I don't know exactly how grok works but I'd bet when it's analyzing a "video" it's actually looking at a grid of images sampled from the video. So it's skipping most of the frames and there's no motion data or anything. Someone running the wrong across a track is exactly the kind of "thing that looks like the training set but not quite" that will screw up an AI.
Since video generation is really taking off, I'd expect full video analysis to be part of the next generation of multi-modal LLMs. Give it about a year.
This is really the perfect case study for how difficult it is to reach normies. Literally everyone has been telling people that AI doesn't think anything since the very first day it became available. There are no popular theories about AI actually thinking. And STILL there are plenty of people who attempt to interact with it as though it thinks.
There are no popular theories about AI actually thinking.
Except that retarded former Google AI engineer running around telling people maybe AI is becoming self-aware and he's scared. I literally facepalmed when Sargon used quotes from the guy on a recent Akkad Daily video. (ultimately I agree with his overall point that the tech could become dangerous, but from misuse like any tech, not from any kind of machine mind self-motivation)
"Grok has nfi what its doing"
Dear Lord... It's becoming more human than ever.
Truth is, no MML has ever had any idea what it was ever doing.
It's jeets all the way down!
As soon as we get rid of the "Garbage in" part it should all work out just fine...
"Ai" being total shit? You don't say...
The non-free AI is doing amazing stuff, can make very realistic looking videos. With free AI you're still largely getting what you paid for.
I don't know exactly how grok works but I'd bet when it's analyzing a "video" it's actually looking at a grid of images sampled from the video. So it's skipping most of the frames and there's no motion data or anything. Someone running the wrong across a track is exactly the kind of "thing that looks like the training set but not quite" that will screw up an AI.
Since video generation is really taking off, I'd expect full video analysis to be part of the next generation of multi-modal LLMs. Give it about a year.
This is really the perfect case study for how difficult it is to reach normies. Literally everyone has been telling people that AI doesn't think anything since the very first day it became available. There are no popular theories about AI actually thinking. And STILL there are plenty of people who attempt to interact with it as though it thinks.
Except that retarded former Google AI engineer running around telling people maybe AI is becoming self-aware and he's scared. I literally facepalmed when Sargon used quotes from the guy on a recent Akkad Daily video. (ultimately I agree with his overall point that the tech could become dangerous, but from misuse like any tech, not from any kind of machine mind self-motivation)
it can't even look up links properly
autistic screeching
It's doesn't even recognize that thing it just sent you is a link!
Nah, it just constantly gives me wrong links. And then tells me they’re correct and then when prompted its wrong, gives me wrong links again
Just think how great it'll be in five years! It'll replace any human endeavor!
/s