Hacker News Reader

Comment by siavosh | Hacker News Reader

A few weeks ago I was skeptical that this technology would get past the emotional response we get from procedurally generated game environments, but I've been convinced otherwise. The emotional response I get from some of the best of these images are novel and thought provoking. Makes me wonder what percent of what makes us human is now algorithmically solved....

View on HN

We still know basically nothing about our Brain/Consciousness. I would say we have a lot more to explore/research

Our brain is apparently just a 4 gb large arrangement of electrical weights.

Not to be pedantic, but we have on the order of 100B neurons, and afaik each of them can be connected to thousands of other neurons. I assume we probably have a ways to go before we're encoding the amount of information a brain can comprehend.

Also, your brain knows how to look up things in external storage, but ML models have to keep everything in their weights even if it's not a good fit.

My wife likes impressionists and sunflowers. "A lone sunflower in a grassy field at sunset oil painting claude monet" plus stable-diffusion and a few minutes of tweaking some settings; she now has a new desktop background.

I actually paint and spend a lot of time looking at 'serious' paintings. AI hasn't even scratched the field to a trained eye.

Doesn't mean I'm not excited though. This kind-of feels like I'm watching the camera or printing press being invented. Everyone is comparing it to fine art, but I think ultimately it's going in a different and bigger direction.

What I did was, IMO, a different and bigger direction to fine art. I mean I could tell that this wasn't an impressionistic painting just given that some areas of the grass were too detailed. It looks "just fine" though to untrained eyes, which are well over 90% of the population.

1. How long would it have taken me to get good enough at painting to exceed what I generated in under an hour? How many people have the motivation to spend that time?

2. How much would I have had to pay an art student to make a painting better than what I generated in under an hour?

Ten million sub-par Monet knock-offs didn't exist, but could exist very shortly at minimal cost. Even if it never gets any better this is already potentially disruptive, and the models are getting better every month.

I've heard this a lot, luckily it's not that hard to test if you can really tell the difference. We need someone to create the 'AI Pepsi challenge' for artists to settle this.

That was fast, looks like artists can't tell, AI wins first prize in an art competition!

https://www.vice.com/en/article/bvmvqm/an-ai-generated-artwo...

This was sort of misreported. He won in the "Digital Arts / Digitally-Manipulated Photography" category, not the entire contest, and it's fair to say AI fits in that category.

This picture is also unusually coherent for Midjourney; if you just ask for a 16:9 image the sides tend to evolve into totally different pictures.

That was the State Fair fine arts competition.. I wouldn't exactly say you're about to fool the MOMA

Ah ok and when they are fooled I wonder what rationalization you will fall back on next.

Those are garbage western judges. Modernism really destroyed representational art in the west. Try a competition with Russians or Italians running the jury. Or ARC-recognized atelier instructors.

Are they particularly good at telling the difference between wet neural nets and dry ones?

Maybe. I've been often reminded lately of Herbert Goldstone's "Virtuoso" (1958):

http://elateachers.weebly.com/uploads/2/7/0/1/27012625/virtu...

Reply on news.ycombinator.com