| ▲ | vunderba 3 days ago | |
From the article: So maybe some people hate AI because they have an artist's eye for small inadequacies and it drives them crazy. This is it 100%. When somebody draws something (in an active fashion), there is a significantly higher level of concentration and thought put towards the final output. By its very nature, GenAI is mostly using an inadequately descriptive medium (e.g. text) which a user then must WAIT until an output that roughly matches your vision "pops" out. Can you get around this? Not entirely, though you can help mitigate this through inpainting, photobashing, layering, controlnets, loras, etc. However, want to wager a guess what 99% of the AI art slop that people throw up all over the internet doesn't use? ANY OF THAT. A conventional artist has an internal visualization that they are constantly mentally referring to as they put brush to canvas - and it shows in the finer details. It's the same danger that LLMs have as coding assistants. You are no longer in the driver's seat - instead you're taking a significantly more passive approach to coding. You're a reviewer with a passivity that may lead to subtle errors later down the line. And if you need any more proof, here's a GenAI image attached to _Karpathy_'s (one of the founding members of openAI) twitter post on founding an education AI lab: | ||