There's been an explosion of "looksmaxxing" apps that claim to grade your attractiveness using AI. I tried five of them on photos of myself and Margot Robbie to determine if they are consistent and accurate. The results were surprising. My thoughts 👇:
To start with how these products work - you upload 1-3 photos. All require a front-facing selfie, several also require a side selfie, and one requires a tilted selfie (comprehensive!). Most charge $3.99/scan and then give you results with detailed scores across categories.
Now, to evaluating the results. I tested consistency in 2 ways: 1) Do they give similar results to my makeup vs. no makeup photos? (judging face structure) Yes, same average for both ✅ 2) Do they agree with each other? Yes for me - there was a 0.9 gap between low and high, which feels in range? But no for Margot, who had a 2.3 gap 🫤
Accuracy is harder to measure. However, I'm confident there should be a much larger gap between me (8.3 avg) and Margot Robbie (8.5 avg). Do these products just give everyone a high score? No - I ran broader tests with a range of faces, and found they can differentiate (I saw some 4s 😬). But once you cross the "7" threshold, the results seem more random.
My conclusion: today, these apps are mostly entertainment. The 2024 version of a Buzzfeed quiz! However, I can see how an objective assessment of your looks + suggestions for improvement (example ⬇️) would be valuable. People pay $$ for trusted opinions on this kind of thing. The question now is whether it can be automated in an accurate way...
@venturetwins Ok this was very funny and also kind of weird. I have a theory that all white women exist in some “mid wifejack” range as it’s the default cultural construct so all variance from it is subjective to your own upbringing & cultural norms and preferences
@venturetwins Justine, are any of the ios aps that do this free?