I got so tired of reading fake Google reviews to find dinner that I built an "anti-choice" restaurant app. Here's how I put it together
I hoped you might find this interesting.
Perhaps I was the only one who just could not decide where to eat. The risk of finding a decent place to eat while traveling was enormous. I never have enough money and I always want to avoid tourist traps.
I was standing on a street corner, hungry, scrolling through endless 4.2-star Google Maps spots, trying to figure out which reviews are paid for, which are bots, and which are from tourists who don't actually know what good food tastes like.
For instance tourists always post amazing reviews because they are more relaxed and on holidays.
Because my background is in location data, I know exactly how these map algorithms work—they are optimized for ad spend and foot traffic, not for a quality meal. There is a lot of money on seo.
I wanted to see if anyone else was feeling this specific type of BS paralysis, so I started building.
It is still early on and no revenue but here is the step-by-step of how I validated and built Hungry But Not Stupid (HBNS):
The Scrappy Validation (F5bot)
Before writing a line of code for the app, I needed to know if people were actually complaining about this. I set up F5bot to monitor Reddit travel subs for exact-match desperation keywords. I didn't use generic terms like "tourist trap." I tracked raw, human frustration strings like "tired of searching," "help me pick a restaurant," and "don't trust reviews." Result: My inbox absolutely blew up. The pain point was massive. People were having real-time meltdowns over decision fatigue.
Designing the rebel UX
Every food app today gives you more options. I wanted to build a brutalist tool that gives you zero options. I decided the UX had to be entirely frictionless and the responses almost narcissistic, like from a good food critic.
No top 10 lists or bs from international sites.
No reviews to read no opinions.
No social feeds.
Just one button that gives you one definitive local mandate. You go there, or you figure it out yourself.
The MVP and Domain
I took a domain I already owned—hungrybutnotstupid.com—and spun up the MVP. The name itself does half the marketing. It sets the tone immediately...imo this isn't a polite lifestyle blog; it’s a rescue tool. Whatever to me it works.
Manual Data Curation (The bottleneck)
The easiest thing to do would be to hook up the Google Places API, but that immediately corrupts the product. It would just feed people the same SEO-gamed garbage they were trying to escape. Instead, I started manually verifying tight data corridors. I focused on the highest-stress areas: major Italian transit hubs (Rome, Florence, Milan) and Swiss zones like Bern, Zurich, and Interlaken. I had to build a whole geolocation api to get this to work, and that was another service entirely with 100gb of data (bwendi.com).
Where I need your feedback:
I'm at a crossroads with scaling the data. F5bot is handing me highly qualified, desperate leads every day, but the manual curation is heavy.
For those of you who have built highly curated or "opinionated" products:
How did you scale your dataset without resorting to the mass-market APIs that ruined the ecosystem in the first place?
Have you ever built a product that intentionally restricts user choice? How did users react to the friction?
Also wow to everyone making so much revenue. What am I missing?
Would love to hear how you'd handle these issues