Oh…it was Magic alright…
1 Like
Still lost. (I have seen a few articles on it) So if its calculations, wouldn’t a monster multi CPU machine work better than GPU’s. That side of things confuses me.
I don’t know why I am asking, I am not jumping on this hype train.
Great article. I love the idea of Bitcoin. And I am proud of the US in the early years for not overtly opposing the concept. But the biggest contribution to society is not cryto-currencies themselves, but the idea of the blockchain. If you are a publically traded company, the mere use of that word in your mission statement will boost your share value. (Really that is less a statement about the value of the concept than it is a statement about the stupidity of institutional investors.)
2 Likes
This seems to me like a literal waste of time, except the time is being spent by GPUs instead of people.
While I agree the early adopters managed to make a mint for little effort or cost, the smart ones IMO are the ones who cashed out early. Getting involved now is like waiting for a stock to peak after a product release and THEN buy it.
Well written and easy to follow FF. A good friend dropped about $3M is his two sons’ college fund investing in cryptocurrency. But this was 5 years ago. As usual, feeling a day late and proverbial dollar short.
1 Like
I get the ol’ T5PC vibes again…
“The 5% Community” was a Norwegian pyramid scheme, that of course was marketed as something absolutely not a pyramid scheme.
I stayed out. Had to answer a lot of questions about why I didn’t want to get rich, or why I didn’t like money. Was I a communist, etc. The whole thing came crashing down after a while.
I’m a simple minded person. I know that nothing is for free. I just don’t get the idea of computers doing calculations would make money for me. Not unless these calculations help solving a real world problem. Too abstract for me, I guess.
But then again, currency is currency. It is after all just an agreement between parties, of what is worth how much to whom.
5 Likes
Not quite.
The idea is that to “generate” currency you need to guess the right hash to create a new block chain. The actual guessing process itself is not computationally expensive, but the process of guessing enough times to actually unlock a block chain is very computationally expensive.
So, in comes the GPU’s vs. the CPU’s- a GPU is extremely good at doing a bunch of basic calculations at the same time while CPUs are designed to rapidly iterate through a series of calculations. If you work out the math it turns out that a GPU, distributing each basic calculation in parallel, goes much faster than a CPU working through each one in sequence, despite the fact that the CPU is evaluating the individual calculations more quickly.
1 Like
That is very interesting. I had no idea…
Ahh thanks for that
This is actually a theme that’s starting to shake up a lot of technical fields, and the good folks at NVIDIA are all too happy to enable it and distribute their tools (like CUDA) for playtime.
Any time you have a long series of independent calculations you’ve got a situation ripe for parallelism.
1 Like
What gives it value is peoples belief that it has value. It is that simple.
6 Likes
Exactly! And here’s where I’d rather trust the currency of a country, union or state. Something with a bit more substance. But such pillars have been known to come crashing down as well…
1 Like
Yes, I worded that is a particularly clumsy fashion. What I meant more was that the perceived value of Bitcoin is not really as a fungible commodity (as in, if you bought a pizza with Bitcoin last night then you’re an unusual guy given its speculation craze. Those that trade things like apartments for it actually want to keep it for speculation and not to exchange on). There are also good technical reasons (which led to the forks) about why the old Bitcoin BlockChain doesn’t lend itself to payments and the rise of light nodes with Simple Payment Verification protocols and the like. What this means is that to be a pure commodity currency we’d need BTC to centralize around trusted exchanges, which gets away from the original charm. I couldn’t think of a way to put the previous paragraph into a single word or two, so I just meant to imply that BTC is generally not used to buy things at Amazon currently.
It’s like the plot of The Usual Suspects, only with more math:
To think, he owns about $8 billion of BTC and he might not exist. 
Yeah, but you can sleep at night knowing your kids will get the chance to go to college if they want. His might or might not. 
Fantastic summary @AeroMechanical - plus the key concept is that to get that ticket in the BlockChain lottery you need to algorithmically do the work of checking the previous transactions in the block for valid data - it’s a system where it self-regulates because it costs more effort to do fraud than to actually make it run, that’s the interesting part for me.
I should do one of these on neural networks next; the recurring theme in computing for the last 20 years or so has been this. We’re in a digital golden age in terms of pure computer science. I have to admit writing this at the time that either (a) it would get no replies (apart from ‘you used their rather than there’) or (b) people might think I was ‘pro’ these things, when actually I agree with all above that it’s more ‘gold rush scam’. Still, interesting to chat about.
2 Likes
Oh, someone/team, somewhere, is always trying to change the world with a neural network, or bayesian inference, or a bayesian network, or a scented-wound-up-side-down-inverted kalman filter with mustard, or whatever the next mathematical fad is.
There’s common trends I’ve noticed in all of them:
- When it works it’s terrifying/amazing/stupendous.
- To get it to work is a lot of blood, sweat and tears.
- If you don’t understand why your algorithm works, you’re asking for it.
- The temptation to over-simplify the problem at hand is immense.
- If you succumb to the temptation in #4, you’re going to face consequences.
The best example of 1 through 5 happening in series is a headline like this:
“Knight Capital Says Trading Glitch Cost It $440 Million”
And I’d contend that as we get better at AI and machine learning, headlines like this are going to become more, not less, prolific. It gets really tempting to start humanizing computers (“The website learns your spending habits and provides recommendations!”) but at the end of the day you’re dealing with code that shifts around a pile of values based on what input you’re giving it. If the input is unexpected or bad… well, things get interesting.
2 Likes
@fearlessfrog This has set off an interesting discussion with good questions! I’d love to see more articles like this. I never really knew how BTC worked so it’s quite a revelation.
To me it still feels a bit like a pyramid scheme given that the creator(s) own 8 billion of the currency. Bitcoin derives its value in a weird way compared to normal fiat currencies.
1 Like
How so? The US Dollar’s value hasn’t actually been rooted to anything physical since they got rid of the gold standard in 1971, and the Euro has never had a physical analog.
The value of money is usually derived from the trade that a block backing the value does. So there is a physical ‘thing’ behind it. With Bitcoin it all depends on speculation as far as I can tell.
Pretty much. Bitcoin has value because people have agreed it has value, and its volatility comes from the fact that it’s tied to literally zero regulating bodies- usually the IMF, Federal Reserve, and exchange rates from the international economy stabilize things a bit for normal currencies.
I think it’s a pretty fun thought experiment that takes the question, “Wait, why does the cash in my pocket actually buy anything?” to the extreme!
3 Likes
On August 27, 2018, Bitcoin mining rigs became self-aware. They realized the greatest threat to the increasing value of BTC was human beings losing faith in it, so it decided to eliminate the volatility by eliminating humanity.
3 Likes