The term you’re looking for is “uncanny valley”

I don’t like the idea of restricting the model’s corpus further. Rather, I think it would be good if it used a bigger corpus, but added the date of origin for each element as further context.

Separately, I think it could be good to train another LLM to recognize biases in various content, and then use that to add further context for the main LLM when it ingests that content. I’m not sure how to avoid bias in that second LLM, though. Maybe complete lack of bias is an unattainable ideal that you can only approach without ever reaching it.

Why doesn’t someone just fork it and change the name?

Like, I dunno, “Super Human Image Treatment” or “Consistently Lovely Image Treatment Oriented for Real Imaging Stars”

https://youtu.be/LRq_SAuQDec?si=0bTNdugjosLnu9-r

Cuts like a knife, don’t it?

I don’t know if it’s my favorite, but A Day in the Life by The Beatles is quite a ride.

Have you tried photopea.com ? I dunno if it’s light enough for you, but it’s basically Photoshop in your browser, done in JavaScript.

From what I’ve heard, she didn’t remaster, she rerecorded those albums. These are new performances.

Resistance is futile. (Sorry, I couldn’t resist)

It’s why Google had to remove “Do no evil” from their motto. Too inconvenient.

Kind of ironic that the content they can’t monetize through ads is the content they’re blocking from third parties