Status of DC and its worth?

[StefanR5R]

Just think what 1 national evening news broadcast telling of a find and how everyone could help.

Like many other more or less closely related animal species, homo sapiens requires rather contiguous gratification in order to stick with doing something continuously which is more involved than low-level life functions such as breathing. Some time after such a one-off news cast, attention would shift away again quite soon. So it would make a short impact, but not a real difference at all. Rather, there would have to be a continuous stream of success stories. (Which would be a remarkable cultural shift in itself, because the sort of news that humans tell each other day in, day out, are predominantly about humanity's conflicts and failures, rarely about its successes.)

Yet computer simulations, numeric data analysis etc. are generally just tools for the scientific process. It'd be rare for a discovery to identify mere simulations and analyses as a primary reason of its emergence.

 

[Assimilator1]

Some time after such a one-off news cast, attention would shift away again quite soon. So it would make a short impact, but not a real difference at all.

IIRC Folding@home was quite widely talked about in a whole host of different non DC websites (at the least, not sure about TV) during the Covid pandemic and it got a massive boost in output, so much so that their WU servers couldn't keep up! lol. Yes it was a temporary boost (about a year or 2 I think), but I'm sure that massive shot in the arm to their project did make a difference, if only to rule out a bunch of negative results in a much shorter time (although I'm pretty sure they did get some positive results, however I don't recall the outcome. I would think such reports would be on their website though, F@H are better at reporting back than many other DC projects).

 

[StefanR5R]

This is a notable case: F@H didn't even announce a success, they only announced that they just started Covid research.

And indeed, massive participation at F@H was following. What was the big motivation? There was a new virus out there, highly contagious, potentially deadly, there was no cure, no vaccine. Practically everybody's life was affected daily in many ways, at times rather drastic. At that point, F@H's announcement provided hope that we can help to end this situation.

The virus was in the daily news for quite some time, but what's more, it actually kept affecting our daily lives directly. Likewise, big participation at F@H held up for quite some time. After the real big peak which lasted weeks or months, participation decreased, IIRC first gradually, then asymptotically towards the current level which is probably about the same as before the pandemic (WRT head count, not WRT computing throughput which is subject to technological advance).

I agree that this participation, while it lasted, must have made a difference, and not just to Covid research and not just to the researchers associated with the F@H consortium. What I want to point out is that this participation must have been influenced, or arguably caused, by how the virus kept influencing our lives and occupying our minds, daily.

Now let's assume there had been an announcement later on that vaccine XYZ is entering test stages and it was based on F@H's research, perhaps even later followed on by the news that vaccine XYZ passed tests successfully and is recommended to the public. What would be the effect of such events on F@H participation? (Replace F@H by DC if you want, or vaccine by treatment.) We don't know the answer. My hypothesis on it is the one in #26.

 

[Exascaletech]

Like many other more or less closely related animal species, homo sapiens requires rather contiguous gratification in order to stick with doing something continuously which is more involved than low-level life functions such as breathing. Some time after such a one-off news cast, attention would shift away again quite soon. So it would make a short impact, but not a real difference at all. Rather, there would have to be a continuous stream of success stories. (Which would be a remarkable cultural shift in itself, because the sort of news that humans tell each other day in, day out, are predominantly about humanity's conflicts and failures, rarely about its successes.)

Yet computer simulations, numeric data analysis etc. are generally just tools for the scientific process. It'd be rare for a discovery to identify mere simulations and analyses as a primary reason of its emergence.

I totally disagree and here's why. DC has never really had much nat'l coverage because there really wasn't much to talk about other than a pie-in-the-sky theory that maybe this will work, we really don't know. And 20 years later, we still don't have much except and people/programs in disarray. The illusion that it has worked is where? It never changed while I was doing it. The interest created by actually having a "find" and coverage of that would create serious interest and maybe finally get some serious funding, a DC option installed by default by some OEM's and with the masses simply following what social media tells them w/o the ability to do any critical reasoning, it would now be super easy for the non-technical, pre-installed, have a cool factor and donations would soar.

I do agree that the short term gratification would be needed to keep the serious cruncher's as that is what they seem to mainly enjoy and really they are doing most of the work with some of the amazing systems they have. For me I'd no doubt build my super rig and join the nat'l pride of giving back to society instead of the social norm of people asking what everyone is going to do for them.

Getting millions who could start that night because it was already installed, it would have a major impact, simply getting one find that every one could understand, the rest would all fall in place.

 

[StefanR5R]

I remain convinced that there would have to be such a "find" every several months or so, e.g. maybe twice a year. People *will* gradually lose interest after one such find. Some quicker, others slower. But the large majority will turn their backs on it again eventually.

But we don't even have a single such "find". And that's because simulations and data analyses are really just small stepping stones in getting difficult answers to difficult questions. Worse, once there are answers, even more questions (of even greater difficulty) arise. It's the nature of the beast. There is no way around it.

In other words, we can't just *compute* a drug.
We can't even just say "Here look: This signal here originates from an alien intelligence."

(Don't trust people who claim that there are simple answers to our big questions.)

 

[Exascaletech]

I remain convinced that there would have to be such a "find" every several months or so, e.g. maybe twice a year. People *will* gradually lose interest after one such find. Some quicker, others slower. But the large majority will turn their backs on it again eventually.

But we don't even have a single such "find". And that's because simulations and data analyses are really just small stepping stones in getting difficult answers to difficult questions. Worse, once there are answers, even more questions (of even greater difficulty) arise. It's the nature of the beast. There is no way around it.

In other words, we can't just *compute* a drug.
We can't even just say "Here look: This signal here originates from an alien intelligence."

(Don't trust people who claim that there are simple answers to our big questions.)

Lets learn to walk before we run. A find twice a year, how about 1 find in the last 20? At least there would be something to believe in. If we're not making things simple enough with our calculations, then we are moving in the wrong direction.

Science for what we are looking for is getting easier as chemistry has been here for centuries, computers have not.

(Simplicity is complex. It's never simple to keep things simple. Simple solutions require the most advanced thinking)

 

[Skillz]

Negative or no result is still a good result.

 

[Ken g6]

GPUGrid at least publishes papers every so often, and even links you to the papers you helped with in your badges.

Has F@H even led to any published papers? (Probably, but the fact that I can't recall any means they don't publicize it enough.)

PrimeGrid, of course, does find primes, for whatever that's worth.

 

[Endgame124]

GPUGrid at least publishes papers every so often, and even links you to the papers you helped with in your badges.

Has F@H even led to any published papers? (Probably, but the fact that I can't recall any means they don't publicize it enough.)

PrimeGrid, of course, does find primes, for whatever that's worth.

F@H has a number of papers. The modeling they did of the corona virus lead to multiple papers and lead to some anti virals as I recall, and the dramatic increase in processing power in 2020 and 2021 lead to a lot more papers than they usually do.

Papers & Results – Folding@home

foldingathome.org

 

[Assimilator1]

I checked out that page, but none of the papers are linked, although if they were it's likely they'd be too technical for me anyway, lol.
That page says their are summaries of the papers on Folding wiki article, but I saw no such summaries there and it simply links back to the F@H paper and results page.
I have so far only read the wiki section "Examples of application in biomedical research", "Alzheimer's disease" and "Huntington's disease" paragraphs, are the summarised papers within the other individual listed diseases?
From a quick scan through them, it didn't seem like it.

 

[Endgame124]

Just saw this paper and, honestly, it’s not an easy read but it’s relevant:

Folding@home: achievements from over twenty years of citizen science herald the exascale era

Simulations of biomolecules have enormous potential to inform our understanding of biology but require extremely demanding calculations. For over twenty years, the Folding@home distributed computing project has pioneered a massively parallel approach ...

www.ncbi.nlm.nih.gov

 

[Exascaletech]

The common thread seems to be unlimited research hours and a spec of sand found here and there. Research funding gives researchers time, nothing found and then need more money. There's zero connection between results and funding. That's not what DC was sold as in the beginning for those of us that were there. Even taking 10 years of research off the table by DC would have given us a find several times over, yet it has not. That means DC by orders of magnitude didn't provide the results anyone expected which makes one wonder how effective this form of research really is. Medical finds seem more frequent in the past than they do today. What looks like new today in drug discovery is not a cure but rebranding older drugs, expanding the label and marketing, marketing and more marketing.

Made a ton of $$$ in biotech stocks in the past and continue to do so. The follow the money rule is saying the interest is in what drug companies want to sell and even if there is a find, who is there to pick it up and run with it? If there was any money to be made from this research, a for profit would be doing it on their own on a very large scale.

Not found a program other than WCG that so far seems remotely viable in the medical sector. Rather than a super rig, now looking at a bunch of the Raspberry-Pi route just for efficiency since I'm not convinced that if the next 20 years are not any better than the last 20 this approach will be abandoned.

 

[cellarnoise]

The common thread seems to be unlimited research hours and a spec of sand found here and there. Research funding gives researchers time, nothing found and then need more money. There's zero connection between results and funding. That's not what DC was sold as in the beginning for those of us that were there. Even taking 10 years of research off the table by DC would have given us a find several times over, yet it has not. That means DC by orders of magnitude didn't provide the results anyone expected which makes one wonder how effective this form of research really is. Medical finds seem more frequent in the past than they do today. What looks like new today in drug discovery is not a cure but rebranding older drugs, expanding the label and marketing, marketing and more marketing.

Made a ton of $$$ in biotech stocks in the past and continue to do so. The follow the money rule is saying the interest is in what drug companies want to sell and even if there is a find, who is there to pick it up and run with it? If there was any money to be made from this research, a for profit would be doing it on their own on a very large scale.

Not found a program other than WCG that so far seems remotely viable in the medical sector. Rather than a super rig, now looking at a bunch of the Raspberry-Pi route just for efficiency since I'm not convinced that if the next 20 years are not any better than the last 20 this approach will be abandoned.

Wow... I'm going to go careful here...

Science is not a linear / literalpath...? All the time? Sometimes it takes sitting under the right tree at the right time? The wheel was easy ?

I need to go find a few trees to sit under and just wait Oh, I think I can only sit under one tree at a time... I might need more fallen threads to sit under?

Maybe something like some more advanced / new era of expiremental sillycone and a few generations after that might help?

I miss my 6809e at .89 MEGAhz . I would like that! In order and wait for the results before doing anything and after asking the RAM for a result... RAM back then was not that fast It could multiply a few bits though Just don't ever think about over 16 bits at a time and all will be well.

 

[Exascaletech]

You should have taken more time to be careful.

 

[Exascaletech]

Listen, I get it. If it was easy the world would never have problems because we could solve them the same day we discovered them. I'm still crunching for WCG right now but am very disappointed there hasn't been more progress inmedical DC in 20 years. If you were around way back then, it was the believe that this type of research would be a near instant game changer. I simply think it would be better if DC could be even 1% of what we thought it was going to be. I was a big player at the time, now I'm a small fry compared to what many have. Really wanted to find a project to build a super rig for. The results were sad. Looks like the odds are even slimmer today. Seriously. I am only interested in medial DC projects so other DC for math/astronomy may be doing a lot better as computing power now is crazy. There is a reason participation is at a 10 year low in medical, lack of results. I really do think they need to re-evaluate how they're using our cycles. It made perfect sense 20 years ago that our contribution would move us ahead tremendously, it just didn't happen. WCG is open source with their research which I like but when the research money is there, not one single DC project will ever turn it down as they know they wont get funding from whoever again if they say no.

Has there ever been any discussion of all the medical DC projects going under one roof to maximize their ability, support, cut costs and take another look at things? The government gives silly amounts of grants to listed stock companies, our tax dollars to for profit organizations with sleazy CEO's making millions that I think we could spend some of that to have the brightest in several fields to see what medical DC may do.

 

[StefanR5R]

Apropos funding. Here is an idea for a new DC project: ApplyForResearchGrants@Home. This is a volunteer computing project that uses Internet-connected computers to advance the creation of applications for research grants. Participate by downloading and running a free program on your computer. It will use spare cycles of your processor to investigate new and ongoing funding programs, compute corresponding research proposals and generate competitive applications for funds. The ultimate goal of this project is that scientists can spend more of their time for research, and less for writing applications.

 

[StefanR5R]

Here is one of the various success stories from Distributed Computing which never seem to make it into the evening TV newscasts:
https://boinc.bakerlab.org/rosetta/forum_thread.php?id=15029
It's from mid 2022 and I had already forgotten about it myself.

This message board post links to a news release of the university, in which the detail that computational simulations played a role in the research is already lost. Nevermind any mention of distributed/ volunteer computing. :-| Edit: Of course it will be in the eye of the beholder how big or how small the computational part played a role in this R&D. Maybe it just goes without saying that there is also computation involved in this line of work.

 

[mmonnin03]

Rosetta’s role in fighting coronavirus - Institute for Protein Design

We are happy to report that the Rosetta molecular modeling suite was recently used to accurately predict the atomic-scale structure of an important coronavirus protein weeks before it could be measured in the lab. Knowledge gained from studying this viral protein is now being used to guide the...

www.ipd.uw.edu