Live data from Hacker News

The Lone Banana Problem in AI

digital-science.com

1–10 of 109 posts

Re: The Lone Banana Problem in AI

#3

> A single banana casting a shadow on a grey background That prompt works fine in DALL·E 2.

It will likely be fixed in the next iteration of MidJourney as well.

But the point isn’t about specific prompts where the error is obvious and relatively easy to fix. This is just a n extremely easy example to understand.

The point is about the subtle biases that are much harder to detect, and which will therefore go unfixed.

Re: The Lone Banana Problem in AI

#4
Seems like we need more actor crific models that evaluate theae outputs on real world physical modeling and accuracy and not just quality of artistic output or token similarity.

But also Midjourney in particular seems trained to be more evocative / stylistic rather than photorealistic or precise.

Re: The Lone Banana Problem in AI

#6
I've seen lots of long prompt suggestions, however single words are fascinating, some words are extremely heavy, for example, we might assume 'topic' and 'subject' are somewhat interchangeable, but when I'm talking to an LLM swapping those words cause a huge change. Same with names, if you ask it to act like Dave, it's not some random personality, it's a distillation of every Dave it has encountered! I don't know how we could visualise and explore this but it's certainly interesting.

Re: The Lone Banana Problem in AI

#7
post #6

I've seen lots of long prompt suggestions, however single words are fascinating, some words are extremely heavy, for example, we might assume 'topic' and 'subject' are somewhat interchangeable, but when I'm talking to an LLM swapping those words cause a huge change. Same with names, if you ask it to act like Dave, it's not some random personality, it's a distillation of every Dave it has encountered! I don't know how…

These are the Daves I know: https://m.youtube.com/watch?v=8nvzEqsZIGo
Post reply on HN