Why food discovery is broken — and what we're doing about it

The average person eats roughly 1,000 meals a year. Of those, most people rotate through the same 10–15 restaurants. Not because they don't want variety — but because the cost of a bad meal (time, money, disappointment) is high enough that they default to the familiar.

We built Taste Twins because we believe that's a discovery problem, not a preference problem. People want to eat well and eat diversely. They just don't have a tool that makes it safe to try something new.

What existing apps get wrong

Star ratings aggregate the opinions of everyone — including people whose palates are nothing like yours. A 4.2-star Thai restaurant might be perfect for someone who loves bold, aromatic, high-spice food and completely wrong for someone who prefers clean, delicate flavors.

Proximity-based recommendations assume you want what's nearby. But the best meal of your life might be 45 minutes away.

Trend-based recommendations assume you want what's popular. But popularity and personal fit are almost entirely uncorrelated.

The social graph we're missing

The most reliable food recommendation you'll ever get is from a friend who eats exactly like you. Not a friend who's a "foodie." Not a friend who's been to a lot of restaurants. A friend whose palate overlaps with yours in a specific, measurable way.

That's what Taste Twins is building: a social graph organized not by who you know, but by who eats like you. Your Flavor Twins are your most trusted food advisors — and the algorithm finds them for you.