6 ms·

I know it's fun, and I think participative science is such a huge shift in paradigm … But considering the global impact of coronavirus, I'm wondering if there s

by Aissen 7y ago

I know it's fun, and I think participative science is such a huge shift in paradigm … But considering the global impact of coronavirus, I'm wondering if there shouldn't be a better way to fund this type of endeavor. Any cloud provider willing to donate compute to researchers?

sschueller 7y ago
How about the government giving access to its huge computing power or commandeering free resources at large providers for the common good?
- gruez 7y ago
  >commandeering free resources at large providers for the common good? Sounds like a good way to get every cloud provider to pull out of your country.
_Wintermute 7y ago
The research institutes I've worked at have all had copious amounts of compute resources usually freely available to reserarchers. I'm not in this field, but I really don't think compute power is the limiting factor in stopping this outbreak.
- usrusr 7y ago
  Yeah, if you could somehow just trade compute for cure... Got a minor chuckle out of the hypothetical that the virus could somehow be defeated with hashrate: poor thing wouldn't know what hit it.
cowsandmilk 7y ago
fold.it is a lot more about the fact that humans still are a lot better at exploring complex spaces than computers. Analysis of results has shown players reaching semi-global minima in folding space much quicker than computers and designing novel folds that express and crystallize that have not been seen in nature before (which means they can explore more space than starting from known templates). Researchers are trying to learn from the types of moves that humans make in the game to improve algorithms, but humans are still better at searching, making very productive large moves in folding space. So basically, this is less about funding and more about the fact that our algorithms aren’t good enough.
- Aissen 7y ago
  Thanks for the context, I remember now, I had completely missed this part. Your comment should be at the top of the thread.
moomin 7y ago
So close, and yet so far... What you're looking for is a well-funded public health system. And to vote out politicians that defund healthcare.
- ppf 7y ago
  Public health systems, well-funded or not, are not a cure-all. I'd be curious to hear if you have any examples - if anything, they are reliant on research funded by the huge amount of money sloshing around the US healthcare system. Given a fixed annual budget, a public health system essentially has to decide what "healthy" or "cured" means, and set some quality of life to cost of treatment ratio. Some examples: my dad, an incredibly fit ultra-runner, suddenly found himself unable to run more than a few miles, and was gasping for air. He was initially told he was fine for his age, and had to fight for further diagnosis. A close friend of my wife's died of bowel cancer - this is a known "loophole" in the diagnosis pathways of our public health care system. Essentially, it was decided that the cost of early diagnosis is too high for the relatively tiny amount of people who have it, which is why most cases are found when they are terminal.
  - nolongerlowres 7y ago
    I’m sorry to hear about your wife’s friend. But both examples you gave (people having to fight for further diagnoses of outside-case health problems) are really common with private healthcare in the United States too - we’ve just shifted the burden of cost of treatment/quality of life decisions to private entities who typically are only interested in furthering shareholder value. To give a personal anecdote as well: A close friend of my wife’s had to fight and fight to get insurance to cover appointments in order to get her breast cancer diagnosed. I would like to understand if the world really does rely on US private-healthcare funded research - I’ve heard this argument mentioned before and have always been curious about it.
    - ppf 7y ago
      Agreed, the US system is flawed. I suppose with public-funded healthcare, everyone is on the same level - the survival rates of breast cancer and bowel cancer is the same for all (unless you can afford to fly to the US and pay for care there).
dekhn 7y ago
The big cloud providers already donate a fair amount of compute (for example, some time ago I ran exacycle, which provided teams like FOlding@Home and Rosetta with extensive computing power. However, the economics of cloud are such that nearly all large-scale compute needs to provide revenue to the cloud provider for the provider to stay competitive. I've worked with the US gov't funding agencies in the past to make it possible to get "cloud credits" instead of grant $$$ (and even sat on panels that selected grants to promote cloud computing over closet clusters). The US works quite slow. IIRC, fold.it is from David Baker, and he has a stable funding pipeline due to the sucess of Rosetta.
- chapium 7y ago
  Governments have a staggering amount of computational power. Good will is nice and all, but are we putting our tax payer funds to an obvious public need?
  - dekhn 7y ago
    Do they? I mean, technically the US gov't leases a ton of desktops (not great for scientific/technical computing), and they own a few supercomputers for specific needs (national defense, weather), but mostly they fund modest computational centers in academic environments. Before cloud, I worked in grid, and grids strapped together the many academic and government clusters, mainly for embarassingly parallel workloads. I don't think the government has shown itself to be a particularly skilled operator of large computing environments compared to the large cloud providers.

by Aissen 7y ago

sschueller 7y ago

How about the government giving access to its huge computing power or commandeering free resources at large providers for the common good?

gruez 7y ago
>commandeering free resources at large providers for the common good? Sounds like a good way to get every cloud provider to pull out of your country.

_Wintermute 7y ago

The research institutes I've worked at have all had copious amounts of compute resources usually freely available to reserarchers. I'm not in this field, but I really don't think compute power is the limiting factor in stopping this outbreak.

usrusr 7y ago
Yeah, if you could somehow just trade compute for cure... Got a minor chuckle out of the hypothetical that the virus could somehow be defeated with hashrate: poor thing wouldn't know what hit it.

cowsandmilk 7y ago

fold.it is a lot more about the fact that humans still are a lot better at exploring complex spaces than computers. Analysis of results has shown players reaching semi-global minima in folding space much quicker than computers and designing novel folds that express and crystallize that have not been seen in nature before (which means they can explore more space than starting from known templates). Researchers are trying to learn from the types of moves that humans make in the game to improve algorithms, but humans are still better at searching, making very productive large moves in folding space. So basically, this is less about funding and more about the fact that our algorithms aren’t good enough.

Aissen 7y ago
Thanks for the context, I remember now, I had completely missed this part. Your comment should be at the top of the thread.

moomin 7y ago

So close, and yet so far... What you're looking for is a well-funded public health system. And to vote out politicians that defund healthcare.

ppf 7y ago
Public health systems, well-funded or not, are not a cure-all. I'd be curious to hear if you have any examples - if anything, they are reliant on research funded by the huge amount of money sloshing around the US healthcare system. Given a fixed annual budget, a public health system essentially has to decide what "healthy" or "cured" means, and set some quality of life to cost of treatment ratio. Some examples: my dad, an incredibly fit ultra-runner, suddenly found himself unable to run more than a few miles, and was gasping for air. He was initially told he was fine for his age, and had to fight for further diagnosis. A close friend of my wife's died of bowel cancer - this is a known "loophole" in the diagnosis pathways of our public health care system. Essentially, it was decided that the cost of early diagnosis is too high for the relatively tiny amount of people who have it, which is why most cases are found when they are terminal.
- nolongerlowres 7y ago
  I’m sorry to hear about your wife’s friend. But both examples you gave (people having to fight for further diagnoses of outside-case health problems) are really common with private healthcare in the United States too - we’ve just shifted the burden of cost of treatment/quality of life decisions to private entities who typically are only interested in furthering shareholder value. To give a personal anecdote as well: A close friend of my wife’s had to fight and fight to get insurance to cover appointments in order to get her breast cancer diagnosed. I would like to understand if the world really does rely on US private-healthcare funded research - I’ve heard this argument mentioned before and have always been curious about it.
  - ppf 7y ago
    Agreed, the US system is flawed. I suppose with public-funded healthcare, everyone is on the same level - the survival rates of breast cancer and bowel cancer is the same for all (unless you can afford to fly to the US and pay for care there).

dekhn 7y ago

The big cloud providers already donate a fair amount of compute (for example, some time ago I ran exacycle, which provided teams like FOlding@Home and Rosetta with extensive computing power. However, the economics of cloud are such that nearly all large-scale compute needs to provide revenue to the cloud provider for the provider to stay competitive. I've worked with the US gov't funding agencies in the past to make it possible to get "cloud credits" instead of grant $$$ (and even sat on panels that selected grants to promote cloud computing over closet clusters). The US works quite slow. IIRC, fold.it is from David Baker, and he has a stable funding pipeline due to the sucess of Rosetta.

chapium 7y ago
Governments have a staggering amount of computational power. Good will is nice and all, but are we putting our tax payer funds to an obvious public need?
- dekhn 7y ago
  Do they? I mean, technically the US gov't leases a ton of desktops (not great for scientific/technical computing), and they own a few supercomputers for specific needs (national defense, weather), but mostly they fund modest computational centers in academic environments. Before cloud, I worked in grid, and grids strapped together the many academic and government clusters, mainly for embarassingly parallel workloads. I don't think the government has shown itself to be a particularly skilled operator of large computing environments compared to the large cloud providers.