multimodal ai laziness
Inspired by Sam's In defense of AI slop?.
AI and ads
Like AI content takeovers of social media, like ads. People don't use ad blockers, let alone think to use them. They don't reject social media services whose companies not only spearhead AI but also train with user data. Complaints peak on exposure then dissipate after the nuisance disappears. Is absorbing content and sharing misery this addictive that people willingly forgo baby steps?
AI art
I imagined a bell curve based from domain experts critiquing others' art in an AI art space.
Experts by social proof frown upon lazy users but for a different reason than non-users. They detect laziness most superficially via commonality and dependency. Think of standardized image resolutions or popular artistic styles like ChatGPT 4o's infamous yellow tint and Ghibli filter as some of the easiest examples of commonality. As for dependency, AI users across all modes tend to settle down with their favorite model, workflow, and settings. With guidance, bespoke and idiosyncratic setups can arise. However, like how people treat AI takeovers and ads, few will bother. Care to follow their lead?
Don't forget about purpose. Much more people could care less whether throwaway memes were AI generated compared to images meant to be taken seriously. How many creatives bypass or disavow AI to prevent audience backlash, only for it to backfire?
Even then, compared to a casuals versus speedrunners state of mind, both experts and novices teach and learn from each other regularly. Casuals with sheer luck created some of the wildest video game speedrun skips, which speedrunners then capitalized upon to achieve speedrun history.
AI music
Having not listened to much AI music nor produced any, I may benefit from higher exposure to detect laziness as above. Common gripes include off-putting voices and vocal melody giveaways that I hear even in songs that Sam referenced. Listening through contexts where it's disclosed as AI also doesn't help. Does a music arena website exist where users can vote on whether none, either, or both songs per round are AI generated or not? I want to fine-tune my ear like how AI artists fine-tuned their eyes above.
Some music producers showcase their workflows and lack of generative AI use. Ironic how their music sometimes returns first for AI music searches. If you want to try this yourself, I suspect less popular or more specific queries work better. Could such music be sought by ai more readily than other examples like websites using Not By AI badges?
Should process outweigh outcome for music? I frequent music I like the sound of. Liking the process adds to that, but doesn't make up for sounds I don't like. Who knows how many more songs we have the privilege of enjoying? I can afford to listen to and produce more of what brings me joy.
Want to reach out? Connect with me however you prefer:
- Email me via your mail client
- Copy my email address or remember it for later:
yoursimperfect@proton.me
- Email me via Letterbird contact form or open it in a new tab