The New Era of Personal Software

Howdy y’all 🤠. It’s been a while since my last blog post and felt time for another.

The other day, I found myself asking a question most people wouldn’t even consider asking or trying to answer: How many meetings have I had with external contacts over the years? Who have I met the most? When did I last meet them? But instead of shrugging it off, I did what I’ve been doing with a lot lately—I quickly built something to find out.

30 minutes with Cursor to build a script

With some assistance from the AI code editor Cursor and a few basic Python tweaks (though, honestly, you can accomplish this without any coding knowledge), I quickly put together a script. This script:

Pulls in Google Calendar events
Parses attendees and counts their occurrences
Uses GPT-4 to extract names and companies from attendee emails

Ultimately, it generates a neat CSV files displaying who I’ve met, how often, and when I last met them. Feel free to use it yourself! Code and readme here.

As a lowly tinkerer, this project would have otherwise taken me a couple hours—beyond what’s feasible. But in < 30 minutes, it was done! With continued advancements in LLMs, I would guess this will take 1-2-minutes 1-year from now.

Other Cursory experiments

Again, this is just one example. I’ve been leaning into building lots of simple tools to scratch random itches. A couple others:

AI Running Plan Generator

This fall, I was training for an ultramarathon with some friends (a 6-man 200-mile relay in Kentucky called Ragnar that we won btw!). I struggle to stick to a running plan, but I do appreciate knowing what to do next based on my goals and Strava history. Typically, all available plans are generic static PDF documents. So, I decided to create my own tool—a generator that takes your distance and pace goal, along with your strava data, and feeds it into GPT to create a customized go-forwards training plan. This allows me to recreate an AI-generated running plan after each run, adapting to progress that naturally deviates from the original plan. If you’re curious, you can see it in action here.

Sneaky Scrapers

Data scraping has always been somewhat of a dark art, especially on websites with stringent bot detection. To address this, I’ve built a variety of scrapers that mimics human behavior to slip through unnoticed.

Entering era of “Personal Software”

At a YC event I attended a couple of months ago, Replit’s CEO, Amjad, discussed the concept of an era of “personal software”—where individuals build their own tools to solve highly specific problems.

Recently, I met a solo entrepreneur who purchased a golf app from Acquired.com. He’s been developing various pieces of personal software to manage his business. For instance, he discovered that off-the-shelf AI customer service solutions didn’t effectively answer his customers’ questions. By feeding his 100,000-line codebase into Gemini’s 1M context window, he was able to create his own customer service bot that perform far better.

These are just a couple examples. X is ablaze with more.

In this new world, I think we’ll start to see ‘software’ become the new amazon product (ie, way to many options!). And in that world, distribution becomes the bottleneck to success. I listened to a podcast this summer with one of the founders of Whop, a ‘software’ marketplace with affiliate marketing built-in, where he talked the personal software era and how they’re built for it and it really resonated (wish I were an investor!). Increasing importance of distribution is why everyone wants to be a creator!

Give it at least a Cursory shot ;)!

There’s something incredibly satisfying about taking an idea and bringing it to life in less than an hour. Once you’ve done it a few times, you become “a hammer that sees hidden nails.” Frankly, it’s so easy that even an 8-year old can do it. Highly recommend watching the embedded ~10 minute video by the way - hilarious and inspiring at the same time.

So, what are you going to build?

As always, if you enjoyed the read, please feel free to share with friends. And let me know your thoughts/reactions/questions and if/what you’ve been building lately — I would love to hear from you. You can reach me at altmbr@gmail.com or on twitter @altmbr. And for anyone new that wants to get future posts in their inbox 👇

Bryan