Hacker News Reader

Comment by rocgf | Hacker News Reader

Does anyone actually believe DALL-E "understands" what it's doing? For any reasonable definition of "understands", I assume most people would be skeptical.

So if we go with that, then yes, it just glues things together without understanding their relationship. I'd just be tempted to say it doesn't really matter that it doesn't understand, except maybe for some philosophical questions. It's still incredible based on its output.

View on HN

> Does anyone actually believe DALL E "understands" what it's doing?

I believe it understands enough to make tens of thousands of people interested and debating its merits. The GANs of 5 years ago were baby toys compared to DALL-E. They were drawing 5 legs to a dog and limited to a few object classes. Now people debate if it "really understands" and if it is "(slightly) conscious".

can we come up with a 'reasonable' definition of understands that isn't couched in human exceptionalism and could still apply to any computer model?

i think dall-e understands, within the sensory domain it's trained from.

I think DALL-E understands the same way an Excel sheet would understand SUM(B3:B6), obviously at a way larger scale and with a more complicated function. It's all just rather simple maths all the way down.

Not similar. The SUM() function was written by humans in a top-down fashion, DALL-E learned from the visual and textual data by a completely different process. DALL-E distilled something about the world and is using that to generate.

Reply on news.ycombinator.com