The COVID Moonshot also published its results free to use for anyone. This is much different than organizations that have billions of funding behind them looking to make money. Hence the slower going research. But perhaps something unique will come out of it.

I think the main issue about the "hard sell" is that nobody is promoting DC. WCG doesn't spend a single minute of the year promoting itself and the projects...and hasn't for 5+ years in my opinion.

So...nobody's gonna buy if they don't know you're selling.

It is my limited understanding that AI cannot simply accelerate the computation of a math problem that needs to be solved. As you know, the data sent to clients is some sort of Model (the big file when project is attached) And work units which are keys. The keys are computed against the model each time.  AI cannot accelerate this, same as if 7-zip  wanted to increase compression / decompression time with AI.

If there is data to infer and compare AI shines. Such as looking at results of computation.


Could Krembil develop a system that attains the same or more useful results without needing the massive iterative compute by utilizing AI?  possibly, but how much time and effort  would that take away from use the data they attain now and research what models and work units to generate? I think they have enough compute, they aren't asking for more  it is about the correct  pace. But if they coild, would  dramatically reduce power consumption  of project. But, they are working on GPU support. It will  be interesting if GPUs with CPU to feed, will perform all compute for Markers work units,  or if there will be stages of compute still CPU only. If CPU will stilp be necessary or a low-end to feed GPU will be a good match.


Lack of new projects.  I do not know how much of it is from lack of awareness of groups in the field that could use the charitable compute, or difficulty in setting up a project. Something  'Dockers' could help, which BOINC developer Vitalii Koshura is on top of. (EDIT https://aenbleidd.blogspot.com/2024/11/boinc-...tatus-report-october.html )Also i believe BOINC is getting  support for networked work unit compute, so-called 'Capability Compute' rather than traditional  Capacity Compute. (EDIT 'Sporadic' https://github.com/BOINC/boinc/wiki/Sporadic-Applications) So many ideas that professionals have requires communication because Work Units are related to each other so much, in my limited understanding.


Founders of BOINC, not just World Community Grid peeps wonder about the future of these platforms. I imagine Folding too.


We can spread public awareness by trying to make viral videos of the topic. WCG BOINC apparel and bumper stickers?

Gridcoin Benefactor contract idea could help awareness as well if implemented well.

Show the DeSci community, don't reinvent the wheel, BOINC is the way to go! 

100% agree on emails and updates of project efforts , would help plant the seed of reputation.
That's all I have for now off the top of my head.

Mini-PCs are a nice, energy-efficient way to crunch. They usually use laptop CPUs but have better cooling. There are a few laptops with heavy-duty cooling, but they tend to be expensive "portable workstation" machines.

One other thing we need to note is that not all of the DC world is English-speaking. The top team other than IBM is Team 2ch (from Japan), and there are teams from France, Germany, China, Taiwan, and plenty of other places in the top 100.

Exactly. Take a look at the points in this thread . The thread isn't that long and I have some great examples of the Mac mini M2 machines with respect to purchase price, performance, electricity, and heat.