Friday, March 27

Tag: Artificial Intelligence

I am a painter with work at MoMA and the Met. I just published 50 years of my work as an open AI dataset. Here is what I learned.
News Feed, Reddit

I am a painter with work at MoMA and the Met. I just published 50 years of my work as an open AI dataset. Here is what I learned.

I am a painter with work at MoMA and the Met. I just published 50 years of my work as an open AI dataset. Here is what I learned. I have been making figurative art since the 1970s. Oil on canvas, works on paper, drawings, etchings, lithographs, and more recently digital works. My paintings are in the collections of the Metropolitan Museum of Art, MoMA, SFMOMA, and the British Museum. Earlier this month I published my entire catalog raisonne as an open dataset on Hugging Face. Roughly 3,000 to 4,000 documented works with full metadata, CC-BY-NC-4.0 licensed. My total output is about double that and I will keep adding to it. In one week the dataset has had over 2,500 downloads. I am not a developer or a researcher. I am an artist who has spent fifty years painting the human figure. I did thi...
A supervisor or “manager” Al agent is the wrong way to control Al
News Feed, Reddit

A supervisor or “manager” Al agent is the wrong way to control Al

I keep seeing more and more companies say that they're going to reduce hallucination and drift and mistakes made by Al by adding supervisor or manager Al on top of them that will review everything that those Al agents are doing. that seems to be the way. another thing I'm seeing is adding multiple Al judges to evaluate the output and those companies are running around touting their low percentage false positives or mistakes adding additional Al agents on top of Al agents reduce mistakes is like wrapping yourself in a wet blanket and then adding more with blankets to keep you warm when you're freezing. you will freeze, it will just take longer, and it's going to use a lot of blankets. I don't understand. the blind warship of pure Al solutions. we have software that can achieve determinism. ...
The AI Report