No, it’s hard because the energy levels that we have to have to test things at the plank scale are much higher than anything we can achieve right now with our current level of technology. Plenty of theories make predictions about quantum gravity, string theory, M theory, lopp quantum gravity. There’s even a few out there theories that just try to modify newtonian gravity.
This reminds me of how someone illustrated the machine learning problem of what I want to say is called “gradient descent”. This was way back in the 2000s before all the more recent AI stuff.
Basically the problem as I remember it being described in a Tedtalk was if you think of a problem like a sphere with a surface and a bunch of tunnels at the surface, where only one leads to the core (answer) of the sphere. Some tunnels might get really close to the core, but only one leads into the core. The AI would get stuck diving down these holes using insane amount of computational power trying to dig for the answer, not realizing that if it backed up a bit and went down the hole next to them they could reach the core (answer).
One way to help this problem was developing the game “Foldit” which allowed regular old users to manipulate the proteins themselves. When people had foldit at home running they would notice that the Screensaver displaying the folding would skip over what seemed to be the right shape and would get frustrated that they couldn’t help guide it.
Trying and failing.
Is it not possible that it’s “hard” because we’re chasing the wrong path.
This isn’t something I alone think. You seem to be under the impression I have a less than Wikipedia level understanding of this. I do not.
No, it’s hard because the energy levels that we have to have to test things at the plank scale are much higher than anything we can achieve right now with our current level of technology. Plenty of theories make predictions about quantum gravity, string theory, M theory, lopp quantum gravity. There’s even a few out there theories that just try to modify newtonian gravity.
It’s “hard” because we didn’t find what we expected at the energy levels we targeted.
There is too much funding behind it now. No one can question the status quo and maintain funding.
As I said, you don’t know what you’re talking about. That’s all there is to this conversation.
This reminds me of how someone illustrated the machine learning problem of what I want to say is called “gradient descent”. This was way back in the 2000s before all the more recent AI stuff.
Basically the problem as I remember it being described in a Tedtalk was if you think of a problem like a sphere with a surface and a bunch of tunnels at the surface, where only one leads to the core (answer) of the sphere. Some tunnels might get really close to the core, but only one leads into the core. The AI would get stuck diving down these holes using insane amount of computational power trying to dig for the answer, not realizing that if it backed up a bit and went down the hole next to them they could reach the core (answer).
One way to help this problem was developing the game “Foldit” which allowed regular old users to manipulate the proteins themselves. When people had foldit at home running they would notice that the Screensaver displaying the folding would skip over what seemed to be the right shape and would get frustrated that they couldn’t help guide it.
This might be a different Ted Talk, but it is about the same subject.
Here is an alternative Piped link(s): https://piped.video/uBA0vKURH3Y?si=NS5Wp4FzjjtvxwGh
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source, check me out at GitHub.