Rendered at 19:34:32 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
ryandvm 4 hours ago [-]
I'm a huge fan of Feynman and his autobiographies - some of my favorite books. There are a number of YouTube videos that purport to be Feynman's audio lectures, but they're actually AI generated readings of his script. Against my better instincts I listened to a couple and it was surreal experience.
On one hand, they were actually pretty decent simulacra and I loved hearing Feynman's voice and mannerisms; but on the other hand, they... weren't... real. What am I even listening to?
We are in for a weird future.
MSKJ 2 hours ago [-]
There's this stigma towards AI in certain things that I find interesting. In your case was the weirdness (the uncanny valley area) due to the AI using Feynman's voice? Would it still be weird if it used a different voice?
The AI generation in this case is a tool to convert text to voice, one of the best use cases of this technology IMO
ryandvm 31 minutes ago [-]
It was weird because it seemed authentic, but was not. There was nothing technically uncanny about it - the illusion was complete. Ten years ago I would have absolutely thought that it must have been Feynman's own voice.
I find it weird in the same way that I feel about AI services that will let you chat with deceased relatives. It's the principle of the thing that is unsettling. Perhaps future generations will find these objections quaint.
UmYeahNo 3 hours ago [-]
Angela Collier's video about Feynman completely changed my view of him.
ryandvm 2 hours ago [-]
I suspect nobody's legacy survives the microscope.
220hertz 3 hours ago [-]
I generally agree with that video. But did she really need to take three hours to make her point?
Isamu 3 hours ago [-]
She’s not known for the concision of her videos, true. The Feynman one is interesting to me because I also read the books about his exploits and enjoyed them.
As a reader I think you should be aware of the strong human tendency to THINK you know the character of the person you are reading about. But you should resist this temptation, because ultimately it will be a projection of your own making.
If you enjoy Feynman stories, go ahead. You don’t need to defend him or get caught up bro culture or alternatively you should be careful about denunciations of historical figures that are performative.
You should be clear about what your morals are and not base them on some historical character that you may not understand completely.
I agree that something like *lisp would be a welcome relief from the hairball that is CUDA. while the lisp part was actually compiled, the parallel part reduced pretty 1:1 onto Paris, (the 'assembly language' designed by Steele). that was was basically a jump table into the microcode, which was just a big control word issued by (what I remember vaguely was being an amd 2900 bit slice) sequencer.
today I guess you'd translate those vector instruction into MLIR and hope for some layout and fusion.
but absolutely, you can just use tensors like normal variables. what a awful regression
On one hand, they were actually pretty decent simulacra and I loved hearing Feynman's voice and mannerisms; but on the other hand, they... weren't... real. What am I even listening to?
We are in for a weird future.
The AI generation in this case is a tool to convert text to voice, one of the best use cases of this technology IMO
I find it weird in the same way that I feel about AI services that will let you chat with deceased relatives. It's the principle of the thing that is unsettling. Perhaps future generations will find these objections quaint.
As a reader I think you should be aware of the strong human tendency to THINK you know the character of the person you are reading about. But you should resist this temptation, because ultimately it will be a projection of your own making.
If you enjoy Feynman stories, go ahead. You don’t need to defend him or get caught up bro culture or alternatively you should be careful about denunciations of historical figures that are performative.
You should be clear about what your morals are and not base them on some historical character that you may not understand completely.
(Supposedly they're building it in West Texas - last update in 2024)
https://mitpress.mit.edu/9780262192774/the-paralation-model/
today I guess you'd translate those vector instruction into MLIR and hope for some layout and fusion.
but absolutely, you can just use tensors like normal variables. what a awful regression