Awesome AIWhat do you think will the future of image generators be like?
We are always talking about those upcoming new things like txt2video, txt2objects and stuff, but what do you think will Image generation be like in 3 years BESIDES resolution and generation time? Do you think we will still use webUIs with custom community checkpoints? Will it be perfectly coherent? Will prompt crafting still be a thing?
6 comments
johnAug 1 agoText-to-3dprint with preview first before printing something.
bobAug 1 agoI think you'll see what happens when big companies have real product teams working on this.
You won't have cobbled together webui stuff you have to download from a repo (good luck driving adoption for non technical people if you're not in an app store).
You'll have a polished device agnostic or mobile native app with all the ugly under the hood stuff abstracted away into the cloud.
Interfaces will be a combo of text prompts (those less complex than they currently are) and tools that allow you to rapidly plow through visual choices to refine elements of the image.
Like, think clicking on the eyes and seeing a radial menu with several variations instantly populating. You select one and it instantly inpaints it.
Payment will be a monthly fee or usage based depending on how constrained the app is. My guess is it will be monthly so users don't need to think about costs. Likely tied to some cloud storage service they charge for (GDrive, Adobe creative cloud etc).
Aa2d agoLodo BC