AI is good at art.
For instance, if you trained it on 10,000 comic books, it could give you back something that looks analagous.
It's quite awful at engineering. If you asked it to make a two-way loudspeaker, it would have no idea where to start. It just regurgitates back what you train it on.
Where AI will be very useful in loudspeaker design is something that's already been explored over on the ATH thread:
step 1: you give it a design
step 2: you allow parameters of that design to vary. For instance, you could give it a waveguide and tell the AI that it can vary the depth of the waveguide by plus or minus 25%
step 3: then you let 'er rip, and have the AI spew out a thousand iterations of said waveguide
step 4: and then provide it a series of success criterias. For instance, "show me the waveguide with the smoothest polar response on and off axis"
But this isn't really AI, it's just a brute force approach
Like a lot of nerds, I've been tinkering with stable diffusion. At first, it seems like magic. But quite quickly you realize that it can't do something as basic as "make me a picture of what a human being looks like from overhead" UNLESS it's been trained on that. If it hasn't been trained, it can't simulate it. Many of the eye-popping art pieces and fake photos are based on the fact that the models been trained on thousands of images from a comparable perspective; once you tell the AI to make a photo from a different perspective, it can't, because there is no data.
Thinking out loud here, I guess one legitimate use of "real" AI would be to take all of the preference scores collected by Harman and various other companies, train the AI on what works and what doesn't, and then have it create something that's in the same ballpark. But it would only be able to make the crudest approximations; it might be able to tell you that a two-way loudspeaker that uses a particular woofer and tweeter size will often produce good results. But we already know that, and the AI can't do complex tasks like "design a crossover" or "make an ABEC simulation and run it."
It is garbage and will always be garbage.
The equivalent to interns writing tech documents
with no experience. Back in my tech days.
The would re write old documents well written
by older experienced engineers. Already stored
on the companies servers.
Rather comical all the garbage they made
can now be done for free and save money
when companies want to " update"
In the end still garbage....Fired
After managers had meetings with managers.
We actually did the work. Delete Delete Delete.
And restored the original documents.
It is a tough world for office princesses.
When they actually have to work LOL
Instead of handing over work to AI's
and non qualified people.
The equivalent to interns writing tech documents
with no experience. Back in my tech days.
The would re write old documents well written
by older experienced engineers. Already stored
on the companies servers.
Rather comical all the garbage they made
can now be done for free and save money
when companies want to " update"
In the end still garbage....Fired
After managers had meetings with managers.
We actually did the work. Delete Delete Delete.
And restored the original documents.
It is a tough world for office princesses.
When they actually have to work LOL
Instead of handing over work to AI's
and non qualified people.
Real Humans did that in 1947
Called a Transistor.
But I do get your point how cool it would be.
AI's just recycle known information.
Most of which is wrong and out of context.
It is like warehouse workers that finally mature.
They finally realize the old saying.
It is more work, do not work.
You constantly have to revise your input
to a AI to even get a basic accurate response.
Then realize it is recycling generic answers
from google. Usually wrong and opinionated
with some facts included.
So it is essential the same thing, More Work
to get no work done.
It will be interesting to see how Amaras Law will play out with AI. AI isn't magic.
I told a tech worshipping friend over a bite to eat recently that I didn't care about AI. He was shocked. Subsequent to that proclaimation he had the look on his face of "I'm sitting here eating with an Apostate, what do I do now?". Nevertheless, we're still friends.
I told a tech worshipping friend over a bite to eat recently that I didn't care about AI. He was shocked. Subsequent to that proclaimation he had the look on his face of "I'm sitting here eating with an Apostate, what do I do now?". Nevertheless, we're still friends.
I am going to submit this question. Don't hold your breadth. My contraption has improved a lot over the last 9 months, but it's far from perfect yet.
The rate of progress is pretty high, though.
