Tim Kellogg

Accountants Should Do Hackathons!

Everyone should do hackathons!

Here’s the thing, it’s so easy to make a software tool these days that the hardest part is finding problems worth solving. And it’s the accountants, executive assistants, sales people, etc. that are most familiar with the biggest opportunities a business can solve.

A hackathon is a goofy thing software engineers came up with 15-20 years ago. The idea is to build software tools just for the sake of building. No objective other than to build. Spend a night, a day, a weekend, and just build.

The crazy thing is how many very cool ideas come out of hackathons. Google embedded it into their culture, and thus we got Gmail, AdSense, and Google News. The thing is, most hackathons only produce personal tools or software dev tooling, because that’s all engineers know.

What Is Hacking?

Hacking is solving problems. Simple as that. Scratch your own itch.

I saw this post where a Russian political analyst was offering a tip for using ChatGPT for proofreading. I, an engineer, was alarmed at the massive waste of using a huge LLM for a task that can be done natively in Mac OS. His response was, “yeah, but I like the ChatGPT voice better”.

Screenshot of two Mastodon posts. The first is from Kevin Rothrock, discussing the use of text-to-speech for proofreading with a demonstration of ChatGPT reading negative, I am a meat popsicle. The second post is from Tim Kellogg, suggesting using a Mac terminal to achieve the same effect by typing 'say i am a meat popsicle or using pbpaste say if the text is on the clipboard.

That’s hacking!

  1. He had a problem
  2. He built a solution

If you want real business problems solved, enable the person experiencing the problem to solve their own problem.

In a corporate setting, if he couldn’t solve his own problem, he’d have to draw up a funding request with ROI figures and then gain alignment from a software team. But how can you calculate ROI before you even know if it works? This project simply wouldn’t have happened.

Product Management Shouldn’t Be a Job

Ah, there’s a hot take! “Product management shouldn’t be a job”. Alternately, “everyone should be a product manager”.

Product managers design products, in the abstract sense. The product is three things:

  1. A problem that a user experiences, fully validated and understood
  2. The solution, a fully verified approach to solving the problem that users agree works for them
  3. A growth hypothesis. How will new users find and adopt this product?
graph LR Problem((Problem))-->Solution((Solution))-->Growth((Growth))-->Problem

When an accountant cobbles together a ChatGPT prompt to solve a problem, they’ve already fully completed step #1 without even thinking about it. Of course it’s a problem, otherwise they wouldn’t try to solve it.

If it works, then great, that’s step #2. If not, then iterate until it does work. Or give up, that’s fine too when you can iterate quickly.

The growth hypothesis is trickier, but it usually boils down to, “how are people going to use it?” In the screenshot above, Kevin blasted out his prompt in a post. That’s a growth hypothesis, and he probably convinced someone to give it a try.

The ChatGPT Store is another option. It’s a decent option because it gives you more tools, like access to data, plus it just plain seems to work better than a in-line chat prompt. Microsoft Copilot is similar. Both of these give you tools for getting it in front of your peers.

Everyone could be a product manager, why aren’t they?

Go Forth And Build

If your a decision maker in IT, give people access to the tools.

If you’re a normie, you might struggle with knowing what to build. That’s a common problem! A decent place to start is

  1. Name 5 pet peeves about your job
  2. Think about how you could solve them given your tools (ChatGPT, Copilot, etc.)
  3. Tweak and iterate

And do it together. That’s what hackathons are. People honestly do have trouble getting started with solving their own problems. Hackathons are a social way of helping each other figure it out. It’s just a bunch of people in a conference room that say, “stop! we’re going to block off 4 hours to hack. also here’s some free beer”.

Got an idea? Ping me. I’d love to help you figure it out too.