WOW. Amazing! Thanks for sharing. I have heard of it as it has progressed. What a huge leap. Amazing that they say the next levels will be 4 times as good.

Truly extraordinary. One thing that I don’t understand is the degree to which the images are predetermined. If you give the same phrase on two different occasions, will the image be identical, or is there built-in randomness?

@LostInParadise
It creates several variations from a single phrase.

Here is a collection of 1000 robot images generated by the AI:
https://archive.org/details/1111101000-robots/page/n27/mode/2up

I’ve used one of these programs awhile ago. Fun to use. But nowhere near as impressive as this program is now.

This is wild.
Thanks for sharing.

Quite incredible. Salvador Dall.E?

So the day we have A.I. creating A.I. will be when the game really changes. Seeing this makes me think that’s coming somewhat soon.

Maybe I’m too conservative, but I’m not as excited as I should. I’m really concerned where this is going and the implication for artists. Would we come to a day when artists or creative people in general are no longer needed? AI has been replacing us in so many fields, but right now most of them are just manual work. Humans are moving more onto the “manager” role, monitoring the AI and brainstorming ideas. The thing we still have advantage over AI is our creativity. But what if AI can master that too? Where will we go then?

I’m sorry to be so pessimistic. I’d love to be proven wrong.

Good question. And yeah it’s surprising stuff – but also likely a cause for concern.

The results are obviously technically impressive, with hardly any real inspiration. In fact the commentary with the video brushes past the truly important questions of “what is art” and “what is creativity” without even trying for an answer.

Aren’t we at a point finally where serious questions about the human future should be asked, and taken seriously ? Don’t we have a right to know whether AI developments like this will really make our lives better, or just end up leaving us more empty? And if the answer is “no one knows”, then what’s this shit for?

This is just remarkable!

@LostInParadise “One thing that I don’t understand is the degree to which the images are predetermined. If you give the same phrase on two different occasions, will the image be identical, or is there built-in randomness?”

It is my understanding that the process is deterministic, and therefore if you specify you want “Chariots pulled by guppies in outer space in an impressionist style” with 10 outputs, it should produce the same 10 outputs every time (I believe). If the training/underlying model changes then the output will change as well.

That is amazing. Thank you for pointing it out.
I would enjoy playing with it.