I will test out DALLE-3 version available to me upon release.
If it can consistently return a polygon from its description for upto and including octagons, then this resolves to yes. If it needs a bit of nudging like specific prompt wording but can still get it right, then also its a yes.
I will try it 10 times, and if it gets it correct 50% of the time for each of the polygons, then it counts for that polygon.
10 times here means 10 prompts. If there are 4 different images to each prompt, we're testing on 40 images.
Related questions
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ285 | |
| 2 | Ṁ225 | |
| 3 | Ṁ37 | |
| 4 | Ṁ30 | |
| 5 | Ṁ27 |
@Shump "up to and including octagons" seems pretty clear to me, what's the argument for excluding them?
@IsaacKing The argument would be that a septagon is probably much harder to generate than an octagon. I'm also checking because it's a possible difference between your market and this one. Your market excludes esoteric ones, which I would take to mean septagons at least. Or did I misunderstand your description?
@Shump I'm confused by how you can look at me talking about all numbers from 3 to 8, and conclude that 7 is not included. Is there a different dialect of English where "from X to Y" means something else?
@IsaacKing Because your market says "Common Polygons" and "It's ok if it fails on more esoteric ones". Idk about you but I would consider septagons to be esoteric. Also, stop being passive aggressive
@IsaacKing I had to! your original polygon market for any AI image generator was the one that inspired mine in the first place.