Maybe this is bulls***, I am legit stoned at the moment so who knows. Also thanks Australian Government for pulling that stick out of your arse about medical cannabis. You guys are alright.
You may have noticed from my earlier post that I've gotten into crypto mining recently. It's going well and I'm looking at installing my first dedicated mining frame next week now that my solar is online. For context, I'm a retired software developer, businessman & computer scientist.
So anyway I think I have an idea that is a genuine opportunity for a person or group of people to become quite wealthy. Maybe its already being pursued somewhere. I'm too old and tired to do it myself so someone else please take it on, I just want to sell hashtime.
In the next few years crypto currency will likely move away from Proof of Work. There's a really interesting thing happening in the scene; the cryptocurrency world relies on the work that GPU and ASIC miners do but actively dislikes the paradoxically useless nature of the work being done. Crypto enthusiasts and developers are evangelical about the technology while the miners are more pragmatic, they just want someone to sell their terahashes of compute time to.
So sometime over the next few years a couple of things seem inevitable. Crypto will move away from solving fake problems to secure their networks and miners will be holding hundreds of millions (billions?) of dollars of programmable, generic computational hardware looking for work to do.
What sort of problems can be solved with such a massive distributed computational network? What kinds of cloud services could it power? What economic value is there in solving these problems or providing these services? Can the AI compute cores in GPUs be used in a distributed fashion? Could it be used to trade on the stock market?
Is this how we get to the Singularity?
That said I am kinda f***en ripped. Who knows :D
No YAPSBOB here!
I was more thinking about a massive, distributed Tensor Flow array that researchers or whoever needs that sort of thing could tap into on the cheap.
We've had similar things like Folding@Home and SETI, just nothing generic and monetised to service all that machine learning.
I'm thinking something like this.
What I'm really thinking is that by the time those networks are secured and the method to harness that distributed processing power is developed it will be rendered obsolete by quantum computing. State-based quantum computing attacks will then consume that distributed network and turn it against it's previous owners, further solidifying the dystopian surveillance state we are hurtling towards.
I still haven't played through Deus :( I had it forever on the old Wii U.
Not convinced that quantum computers will do much other than annoy the authors of SSL libraries and web servers but I'm likely wrong. My understanding though is that these machines do particular types of calculations extremely well, but won't be useful for other work. Also there are types of encryption that are not exposed to their sneaky strange computation.