@corbeaucat

If you go to other sites like Tumblr, there is a very zero-tolerance, moralistic attitude toward generative AI. It feels very moral panicky and many people are proud to say they never used it intentionally. This goes for both LLMs and generative images. I think this is a rather extreme take. LLMs can be useful if approached properly. I find Claude is best for this approach where you ask a question, then ask for sources, look at the sources and check and then bring up more and more things. The result is an active dialogue rather than uncritically accepting what the LLM says. Generative images are something that is best used for quick, cheap things like roleplaying game character portraits. They can also be used as reference for art like lighting reference and testing color schemes. It’s like previewing an idea before going forward with it. Do note most of what you get from generators like Wombo is rather frustratingly generic and often not what you visualize if you have a specific vision in mind.

1 reply

0 recast

2 reactions