Yeah, this is the problem with frankensteining two systems together. Giving an LLM a prompt, and giving it a module that can interpret images for it, leads to this.

The image parser goes “a crossword, with the following hints”, when what the AI needs to do the job is an actual understanding of the grid. If one singular system understood both images and text, it could hypothetically understand the task well enough to fetch the information it needed from the image. But LLMs aren’t really an approach to any true “intelligence”, so they’ll forever be unable to do that as one piece.

Same honestly. And if I ever ask a question that someone might think is a duplicate, I link to that question and say something like “I found X, but the answers here don’t reflect Y”.

Surprisingly legible, but feels like I can only read it with momentum, flitting past it and letting my subconscious tell me where the word breaks are. The moment I get confused and look more closely, it becomes almost impossible to read.

Exactly the mistake threads just made, trying to capitalize on twitter’s rate limiting fiasco. The “general public” is extremely fickle, and Reddit will give us more opportunities.