’s Rocket Sushi and it’s literally just a like a sushi place where the food gets served on a rocket conveyor belt. They have um the most delicious, I think it’s called Hiyayakko. It’s cold tofu with um some kind of some sort of like cold broth. It depends on what it is, but yeah, it’s delicious. And then Saibo’s Front Room, which is just like a topping food place.

Claire Vo: 21:52 I love it. And you know what I love about this idea of the API? You’re very adorable in that you think that, you know, humans are going to hit this API. But if I’m playing like futurist, I think the reality is is agents are going to be doing a lot of this like kind of work for us. Like I can imagine my OpenClaw where I’m like, ‘Hey, I had such a great podcast with Maddie. Go hit her API, figure out when she’s coming to San Francisco next and like let’s make a reservation at her favorite restaurant.’

Maddie Reese: 22:23 Exactly.

Claire Vo: 22:24 That all just the the agent can do it and then if you combine that with some of the like commerce stuff, the agentic commerce stuff people are doing, it’s like a very interesting flow and it’s giving me some ideas of like what could be an API. And again we’ve like moved to this world where everything is a CLI, everything’s an API, so like why not why not yourself?

Maddie Reese: 22:46 Exactly.

Claire Vo: 22:47 You know what you need? You need to add a post to your API where we can post to your pager or post to your to your printer.

Maddie Reese: 22:52 Oh, you’re right. You’re right. That’d be exactly.

Claire Vo: 22:54 It doesn’t have to be read-only. Not only gets. We need we need some posts.

Maddie Reese: 22:57 Exactly.

Claire Vo: 22:58 I love this so much. This is super cool, Maddie. I just I appreciate just to recap your story for people: like jealous of people who can code, Dad says ‘Hey, have you heard of Lovable?’, you go super deep, it opens up your mind and then you’re like imagining just fun things that you can surround yourself with that spark joy. And what I love about what you’re doing is like it’s not disconnecting you from people. It is actually creating connection points between you and and the world in a really human way. I sent you a very sweet note, I sent you a pager, like I’m going to page you every now and then because I think this is so fun and it’s just this like really novel way to come up with new connection points and also be creative, which is one of the things that I just love the most about about AI. Well, let’s get you to two lightning round questions and then we will get you back to your projects. My first one is if you had or could build any dream hardware project, what’s on your mind right now?

Maddie Reese: 23:59 I want a dot matrix… sprinter, the ones that can do like the big banners. I’m envisioning something where people send me banners, but um, no. So I’m gonna have to come up with something with something else um to do with it, but I know I want one. So, one of these days I’m gonna get it.

Claire Vo: 24:19 I mean, it’s just the sound is so so so enjoyable. Is this zzz zzz zzz. So you’re like really into these like retro hardware pieces where you just want to like surround yourself with like kind of fun retro pieces but then intersect them in in a new way. I know that um the Anthropic team, there was a a Monet exhibition here in San Francisco and they like hooked up the typewriter.

Maddie Reese: 24:41 With the typewriter.

Claire Vo: 24:43 And I just think those things are so delightful. Oh my god, yeah. Okay, so we’re gonna we’re gonna have to brainstorm something for your dot matrix printer. Um, you know what you should print you should have cursor print out an explanation of the code it wrote for you so you can like put it on and read it. Something fun, I don’t know, something fun. Or banners, I don’t know, banners could be really great. Every meeting you could print out a banner and then pin it up behind you on like what the topic is. Like, hey Claire, welcome to How I AI. It could be very very hilarious. Okay, second question I ask everybody, um, you seem very very sweet, but when AI is not giving you what you want or it’s sending you down a hardware path you do not enjoy, what is your response? How do you prompt?

Maddie Reese: 25:34 Just telling it, you know, like, are you sure this is right? Yep. Um, like I need you I need you to double check. And then if that if that doesn’t quite quite get there, um, well first I do a lot of like um troubleshooting. I I like to get deep in there and be like, okay, well is it is it the Metrons or is it is it the AI? Sometimes it’s me. Sometimes it’s it’s the AI, um, but so once I get past that, then I go through like, okay, I need you to double check. I need you to make sure that this is this is what you’re intending to do. I need you to make sure that this aligns with what we’re doing. I don’t normally get too upset because I recognize that it’s not like conscious yet. I don’t know, I don’t know how this is going. Um, but I do get kind of like, okay, like come on. Come on. So, but that’s typically where where it kind of plateaus. I’m not a very angry person so…

Claire Vo: 26:35 Okay, so you just kind of get like minorly frustrated.

Maddie Reese: 26:36 Yeah.

Claire Vo: 26:38 You seem very chill. You’re gonna you’re gonna be fine in the old you know in the inevitable AI uprising. They are not gonna come for you.

Maddie Reese: 26:44 Thank you. Thank you. That’s a great compliment.

Claire Vo: 26:46 Well Maddie, this has been so fun. Where can we find you and how can we be helpful to you?

Maddie Reese: 26:50 Yeah, so I am at Maddie D Reese um pretty much everywhere. My website is maddiedreese.com. On Twitter I’m Maddie D Reese. Um, I’m Maddie D Reese pretty much everywhere I can get it. Please don’t steal it if I

Claire Vo: 27:00 I haven’t gotten the summary yet. So that’s all I ask. If it’s home.

Maddie Reese: 27:03 Yeah, I’m really just exploring kind of everything right now. I mean, we didn’t get into it, but I have a startup, um, but I’m kind of in an advisory position, um, kind of left one of my other businesses too, to just explore a lot in the AI world. So that’s just what I’m interested in doing. If anyone has any, any opportunities for like exploring stuff, I would love to chat.

Claire Vo: 27:29 Amazing. I love it. Well, Maddie, thank you for joining How I AI. This has been really fun.

Maddie Reese: 27:34 Thank you! I appreciate it. This was great.

Claire Vo: 27:36 Thanks so much for watching. If you enjoyed the show, please like and subscribe here on YouTube or even better, leave us a comment with your thoughts. You can also find this podcast on Apple Podcasts, Spotify, or your favorite podcast app. Please consider leaving us a rating and review, which will help others find the show. You can see all our episodes and learn more about the show at howiaipod.com. See you next time.