Blorgals and Fleebs

-=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- (c) WidthPadding Industries 1987 0\|429\|0 -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=- -=+=-
Socoder -> A.I. Stuff -> Blorgals and Fleebs
Tue, 17 Jun 2025, 15:55
Pixel_Outlaw	Blorgals and Fleebs Some telling results. It's interesting that a child can classify with a single image. There have been some concerns that we're "doin' it wrong" when it comes to the way AI is currently being implemented. Here is an interesting test I laid out for ShatGipitee. I did appreciate Dr. Sussman's talk on "We Really Don't Know How to Compute" View on YouTube
Tue, 17 Jun 2025, 16:11
Jayenkai	I don't think it did "that" badly. It only really miscounted, if anything. And we've seen the AI's not only miscount on a bazillion occasions, but also then be pretty damn stubborn about how it absolutely didn't miscount, and you must be wrong. How many R's in Strawberry being the common one. It did identify which shape was which, and told you how to distinguish between them. It understood that much, and figured it out itself. It just didn't count them right at the end. -=-=- ''Load, Next List!''
Tue, 17 Jun 2025, 16:30
Pixel_Outlaw	Counting is the basis for computation.
Tue, 17 Jun 2025, 17:37
Jayenkai	View on YouTube