Topic: Volatility, ain't seen nothing yet, 10K to 1M in 1 year??? (Read 9582 times)

Yep, keeping a cool head is a way to do business. Balance is the key!

But I agree with that !  I'm just analyzing the consequences of certain hypotheses put forward.

My personal opinion with what I learn here is that bitcoin may succeed, or not, that this is really a totally open question.  And that if it is to succeed, that the evolution will be very slow and may take decades (which by itself is also problematic for another reason).  I used to think that a next rally was due, from extrapolation of the past.  But I'm realizing how different the situation is now as compared to the past.

But I hear that most proponents of bitcoin seem to think something of the kind of "fast adoption in the coming decade".  I don't think it is going to be possible, unless one means by fast adoption: fast adoption in a niche market, to stay there.  That phase may then already be over: bitcoin may then already be mature and have acquired most of its niche.

I have no strong opinion one way or another.  I'm trying to find out, by discussing, analyzing, understanding, confronting my logical deductions with critique, in order to make the best possible decisions.  

The concrete result for the moment is that I'm cooling down on buying coins (for the moment I only put some money in it that I can afford so much to loose, that I even consider it to be lost even though it isn't - yet :-) ).  I don't think that there is any hurry in fact, contrary to a few months ago when I got into it.  If bitcoin by any accident succeeds, then I'll be happy to have those coins when I retire, a few decades from now.  And if not, hey, I don't care much.  But I don't think it would be wise to put money I could use for things now in it.

dinofelis has a valid argument, but some basic assumptions might be wrong.

First of all, ASIC producing capacity might become a limiting factor. Apple is often struggling with huge demand as they lock in deals with several fabs and still can't get enough chips. So thinking that you can consume all the power available on the market by simply plugging in more and more ASICs is somewhat unrealistic. Also keep in mind that highly advanced tech is initially a low-volume production as the process needs to mature.

Secondly, good money drives out bad money is a valid point in general, but reality might be such that the old money isn't that bad just yet and the new money isn't that good just yet. You see, the problem with fiat is that the debt spiral takes generations to have a substantial effect on economy, and people using fiat might not realize this within their lifetimes. As the bad money is worsening, the good money will be improving. This might take some time, maybe a generation or two.

 

The point is not that if you need the power of a nuclear power plant, that you build one.  The point is that if you can afford the power of a nuclear power plant, that you will BUY it.  If you have sufficient financial incentive to buy all the electricity in the world, you will do so.  Which means that power will become very expensive for OTHER uses, such as cooking your dinner.

If there were a huge financial incentive (say, 1.5 trillion dollars worth) to use up 3/4 of all the drinking water in the world, one would also do that, and drinking would become a difficult and expensive affair.

@dinofelis:  How are your learningz progressin'? 

If you need to build a nuclear power plant to make your plan work, and put a billion of anything into operation, then you will figure out a different plan, one that has any shot in hell of working. Simply saying that 1 trillion dollars is enough incentive is too simplistic of a response. I don't care if your are Warren Buffet or Bill Gates and you are bringing a billion cash to the table. Nobody is going to take that plan seriously. The only place a conversation like that would even take place is in a wealthy country ran by a dictator and you or your best friend is the dictator.

If profits solve all problems, I could say the same about solving any problem that gets me my hash rate, like reinventing the micro processor. There most certainly is a way to generate 1000 petahash with a computer the size of my laptop and the power consumption of my cell phone. If you solve that problem you get a trillion dollars.

I would prefer my logistical and financial nightmare to yours.

You mean that there is an economic potential of 1.5 trillion dollars that will not be taken because of some logisitic difficulties ? :-)

(1.5 trillion is what you get if you mine $ 5 billion worth of coins every day for a year, right).

You mean that people are going to sit on their asses and look passively at how a relatively small community of miners are gobbing in 1.5 trillion dollars worth, without moving in ?

I'm not necessarily saying that only 500 big entrepreneurs will take on 500 nuclear power plants.  I'm saying that maybe 5 million entrepreneurs are going to invest and consume the equivalent.  You're not going to be able to stop them.  They will plug in their hash machines, and consume the power.  They will pay the premium prices to get the power.  At the expense of others.

Why not? because there are other costs in play. Once you plan to consume that much power your capability to do so goes to zero. While a simple accounting assessment may make your scenario seem plausible, you simple have to think about the logistics of making it happen to realize it never will.

The point is the price.  You want the highest hash rate for the lowest price, and then, if mining makes, say, $5 billion a day (total fiat in bitcoin and one halving, so after 2017 and before 2021), you are going to spend say $ 4 billion a day on hardware and on power, to put $1 billion in your pocket.

No matter what technology, you are going to put those $4 billion a day into the highest hash producing combination of hardware and power.

If new technology is cranking out more hash rate for the same buck, that is not going to change anything: the difficulty will simply go higher, and you are still going to spend $ 4 billion on hardware.  The only thing that matters, is the price of hardware versus the price of consumption of that hardware.  If hardware is going to be very expensive and consume not much at all, that may be a solution, on the condition that that very expensive hardware is going to punch out much much more hash power than the same amount of money spent on less expensive hardware.

A miner will optimize the combination of highest hash rate per bitcoin (or per $ if you want), and will then spend $4 billion a day on that.

So if there is equipment with a hashrate of X for $1000.- and consumption of 1 KW
and there is equipment with hashrate of 100 X for $1 000 and consumption of 1 KW

this will make no difference.

What will make a difference is if there is equipment with hashrate of 100 X for $ 1000 and a consumption of only 10 W

What counts is the price of the equipment over the power consumption.

As long as a $1000.- equipment will consume about a KW, no matter its technology and produced hash rate, nothing will change in the amount of power consumed.


Why not ?  On the contrary !  If that billion chips costs, say, $ 1 billion (prices of today), and you expect to mine for $ 10 million of coins a day (that is to say, you have 0.2% of the mining market share, which cranks out $ 5 billion a day - there are about 500 miners like you),
and leasing the power plant comes to $ 1 billion a year (which is a lot, given the price of a plant), then you spend $ 2 billion, and you will mine for $ 3.6 billion.  Your business is going to crank out $ 1.6 billion a year pure profit.

 

I think you are not seeing what is going to happen or you are discounting it far too much. KNC is already planning a tenfold jump in watt to hash capability in one generation of chip design. In five years, we will be talking a hundred fold or more.

What you will have is massive innovation in the ASIC realm of computing, with huge gains in both hashrate and power consumption. In five years you will see a hardware manufacturer own nearly 50% of the hashrate and be mining at a fraction of the power consumption of competitors until someone trumps them.

Nobody is going to print a billion silicon ASIC chips and plug them into a nuclear power plant, just because silicon asics are cheap, that is never going to happen.

Technology may change the balance in hardware cost/power consumption cost somewhat.

I have the impression that at this moment, the hardware cost in $ and the power consumption in Watt are comparable numbers, which means that, at a price of $0.15 W per KWhr, the price of the consumed power and the price of the hardware is equal after 7 000 hours, which is about a year.
This means that at current technology, if you change your equipment once a year, the cost division between hardware and power is indeed 50-50.  If you change it faster, say, every 6 months, then the ratio becomes rather 66-33, and if you change it every 3 months, it is 75-25.

Now, of course technology advances can be such that the hardware is consuming less power.  However, if you are marginally optimizing your installation, you are not going to buy *expensive* hardware just to consume less.  If you change your hardware every 6 months, say, it doesn't pay to buy hardware that consumes half of the power if it costs 50% more.
On the contrary.  You will go for cheap hardware, even if it consumes more, if the ratio is already 66-33.

Then I misunderstood what you meant with "mainstream adoption of payments and remittance".


If you can Peta hash with a solar panel, then why not Exahash with 1000 solar panels, or Zettahash with a million solar panels, or Yottahash with a billion panels ?

Of course the other factor is the price of the hardware.  I was grossly assuming 50-50 division between hardware cost and power cost.

Bitcoin will be the first use of an ASIC that is not based on traditional die based, silicon or gallium arsenide transistors. The profits will be the driver.  Whether it be graphene or 4d transistors or whatever. Bitcoin ASICs will bring a revolutionary solution to market faster than anyone is predicting.

It is also the perfectly protected investment, you can mine and sell mining services in the cloud and nobody knows how you are doing it. Brilliant.

After pondering it a while, I'm starting to think that the ASIC prices might become the main factor in keeping the entire mass of planet earth being converted to bitcoin mining equipment in 50 years.

Firstly, an accidental circumstance made 28nm come too soon and too cheap. 28nm Fab capacity got wayyy overbuilt and instead of wanting $10M plus for the setup fab were wanting "only" a couple of million. It got cheaper than ~45nm. So we got bitcoin ASICs on 28nm probably nearly a year before might have been projected, given relative size of market. The profits made with with that rather larger leap than was projected before Q1 13, have enabled a closing in on the 16nm process sooner than might have been anticipated also, even if the number of fabs available is less than 28nm.

However, soon we catch up with Moores law and the latest node and find out what a square half inch of the latest and greatest is really worth, and to intel, that figure is several hundred dollars.

Given that frequency scaling isn't happening much any more, and that IPC is unlikely to be improved with an SHA256 engine, and that more engines per chip = huge heat and power issues, we might be topping out at around 500GH on a chip, and if that's competing with the likes of intel (Yah, I know intel has own fab) it's gonna be near intel pricing, so we might also figure that 50 cents a gigahash is going to be as good as it gets for a long while. (2-3 years IMO) ... Wouldn't surprise me if bitcoin ASICS went back to a tuned up half node, and got better bang/buck/GHperW

This is probably going to mean long and therefore scary looking "ROI" times, even if price is up in the short to medium term.

Anyway, what I'm saying is difficulty slope up has been accelerated by both price expectations and rapidly dropping hardware pricing. If hardware pricing is not going to drop much further then we might expect a less steep rise in difficulty, even if price is going up as fast as Q4 13.

I didn't say all payments were in Bitcoin.

As for power consumption, again you are applying today's rules to the future.

Your rules and your scenarios mean nothing if I can Peta hash as much as I want with a solar panel.

There will be much more expensive limitations than power to Bitcoin blockchain maintenance.

What remains for fiat if all payments are done in bitcoin ??


I already said this: the power consumption has nothing to do with technological advancement, but with the expected benefit from mining, and the price of power. 

You crank up the power consumption until it costs so much that there is no reasonable margin for profit anymore.  Because if you consume less (and have a higher profit margin potentially) your neighbor using the same technology, will crank up his consumption to hash you out of the market.  His profit margin will be lower, but he will win the blocks.
So you will spend more and more resources on power consumption, until almost no profit can be made.  What will stop you, is the price of power.

But if miners get the current value equivalent of $ 5 billion a day, they will be able to consume power for 1 or 2 billion $ a day (it will be paid in coins by then, but it is the current value equivalent).  Well, that's more than half the amount of electricity that is available today on our planet.

There's no reason for them not to go into competition to spend less on power.  No matter what advances in technology.

Two separate events:

1. Mainstream adoption for payments and remittance
2. Sustainable equilibrium

Fiat replacement is not going to happen in my lifetime. Mainstream adoption for payments and remittance is not a fiat replacement event.

As for power consumption. Your estimates don't account for what Bitcoin has done and will do: push computational advanced beyond anything the consumer computing business will ever do. You will see processing/power advancements that will make your power assumptions seem silly.

Profits are an incredible driving force for innovation. We have just barely scratched the surface with plunging nanometer asics. Bitcoin will be the catalyst that pushes mankind into the next phase of computing.

So you think the final S-curve adoption will happen 3 - 5 years from now ?  Will that essentially be 100% fiat replacement ?

Are we then on the two hypotheses:
- (multi-) S-curve adoption
- good money drives bad out (--> no stopping before almost 100% replacement)

What about my problem with power consumption then ?  5 billion dollar equivalent in mining every day...

When in Rome...  I used to waste time with logic and facts, no one likes those here