I had a few people ask me last week what exactly happens at a hackathon - given I was in San Francisco, natural curiosity. Long story short, you're in a room with a bunch of other nerds and you build something from scratch in a day (or less).
There's a few examples on AWS' Builder Loft calendar, but let me give you the skinny on how to be very, very dangerous in just a few hours with no advance notice.
Namely: idiosyncrasy. Domain knowledge in particular is highly difficult to replicate and builds a great story. Several of my hackathon projects were around exposing and using United States patent office data - a format I am both intimately familiar with, and have personally engaged with (given I've filed my own provisional patents).
Sitting next to me at one hackathon, there was an attendee by the name of Naol; he came to SF wanting to build a coffee empire and his own shops. He was debating between a coffee grading app (i.e. take a picture of coffee beans and have them examined/graded), or a logistical analysis app to monitor coffee supply chain issues (and price spikes when he buys coffee).
Guess which one he chose?
If you guessed the logistics analysis app, you were correct. It makes the stronger story, requires more domain knowledge, and importantly, it's something that nobody else can replicate without the same amount of time or knowledge investment as him. (He's taking the selfie)
Really, the best way to build something is to 1. examine what unfair advantage you already have (there is no need to reinvent the wheel) and 2. find out what off-the-shelf tools are available to do so. I could yap your ear off all day about the optimal ways to use Claude and Codex and all sorts of product engineering tools for neat apps, but only you know what makes you special.
If you measure what you think about the most, and build around that thing, you will find greater success as a baseline.
Speaking of patent data: I'm looking at putting out a second newsletter that looks at big tech patents and filings from big companies. I haven't gotten around to publishing it because of bandwidth, but reply to this email with what you'd like to know about big tech (or anything else).
Be well,
Michael Kirsanov