Mapping Cancer Markers is a sub project of World Community Grid,a scientific research initiative from the Krembil Institute which uses the computers of volunteers to better understand and eventually develop treatments for lung and other types of cancer. Anybody with a computer can help them process data, no need to have a biochem or computer science degree.

If you're interested in contributing your computer's spare processing power, join us at /r/BOINC4Science. It can be configured to run only when your computer is otherwise idle, so it doesn't slow down your machine.

This is an excellent explanation. Also worth adding that Ethereum moved to "proof of stake" a while ago, so nobody is buying GPUs to mine Eth at this point, whatever effect they have (if any) on current GPU prices has a time limit on it. There are other cryptos which use GPU mining, but they are not particularly popular and couldn't absorb much of the miner exodus when Ethereum switched to PoS.

We don't have to wait on the government to implement basic UBI, cryptocurrency can make such a scheme turnkey. For example, you can make 1% of transaction fees get redistributed to users, you can make rules based on wealth of different users, etc. You can vote on-chain about changes to the rules. For example, 1% of fees can be redistributed to users, 10% of fees can go into a common healthcare insurance fund which anybody can request funds for and be reimbursed from, etc. 100% transparent, no middlemen, free to anybody who wants to join it, and completely uncensorable. There are cryptos already working on a model very similar to this. This is one of many reasons why I am very excited to see what DLT brings us in the next 10-20 years, it has the power to fundamentally change society if people can get past "I saw that crypto was a scam in an article title once".

Instead of all being stuck in whatever economy we are born in, we can choose to participate in an economy that reflects our values, whatever they may be, and we can be in multiple economies at once.

Unfortunately AI as it stands strongly favors those with large resources, ML training is not something that works to distribute well because it benefits greatly from low latency. AI isn't a map-reduce problem or any other problem type which works well to split up into discrete parts to distribute for computation. MLC@Home if I understand correctly was training small single models on each machine/work unit, not training models in a distributed fashion.

been running BOINC for years, love it!