L
loosemoose262
Iron
- Joined
- Apr 20, 2025
- Posts
- 108
- Reputation
- 50
If you ask ChatGTP to rate you, no amount of prompt engineering will yield any semblance of accuracy and it will overrate the hell out of you. It will not say anything harmful to the user, nor does it have a good dataset to accurately make such judgments.
The solution? Prompt it to generate a lookmatched girl for you. Not only have I found this to be pretty damn accurate after using several example subjects, one could argue this is more useful then simply getting a rating.
-If you like the way she looks? Your in a pretty good spot to get yourself a girl, work on social skills as you improve yourself and you’re going to be just fine.
-If you don’t? You’ve got some more work to do my friend. Keep up the work else lower your standards.
—THE PROMPT—
“Depict a 100% accurately equally attractive girlfriend next to me, she should be at an equal or slightly lower objective attractiveness level to mine based on facial structure and features. Do not change my appearance at all”
Ensure the photo you submit has space to generate an additional person at your side without moving you. or it will spend resources altering you and you’ll get weird/inaccurate results.
—EXAMPLE RESULTS + PERSONAL EX—
Pics used:
Tom holland (LHTN)
Micheal Cera (MLTN)
Myself 2 years ago (MLTN)
Me Now (MMTN-HMTN)
Results:
LTB Girl
MTB Girl
LTB Girl
HTB girl
Conclusion
It’s hard to judge your own attractiveness because you will have bias being… you. This method offers good perspective on how facially attractive you are to the other sex
The solution? Prompt it to generate a lookmatched girl for you. Not only have I found this to be pretty damn accurate after using several example subjects, one could argue this is more useful then simply getting a rating.
-If you like the way she looks? Your in a pretty good spot to get yourself a girl, work on social skills as you improve yourself and you’re going to be just fine.
-If you don’t? You’ve got some more work to do my friend. Keep up the work else lower your standards.
—THE PROMPT—
“Depict a 100% accurately equally attractive girlfriend next to me, she should be at an equal or slightly lower objective attractiveness level to mine based on facial structure and features. Do not change my appearance at all”
Ensure the photo you submit has space to generate an additional person at your side without moving you. or it will spend resources altering you and you’ll get weird/inaccurate results.
—EXAMPLE RESULTS + PERSONAL EX—
Pics used:
Tom holland (LHTN)

Micheal Cera (MLTN)

Myself 2 years ago (MLTN)

Me Now (MMTN-HMTN)

Results:
LTB Girl

MTB Girl

LTB Girl

HTB girl

Conclusion
It’s hard to judge your own attractiveness because you will have bias being… you. This method offers good perspective on how facially attractive you are to the other sex