How I AI: Maddie Reese on Building a Pager, Printer, and Personal API with AI

In this episode, I was so excited to sit down with Maddie Reese, a builder who is creating some of the most fun and inventive AI projects I’ve seen. What I love about her work is how she combines software and hardware in a way that, until recently, felt completely out of reach for most of us. She’s taking digital interactions and giving them a physical, tangible form, reminding us that AI can be playful, personal, and a source of genuine human connection.

Maddie’s story is one I think will resonate with a lot of people. She told me she had always been jealous of those who could code. It felt like a locked door. Then, about a year ago, her dad introduced her to AI tools, and that door was blown wide open. Suddenly, she was building things she never thought possible, diving headfirst into the world of hardware hacking, not as a seasoned engineer, but as a curious and creative explorer armed with tools like Lovable and Cursor.

Her journey is a testament to how AI can level the playing field, turning grand ideas into reality without years of traditional coding experience. In our conversation, Maddie walked me through three of her amazing projects: a thermal receipt printer that anyone in the world can send a message to, a resurrected 90s pager that delivers her Twitter/X notifications, and a personal API that catalogs her real-life preferences for friends (and their AI agents). These aren't just quirky gadgets; they're new channels for connection, built with creativity and a little help from AI.

Workflow 1: Building a Physical Inbox with a Receipt Printer

Maddie’s first project tackles a simple, universal desire: to make online interactions feel more real. She wanted a way to bring the conversations she was having with people online into her physical space. The result is a tiny thermal receipt printer on her desk that instantly prints messages sent from a public webpage.

Step 1: Brainstorming with an AI Coder

Instead of starting with a complex technical plan, Maddie began with a simple brain dump into Cursor, her AI code editor of choice. Her initial prompt was wonderfully direct and outcome-focused:

I want to create a printer where anyone in the world can send me a message from a website, how best would I implement this? Let's brainstorm.

From there, she used an interactive process, asking the AI to interview her and ask clarifying questions until the entire plan was mapped out. This collaborative approach helped her refine the idea and identify all the necessary components without getting bogged down in technical details upfront.

Step 2: Sourcing the Hardware

Once the concept was solid, the AI helped generate a shopping list. Maddie’s main components were a mini thermal receipt printer and a Raspberry Pi to act as the brains of the operation. She shared a crucial tip for anyone working with hardware: trust, but verify. Before buying anything, she triple-checks the AI's suggestions to make sure the parts are compatible and correct, a step that has saved her from purchasing the wrong components a few times.

Step 3: Assembling the System

The final architecture is elegantly simple:

1. A public webpage at maddiedreese.com/message provides a form for anyone to type a name and a message.
2. When a user clicks "Print Ticket," the message is sent to a Convex database (for logging) and to a service running on the Raspberry Pi on her desk.
3. The Raspberry Pi connects to the thermal printer via Bluetooth and prints the message out on a receipt.

The result is almost instantaneous. I sent her a message during the show, and it printed on her desk in a second. It's incredibly satisfying!

The Outcome: A Joyful Connection Point

Maddie has received a ton of messages from people all over the world—from mountaintops to ships in the middle of the ocean. The project has been so popular that the first printer actually broke from being used continuously for months! It's a perfect example of how a simple piece of hardware, powered by AI-generated code, can create a delightful and tangible connection to the global community.

Workflow 2: Resurrecting a 90s Pager for Twitter Notifications

If the printer was about tangible connection, the next project is about the pure joy of hacking and embracing the impractical. Inspired by the 90s-style childhoods many parents are trying to recreate, Maddie decided she wanted a pager. Not just any pager, but one that would buzz with her Twitter/X notifications.

Step 1: Navigating the "Pager Mafia"

Getting a pager in the 2020s is not straightforward. The industry runs on a single major network called Spok, and you can't just buy a pager off the shelf. Maddie had to go through an authorized reseller—a company that typically sells to hospitals and doctors, with a website that looked like it was from the late 90s and an email address of pagers123@gmail.com. It was an adventure in itself, but she eventually got the pager, complete with its own phone number and email address.

Step 2: Architecting a Rube Goldberg Machine for Notifications

Getting a modern notification from the Twitter API onto a piece of retro hardware required a hilariously complex chain of services. As Maddie explained, "it's okay to do weird things because AI is coding it and... it's fun." Here’s how the data travels:

1. An interaction on her Twitter account (a reply, like, etc.) triggers an API event.
2. The event is sent to a Cloudflare Worker, a serverless function that processes the data.
3. The worker formats the notification into an email and sends it using Resend.
4. Here's the trick: The pager’s email service is so old it won't accept emails from new custom domains. So, Resend sends the email to a modern Gmail account.
5. The Gmail account then forwards the email to the pager’s ancient, restrictive email address (USA*.net).
6. Finally, the message appears on Maddie's pager.

This workflow is a masterpiece of creative problem-solving. It’s not practical, but it works, and it highlights a key aspect of building with AI: you can chain together whatever tools you need to get to a fun outcome, even if the path is a bit absurd.

Workflow 3: Creating a Personal API for Real-Life Interactions

Maddie's third project, the "Maddie API," might be the most forward-thinking. It started with a simple problem: you want to do something nice for a friend, like get them coffee or make a dinner reservation, but you can’t remember their order or favorite restaurant. You don't want to bug them and ruin the surprise, so you end up doing nothing.

Her solution? An API for her personal preferences. Instead of asking her, a friend (or an AI agent) could just hit an endpoint to get the information they need. What started as a potentially "too clinical" idea quickly revealed itself as a way to reduce social friction and enable more acts of kindness.

What's in the API?

Maddie has populated her API with all sorts of useful, and charming, data points:

• Coffee Order: An essential for any social API.
• Pet Info: Details about her pets, Mabel, Lola, and Moon.
• Favorite Snacks: This gets specific: Orville Redenbacher's simply salted popcorn coated in Bragg's nutritional yeast.
• Time Zone: A practical endpoint for scheduling.
• Favorite SF Restaurants: Including One's Rocket Sushi (where food arrives on a rocket conveyor belt) and Abel's Front Room.

The Future: An API for Agents

While Maddie's idea is sweet, I think the true power lies in what's next: AI agents. I can easily imagine telling my own agent, "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." The agent could handle the entire flow. We even joked that she should add POST endpoints to her API, allowing someone to directly send a message to her printer or pager. Everything is an API, so why not yourself?

Making AI Fun and Human

Maddie's projects are a fantastic reminder that AI development doesn’t have to be about building the next big SaaS platform. It can be about sparking joy, creating novel ways to connect, and simply having fun. Her journey from being "jealous of coders" to confidently hacking together hardware and software is an inspiration.

She proves that with today's AI tools, anyone with a curious mind and a fun idea can build amazing things. It’s about being creative, embracing the weird and impractical, and seeing where these new capabilities can take you. I left our conversation buzzing with ideas, and I hope you will too.

---

Find Maddie

You can find Maddie and her projects on her website, maddiedreese.com, and follow her on Twitter/X at @MaddieDReese.

Episode Links

• Listen on Apple Podcasts
• Listen on Spotify
• Watch on YouTube