• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

New AI tech will replace “golden-ear” audio reviewers soon!

SHB

Active Member
Joined
Sep 1, 2025
Messages
233
Likes
492
GRAS Sound has just released new, AI-powered, virtual audio reviewer technology. Michael Fremer, Jim Austin, Steve Huff and the rest of the dilettantes in subjective audio journalism better figure out what their next careers will be because their days are numbered.

This is truly a modern marvel. Check out the hardware for the virtual reviewer:
1773864264570.jpeg

Personally, I think they’ve nailed it.

The mouth is gaping in astonishment at how 3 dimensional the imaging is. The eyes have the blank, unknowing look of a listener who doesn’t understand any of the science of sound reproduction. And the ears are available in any hue of gold you like (shown here with the pink colored prototypes).

This is hooked up via a 400Tb/sec optical WAN to the 6,100-qubit neutral-atom array developed by researchers at Caltech.

Once the dataset collected from the listening sessions using the Audiophile-Head Origin Listening Emulator (A-HOLE) is transferred to the quantum computer, it takes anywhere from a week to a couple months to process the data and turn it into a semi-coherent stream of thought that is the audio review.

Much of the processing power is consumed by the logic dither algorithm that tries to make sense of considering itself to be an expert when it actually knows so little. Apparently, making an AI both arrogant and ignorant is exceedingly complex.

At present, it draws the electrical equivalent of a small city to write one review, but so what? If it puts Fremer and his ilk out of business, I say it’s worth it.
 
Last edited:
Shouldn't the ears be gold instead of the head/upper torso?

Edit: Nevermind - I failed to read the second half of the post :facepalm:
 
to the 6,100-qubit neutral-atom array developed by researchers at Caltech.
Thats pretty much waste of resources. Humans are able to review directional cable, write hundred words about it, get paid for it, debate on forums, with just one neural thingie.
 
Thats pretty much waste of resources. Humans are able to review directional cable, write hundred words about it, get paid for it, debate on forums, with just one neural thingie.

Sorry, but you’re standing in the way of progress. Accept the future or be crushed by it!
 
It needs a veil to cover the head and a motor to lift it during reviews.
 
Sorry I am all for progress. I have visited the GRAS Sound Co.. Their AI research to find new 'eloquent' words describing sound quality is quite rudimentary.

Aah, yes, one of the early prototypes. They were experimenting with the Million Monkeys at a Million Typewriters theorem. Unfortunately, it was less than successful.
 
Last edited:
Op. You have delivered a knockout blow, to the none ASR hifi reviewing community! But are they listening? :)
 
It will be released officially April 1st. Someone obviously leaked...

It would be nice to get independent AI reviews of dacs, cables, amps etc. too!
 
GRAS Sound has just released new, AI-powered, virtual audio reviewer technology. Michael Fremer, Jim Austin, Steve Huff and the rest of the dilettantes in subjective audio journalism better figure out what their next careers will be because their days are numbered.

This is truly a modern marvel. Check out the hardware for the virtual reviewer:
View attachment 518704
Personally, I think they’ve nailed it.

The mouth is gaping in astonishment at how 3 dimensional the imaging is. The eyes have the blank, unknowing look of a listener who doesn’t understand any of the science of sound reproduction. And the ears are available in any hue of gold you like (shown here with the pink colored prototypes).

This is hooked up via a 400Tb/sec optical WAN to the 6,100-qubit neutral-atom array developed by researchers at Caltech.

Once the dataset collected from the listening sessions using the Audiophile-Head Origin Listening Emulator (A-HOLE) is transferred to the quantum computer, it takes anywhere from a week to a couple months to process the data and turn it into a semi-coherent stream of thought that is the audio review.

Much of the processing power is consumed by the logic dither algorithm that tries to make sense of considering itself to be an expert when it actually knows so little. Apparently, making an AI both arrogant and ignorant is exceedingly complex.

At present, it draws the electrical equivalent of a small city to write one review, but so what? If it puts Fremer and his ilk out of business, I say it’s worth it.
They need to add a few more things:
The jaw must be able to drop to the floor
Then a device/contraption to pick it up from the floor and re-place it .. Just gaping is not enough ..
Still there is an issue that AI may never, yes, never solve; That of the " even the spouse in the kitchen" ...
 
Back
Top Bottom