Comment by ge96
21 hours ago
My naive question, can image generation make something novel eg. "show me a DNA structure that cures cancer" can it do that, or it has to have seen something before to generate it.
Just think we conceptually know what a brushless motor design looks like and it's just pixels. I guess even if it did produce the image we wouldn't know what it means.
All image models can generate images that were not in its training dataset, but it can't generate reductive extreme cases like your example.
What about it is extreme? It's a concept, like "generate an xray image" eventually hopefully the cure to cancer could be represented as a simple molecule or whatever, I'm not saying I know.
There is currently no knowledge nor progress for what a cure for cancer, and nothing a LLM can draw upon.
You could generate "pregnant Elon Musk with four arms and three eyes doing yoga poses" because the image models have enough visual concepts of each of those individual things, but that specific image is (likely) not in any training dataset.
1 reply →
You are overestimating it's intelligence, but I bet it would hallicinate some result, why not try it yourself?
I don't know what the cure of cancer would look like ha (not an organic chemist or biology, genomist... not even sure what field that would be).
But yeah I am slowly trying to incorporate AI into my life (the delegation, work in my sleep part). I develop it is the funny thing (RAG agents) but yeah. Sometimes I get sold on it like "wait a minute maybe it can do that" but no. Can probably tell I don't get deep into the technical part I'm an API consumer. That's the thing I realize too, can only know so much about a topic if you're spread thin/a generalist.