5 ms·

It feels like string theory will be a cautionary tail of being too myopic on a single promising thread of inquiry. We've probably wasted 20+ years after we shou

It feels like string theory will be a cautionary tail of being too myopic on a single promising thread of inquiry. We've probably wasted 20+ years after we should have jumped tracks (or at least stared changing course). Religions of belief tend to pop up around attractive academic ideas that can't be falsified but also have great explanatory power.

bernulli 5y ago
It's a really tough question of when to pull the plug. Marvin Minsky wrote a whole book in 1969 'proving' that neural nets cannot, fundamentally, do anything interesting - yet, here we are.
- otabdeveloper4 5y ago
  He was mostly right. Neural nets are mostly an artistic tool for creating awesome generative art. They don't give any scientific or mathematical answers.
  - civilized 5y ago
    > They don't give any scientific or mathematical answers. Dead wrong. Predicting protein folding has been a scientific holy grail for decades, and neural network-based AlphaFold is the leader. No non-neural approach comes within a country mile of its performance. They created a fold database of unprecedented coverage which is being used in research.
  - bernulli 5y ago
    He was decidedly not right. His claim was that neural networks are fundamentally not able to perform an XOR step, limiting any relevant computational capabilities. That claim was easily remedied later by slightly changing network architectures later on.
RandomWorker 5y ago
It’s like imaginary numbers are fake and don’t exist, but many calculations for electric circuits are easily done using these numbers. Sting theory is that to the nth degree. And, when done well it allows you describe the largest distances and the smallest particles using one mathematical principle. It’s useful in a sense that we can explain the existing observations. Though the predictions will be (currently) impossible to verity because where we have working theories in easy domain, planet scale -1g -1atm, nothings gets disproven. It’s that larger scale at insane conditions we may conclude something interesting. Though testing it will take another possibly larger hydrogen collider, or gravity wave measurement device. The quote They fight you and call you crazy, until your not them you are a genius — has never been more true. I say let the ideas be worked out, and the market place of ideas will solve if it’s relevant. The authors assumption that particle theoretical physics is dead is imo completely wrong, there are still many great people working on it.
- mgamache 5y ago
  Eventually it will sort out yes. There's a cost to beating a dead horse in time and resources. We've been stuck for 40 years: Physicists today can happily make career by writing papers about things no one has ever observed, and never will observe. This continues to go on because there is nothing and no one that can stop it. https://iai.tv/articles/why-physics-has-made-no-progress-in-50-years-auid-1292 https://iai.tv/articles/why-physics-has-made-no-progress-in-...
  - throwaway81523 5y ago
    Nobody has ever observed the interior of a black hole and reported the results to the outside, and nobody ever will (unless GR is drastically wrong). And yet we consider our theory of BH interiors to be basically valid except at the center, where it breaks down. That's perfectly good physics, so deal with it.
- sorokod 5y ago
  > imaginary numbers are fake and don’t exist, In what relevant sense do imagenary numbers differ from real numbers? Being a mathematical abstraction they are in no way required to help approximate phisical reality. That they do is a source of wonder to many, but that is beside the point.
tim333 5y ago
"Science progresses one funeral at a time" I think the string theory guys will stick with it while they can.

bernulli 5y ago

It's a really tough question of when to pull the plug. Marvin Minsky wrote a whole book in 1969 'proving' that neural nets cannot, fundamentally, do anything interesting - yet, here we are.

otabdeveloper4 5y ago
He was mostly right. Neural nets are mostly an artistic tool for creating awesome generative art. They don't give any scientific or mathematical answers.
- civilized 5y ago
  > They don't give any scientific or mathematical answers. Dead wrong. Predicting protein folding has been a scientific holy grail for decades, and neural network-based AlphaFold is the leader. No non-neural approach comes within a country mile of its performance. They created a fold database of unprecedented coverage which is being used in research.
- bernulli 5y ago
  He was decidedly not right. His claim was that neural networks are fundamentally not able to perform an XOR step, limiting any relevant computational capabilities. That claim was easily remedied later by slightly changing network architectures later on.

RandomWorker 5y ago

It’s like imaginary numbers are fake and don’t exist, but many calculations for electric circuits are easily done using these numbers. Sting theory is that to the nth degree. And, when done well it allows you describe the largest distances and the smallest particles using one mathematical principle. It’s useful in a sense that we can explain the existing observations. Though the predictions will be (currently) impossible to verity because where we have working theories in easy domain, planet scale -1g -1atm, nothings gets disproven. It’s that larger scale at insane conditions we may conclude something interesting. Though testing it will take another possibly larger hydrogen collider, or gravity wave measurement device. The quote They fight you and call you crazy, until your not them you are a genius — has never been more true. I say let the ideas be worked out, and the market place of ideas will solve if it’s relevant. The authors assumption that particle theoretical physics is dead is imo completely wrong, there are still many great people working on it.

mgamache 5y ago
Eventually it will sort out yes. There's a cost to beating a dead horse in time and resources. We've been stuck for 40 years: Physicists today can happily make career by writing papers about things no one has ever observed, and never will observe. This continues to go on because there is nothing and no one that can stop it. https://iai.tv/articles/why-physics-has-made-no-progress-in-50-years-auid-1292 https://iai.tv/articles/why-physics-has-made-no-progress-in-...
- throwaway81523 5y ago
  Nobody has ever observed the interior of a black hole and reported the results to the outside, and nobody ever will (unless GR is drastically wrong). And yet we consider our theory of BH interiors to be basically valid except at the center, where it breaks down. That's perfectly good physics, so deal with it.
sorokod 5y ago
> imaginary numbers are fake and don’t exist, In what relevant sense do imagenary numbers differ from real numbers? Being a mathematical abstraction they are in no way required to help approximate phisical reality. That they do is a source of wonder to many, but that is beside the point.

tim333 5y ago

"Science progresses one funeral at a time" I think the string theory guys will stick with it while they can.