Content
@
0 reply
0 recast
0 reaction
Snuts 🎩
@chd.eth
Did you noticed that DegenerHats are a bit different than other AI generated collections? How many varieties in texture and color there are, while keeping the same EXACT position and aesthetic across 10,000?! smol thread ⌄
3 replies
3 recasts
20 reactions
Snuts 🎩
@chd.eth
Well, you should have noticed, because it is very cool… and you can be cool too!
 We’re building a tool to let anybody ship similar collections with great ease - and more user control over what they get. Rough WIP:
0 reply
0 recast
3 reactions
Snuts 🎩
@chd.eth
I used the $Degen hat as a paradigm to rethink PFP collections and dive into multi-step prompts via API that you can only do with Glif. The hat allowed me to create fun PFP material for the new chain, while seeing what consistency I could extract from the process.
0 reply
0 recast
3 reactions
Snuts 🎩
@chd.eth
The workflow I created through Glif was: > Random color select > Random Texture Select > Random mood Select > Stable Diffusion prompt to combine > ControlNet conditioned on a Top Hat image Then off to GPT to check if it looks like a hat.
0 reply
0 recast
2 reactions
Snuts 🎩
@chd.eth
To get the extra special rare hats, I modified the above flow to allow custom text fields to enter in unique ideas, like the Nebulars and Crowns… there’s even a Whale and a DegeneRAT yet to be minted.
0 reply
0 recast
3 reactions
Snuts 🎩
@chd.eth
This process usually takes a TON of GPU and time running by hand. But I could do this through a single Glif API call quite simply. As we build out the app, I’d like to: - Remove the need to quality check each output - Keep random lottery feel, but give user some control - Make multiplayer, fun to do with others
1 reply
0 recast
3 reactions
