Blog
Notes from the build
The case against manual calorie logging
Manual calorie logging is dial-up internet — it works, but no one would design it from scratch in 2026. Here's why AI input has crossed the threshold, and where database-search apps still hold on.
AI calorie tracker accuracy — what's real, what's marketing
"97% accurate" is a marketing number, not a benchmark. Here's what accuracy actually means in AI calorie tracking, why consistency beats precision, and how to evaluate the claims.
Habit-forming, not habit-shaming — what calorie apps got wrong about behavior change
Streaks, red numbers, and "you went over" backfire. The behavioral science says calorie tracking should feel like brushing teeth, not like a daily report card.
The 3-second test — how an AI calorie tracker should actually work
A simple benchmark for any AI calorie tracker — if logging takes longer than your first bite, the app failed. Here's what actually meets the bar, and what most "AI" trackers still get wrong.
Why I keep deleting calorie tracking apps (and what AI finally fixes)
I've installed and deleted every calorie tracker on the App Store. The pattern wasn't willpower — it was friction. Here's what AI actually fixes about logging food, and what it doesn't.
I benchmarked Cali against 200 real meals
How accurate is an AI calorie tracker, really? I ran a public benchmark on 200 photographed meals. Here are the honest numbers.