I used AI tools of our own making, QR codes and paper to get around Social Media, making money, promoting myself and other people's businesses.

- link is an infographic

Backstory:

We started an AI company (fieldgenie.ai) focused on using voice (ASR,TTS) and LLMs to make IRL data collection easy. I used our software to document the deployment of IRL posters, and have used the data to refine my approach to postering.

The runway was tight and I didn't know how to raise money. Somewhere in there I freaked about money and decided to make money by doing what I do best: solving complex problems. I started offering handyman services. But I had a problem – How do I find clients? FB ads take way too much work. All I got from Nextdoor was spam and corporations like Angie's List offering me predatory ad packages.

I call 'em Geotagged Posters

I used our app fieldgenie.ai to document where I put posters and set up a forwarding server, so when someone scans we know the poster got scanned. Boom, I got more handyman work than I could handle. People around here are drowning in infrastructure issues related to their single family money pits, and they need a general ops guy like me.

My friends liked the idea, so I offered them the service and have now deployed a lot of posters in my area. Over time I have built up a map of what locations are good and which less so.

Some Learnings:

Location quality has as much to do with the service being advertised as with the geography/environment. I do A/B testing on all posters. For example, you may be tempted to put up posters in high-traffic areas, but posters don't last long there. There's competition from other posters, and municipalities "keep things clean."

Certain side streets are better for getting more lifetime scans. High-traffic poles last 1-3 days before getting torn down. Side street poles last 2-4 weeks. Looking at the scan data, Monday-Wednesday have the lowest scans, about 30% lower than Thursday-Sunday. Peak scan windows are 12-2pm and 4-6pm. The utility pole outside a climbing gym (#1100363642, by the bike racks) consistently outperforms bulletin boards inside. One early lesson: what you want is small and simple QR codes. Large QR codes draw the eye but have a major flaw when deployed on posters: if the QR code is not perfectly flat, posted on a curved surface like a utility pole, then the depth of field means that people have trouble scanning them.

Ideally: --The QR code should be .5"-1" wide and placed 1" or more from the page margins. --The QR code should be placed in a larger design element with good spacing from text, ensuring good contrast.

I could go on for pages about my learnings, but I'll highlight how fun this whole thing has been. Money's great and all, but living an adventurous life is more rewarding. Isn't this why we do start ups? I bike around, scout and deploy posters, I learn little neighborhood secrets, I see patterns in how people travel and move about my environment.

I didn't get funding for fieldgenie.ai, I got too distracted with handyman work and postering, but AI has a lot of money in it right now so I decided to try my hand at job hunting. Sick of the absurdity of applying online, I made a poster with my resume on it and put it up in strategic locations. This has landed me 2 conversations, one of which has been mind-blowingly amazing.

This is all to say that I think real world ads are gaining in value. From a game theory perspective it's pretty obvious why. It takes real work to put them up, real people have to do it. AI is amazing tech, but it's being abused. As AI trash poisons the net, real world interactions gain value. I hope you found this interesting.