Topic: Why I think the 21Million hard limit will never be reached - deflationary spiral (Read 5975 times)

Yeah, if we define AI as human, then the only AIs are humans.  Pretty cheap to make and also pass the Turing test.

BTW phototropism implies consciousness in plants. 
A rock does a pretty good job at sensing its environment and doing just what it wants to do.  For example the moon can sense the local space-time metric and wants to orbit the earth. 

http://en.wikipedia.org/wiki/Moravec's_paradox

Creating an AI capable of exceeding human limits of intelligence is almost trivial, particularly when confined to particular definitions of intelligence such as aptitude at the game of Chess.  But until we understand ourselves, we can't hope to teach a computer to do things that come naturally to an infant.

http://cvdazzle.com/

Therefore, the architect is as doomed by smart computer systems as the Disney animator, but the skilled laborer is irreplacable.

Well I think a lot of people in this forum expect flops per USD to eventually go down, as USDs in circulation grow faster than transistors per square mm.  There are some physical limits to the number of transistors one can build with a few atoms.. whereas there is no physical limit to the amount of USD in your bank account. 

Perhaps the singularity you refer to is best represented by the amount of memory used by my music player.  Soon my MP3 player, with decreased functionality, will use terabytes of ram.  Great!  Technological progress!  Oh, did I mention the music will also be at best the same? 

One could easily argue that an A.I. more intelligent than humans was created 40 years ago.  Certainly multiplication of large numbers was handled much faster by machines 40 years ago.  And yes, we use computers to design the next generation of computers.  So does this mean the singularity already happened?  What explosion or event horizon did we see? 

I think there's a theoretical reason to add bits to the protocol but it isn't at all related to the block reward.

Think of the following:

A: There's a hard limit on the total amount of atomary value units in the bitcoin system.
B: Every year some bitcoins are lost forever.
C: It's very likely that the proportion of all bitcoins left that are lost in a time unit (eg. one year) has a lower limit.

If A & B & C, eventually all bitcoins will be lost.

Of course, this will take a long time. And by adding bits the lifetime of bitcoin can be expanded indefinitely. So, we need to add one bit per "half of all bitcoins lost" time.

I'm just referencing an already documented theory.

technological growth can be measured in a few ways, for example: computational capacity per $. The interesting thing is, whichever metric you use, the growth is exponential.

The basic idea is a self aware concious A.I more intelligent than humans will be created within the next forty years.  Which could invent a superior version of its self which in turn will invent an even cleverer version and so on.  An intelligence explosion and event horizon for mankind.  Just as Moore's law is predicted to fall off the singularity will kick in.

Wow that is a lot of hand waving.  Maybe we can fly soon.

What do you mean by "technological growth"?  it looks like you are using a scalar quantity "f" in your equation.  What does it represent? 

I don't see it mentioned or I may have missed it, but don't forget the transaction fees. They may one day be more than the mining block reward.

It may or may not be, but I believe the sheer amount of computing that will be required to get that last BTC out of the system will be impressive.

Off-topic posts serve purposes...
I was racking my brain attempting to come up with the explanation you just put forth.  DUR; Kurzweil's "singularity" is not meant to be a singularity in the strict physics sense

Well... a few weeks after that point, it will be x2 per second. So, close enough. It's just a trend line, not an exact mathematical exponential function.

I don't think "singularity" is meant to be taken literally. The idea is that growth, at some point, will be so absurdly fast that it will represent an unimaginably profound shift in society/human existence. Think of it in terms of "paradigm shifting events". See this resource to get an impression of this:
http://www.singularity.com/charts/page17.html

All it's really saying is that the next "event" is coming within our lifetime. Considering the exponential nature of technological growth (growth is proportional to size. f(x) = df/dx), it seems pretty inevitable to me.

Sorry for being so off topic.

I know this isn't the speculation board but I can bitcoin spiking to at least $9 and maybe as high as $14 in the start of next year before settling to $7>$9 and then slowly growing again.  That's just how it looks to me tho.

I think the reward drop is a strong incentive to some people to invest more in to bitcoin then they normally would.  Like the long term bitcoin investors as its better to buy more bitcoins while there only $5 then potentially $10.  Although I suspect a lot of the new buying power that is keeping bitcoin at $5 will be speculators looking for a quick profit early next year.  So it may take until after Easter next year for price to settle after the volatility of speculators dumping a lot of bitcoins as soon as bitcoin shows any major flattening of growth after January. 
So to cut it short I think the bitcoin price will become very volatile in the first few months of next year.  As it seems to have been pretty stable over the last twelve weeks and around $5 since the start of the year.     

The new coins each day are being absorbed by new demand and the price is still for quite some time. Assuming new demand keeps showing up at about the same rate the halving will have a strong effect. The only way I see this not happening is if a lot of current demand is thinking this way. In that case demand would drop around the halving and people thinking like that will start to unload.

But I think it all gets swamped by the demand that's going to show up as people realize that bitcoin is here to stay.

I hope it is $15 but the block reward half'ing in January will take a while to have a hit on the amount of bitcoins available.  I know it will half the supply of new bitcoins entering the market but they will still already be 9,000,000 of them in circulation.  So I think it will take a good few months to filter through but it might make speculators buy more in the months before therefore increasing the price now.  What I think bitcoin needs is to be used to back local community currency's, in places like Africa for instance were they currently use Mpesa and also for something like BitInstant to really compete with PayPal or Western Union and eventually SWIFT and/or VISA  

Alright then.
Keep in mind that the site just quotes the book on the subject. So this is an "official" position.

I can't tell if it is possible.
It also could 12233x per month or 0.12x per second or whatever.

But that's irrelevant hence the above definition of it. What you are proposing is something else, just a point on a slope not a singularity (hence the name (!) )

The block reward cuts in half in Januarry of 2013, so the inflation rate literally cuts in half.  I suspect, then if things just keep plodding along as they have lately, somthing that I consider very unlikely, then the value of a bitcoin in December of 2013 would be closer to $15.

I put bitcoin and a city of 235,00 people called Salford which is within a larger conurbation (city) of Greater Manchester of 2.4million people and the 235,000 population city of Salford had a higher average trend then bitcoin.  I currently think the amount of people holding some bitcoins is below 100,000 and all the people who have ever even heard of bitcoin well below 10,000,000.  Long way to go but if bitcoin did catch on as a major form of payment transfer rather than mainly just a commodity that it is now then each bitcoin could end up worth over £10,000 (not including inflation adjustment) each within our lifetimes.  If bitcoin adoption just keeps continuing at the rate it is now then it should still beat the inflation of the £/$.  I'm guessing a bitcoin will be worth up to and below $7 at the end of the year and up to and below $10 at the end of next year if things just keep plodding on as they are.

Ever tried to google "asymptotically approaching" ?

That's a damn big IF.

Very good observation by OP, if we could continously add decimal places after each block reward drop, and BTC value doubles at the same time, then the amount of BTC value added each year will be constant

At the same time, it means that the existing BTC holders will get their holding value doubled at each block reward drop, which gives them even more incentive to hold BTC as an electronic medium of saving, and when they accumulated enough amount of BTC and go for retire, then they will get ever-increasing value of saving to spend

For example, if I'm holding 1000 BTC as my saving (worth 1million dollar in future), and every four years I spent half of them, then my consumption power will keep the same all the time

Then very natually, many pension fund will consider their investment in BTC, because the scarcity of this medium will make sure the value will be ever increasing, at least better than some government bonds, 100% return in 4 years are not bad

No matter what, BTC has the potential to bring some craziness, but first it has to gain more acceptance and understanding

You're only thinking in one dimension. For example, we could be mass producing cheap human equivalent AIs, which communicate with each other through networks. They could, in turn, produce who-knows-what.

If there is a maximum intelligence, then the singularity is impossible. This is a tautology.

@mucus
"Technological Progress Grows Exponentially and Reaches Infinity in Finite Time". This is a contradiction. The exponential curve reaches y=infinity @ x=infinity. So the referenced page fails to define exponential growth.

You can't have "infinite progress in finite time". I agree that this is impossible.

Lets define the singularity as "technology growth reaches the rate of x2 per day.". Do you think that this is impossible?

In this reasoning you're implying that the value of bitcoin will, in average, double at each 4 years, so that the block reward inflation keeps having the same value.
That sounds too much optimistic to me.

This is not to agree or disagree with either side in this argument, but I believe that the "smarter" part is important because of this reasoning (and "smarter" is used to mean "more capable of designing artificial intelligences):

If humans of intelligence level X, manage to design an AI of intelligence level "X+1", then the machine of intelligence level "X+1" would have been able to design itself (since it is smarter than the humans that did design it).  By implication then, it would be able to design an intelligence of level "X+2", etc, etc.

Personally I find the logic of a technological singularity lacking; I don't really see that the implication is true: who says that "X" intelligence is always capable of designing "X+1" intelligence AIs?  Perhaps it gets exponentially harder to make smarter and smarter designs, in which case there would be no rush to "infinite" intelligence, but instead a rush to "maximum intelligence".  What's more, what if humanity is already "maximum intelligence"?  The assumption that we can design an intelligence cleverer than us is just that: an assumption.

I cannot appreciate the implications.  The reason why is that "smarter" is not clear.  Does this mean better at doing fast mulitplication?  Machines are already smarter.  Holding memory?  Computing hashes?  Already smarter. 

And what is this so-called progress?  Desertification?  Toxification?  Squandering of resources?  Or just the spread of mental illness?   

it does mean infinite progress in a finite amount of time.
The theory states that literally, the term is not by me.


sauce: http://www.mindfully.org/Technology/2003/Singularity-Bell1may03.htm

The Technological Singularity is to Cybernetics what the Perpetuum Mobile is to Physics.

The technological singularity does NOT state "infinite progress". You made a nice mess of that straw man though while dodging the actual question. Do you think we will not soon produce machines that are smarter than humans? Can you not appreciate the implications of this?

The singularity has already long happened. [vid]

The singularity states that infinite progress will be reached within a finite amount of time.
While I agree that the potential for progress is infinite, I see the above statement as impossible.

If you lookup the term I mentioned above (The Omega Point) there is but one possibility this can happen.
But this is metaphysics, not science.

For what reason do you believing the singularity won't happen? We are already developing rapidly in the field of nanotechnology, AI, 3D printing, genetics, neuroscience ...

As far as I'm concerned, the the only necessary element is production of AI machines at or greater than the level of humans. Once we can cheaply produce machines that are smarter than ourselves, or at least augment humans to be much smarter, technology will explode and I don't see any limit.

I've bookmarked it but on the Pilsner at the moment and not into heavy reading.  Thanks.

idk

here you go knock yourself out:
http://humanityplus.org/

"Pussy out" of what trying to prove I'm right and your wrong?  Or should I just bow down to your genius   

Having an idea and using it as a world view are two different things.
The people who make predictions on a time period where that should occur have no academic relevance.

OP has been thoroughly answered above, don't pussy out of it now.

Alan Turing who invented the modern computer was no failed academic and was probably the first person to come up with the idea.  Anyway I don't want to get too off-topic talking just about the Technological Singularity.  

This has been predicted numerous times and revised. It's not gonna happen and is more of a pseudoscientifc joke to anyone involved in natural sciences or cybernetics.
If you wanna have a entertaining read look up The Omega Point, aka The Singularity.
The idea originates in the transhumanist movement, a circle jerk of wanna-be elitists who couldn't make it in academia.

Yes but I think people have said that the processing and storage needs of the blockchain will be able to keep increasing due to Moore's Law.  Although I don't think Moore's Law can continue as it is and will inflex (excuse my terminology[tho could be correct long time since I did any maths]) during the Technological Singularity (https://en.wikipedia.org/wiki/Technological_singularity) predicted sometime after 2030-2040.

Only if there were an infinite number of bits available for storage.  When the block reward is 1 satoshi; then after 210000 blocks, that 1 will be shifted right and the reward will become zero.

Perhaps by then the developers will allocate more bits, but it's still not infinite.

All that being said: I completely agree, this still doesn't equal deflation or inflation.  There will be, near as makes no difference, a fixed number of bitcoins in the economy.  If the economy becomes more productive, then (unlike now) that productivity increase will be reflected in lower prices, as everybody's supplier will be able to supply more for less.

The main point I made this thread as at first glance it seemed to me and many others have had a similar view that as the block reward gets very small that it equals no new bitcoins therefore hyper-inflationary spiral.  The other point being as long as bitcoins value keeps increasing (maybe like gold circa up to 1918AD [start of then end of the gold standard]) then a valuable amount of new bitcoins will keep entering the market.  Therefore killing of the idea of hyper-inflationary spiral once the block reward is very small.

Yes I see after infinite amount of time it will only get closer and closer to 21million.  The main point of the OP is the block reward never ending even after a block reward of only 1 satoshi and bitcoin value ever increasing.  Therefore no hyper-deflation.

Edit:  As a valuable amount of new bitcoins/satoshis always entering the market.

Absolutely right.

The 21M comes from summing the geometric series I mentioned above.  http://en.wikipedia.org/wiki/Geometric_series

1/2 + 1/4 + 1/8 + 1/16 + ... + 1/2^i + ... = 1 = S

Each term in this series maps to one block reward period.  i.e. 210000 blocks.  Now that first term 1/2 maps to the 50 coin reward, so we want to multiply by 100 as well.

Therefore we get S*210000*100 = 21M.

Being that S is the sum of an infinite series, we will never quite reach 21M as we will never have an infinite number of terms.

The only reason I mention adding more decimal places is for when the block reward is less then 1 satoshi (edit: and as bitcoin becomes more valuable) and yes its the deflationary spiral I'm on about not general deflation.

Well no none of that is true.

I don't think deflation will be a problem but your reasons don't make sense.  Adding decimal places isn't increasing the money supply.  Deflation WILL occur I just don't think the deflationary death spiral idea has ever been proven.

Also I don't think there will ever be a time when 1 satoshi will be equal to 1 BTC now.

That would put 1 satoshi = $5 and the Bitcoin money supply at ~ 10 quadrillion dollars (or ~100x the size of entire global money supply).

In the future if 1 satoshis is as valuable as 1 bitcoin now then the block reward will still be valuable when the block reward is only 50 satoshis.  The point of the OP was as long a bitcoin keeps increasing in value and more decimals keeping getting added then deflation will never be a problem.

MoonShadow is correct.

The method used for determining the reward is to do a right shift on 50,000,000 satoshis every 210K blocks.  That truncates any fractional satoshis.
There are some truncation along the way.  For example starting in block 6,090,000 the reward is 9 satoshis.  In block 6,300,000 it becomes 4 satoshis so the block reward declines by more than 50%. Block 6,930,000 will drop the reward to 0 satoshis however only  2,099,999,997,690,000 so it looks like we will be eternally  2,310,000 satoshis (0.0231 BTC) short of the magical 21M.

If Satoshi had wanted it to be an even number the starting block reward should have been a multiple of 2.  For example is the first block reward had been 2^32 satoshis (~42.94 BTC) then we would have had 32 perfect halvings of the reward.  2^32 down to 2^0.  A bit of trivia, the "21M" appears nowhere in the code.  The entire subsidy calculation is just a handful of lines.



Even IF decimal places are added (a move of dubious value) we will hit 0 block reward again before hitting 21M total minted.  I doubt decimal places will be added.  If the entire global money supply was in Bitcoins and no fiat anywhere on the planet existed 1 satoshi would be worth about 5 cents which isn't a limit on commerce. In any scenario short of single global currency 1 satoshi ends up being worth various fractions of a single cent.

Lastly adding decimial places wouldn't change the rate of price deflation by any material amount.  Even if you added decimal 4 places that would allow you to pay out ~ 2.1 million satoshis over the next 3 decades.    99.9999999% will still be issued by the time the block reward originally would have hits 0.  Spreading the remaining 0.00000001% over a couple decades means the money supply growth rate is something like 1 billionth of 1% per year.

0% vs 0.00000000037% = totally meaningless a token change at best.


 Essentially 0% (which is what it will be when block reward is 0).

This argument is a matter of semantics.  The 21 million btc limit is a limit in the calculus sense.  technically the limit will never be hit no matter how many decimals are added to the current protocol, but for any practical reasoning it matters little.

Not necessarily; 1/2 + 1/4 + 1/8 + 1/16 + ... is famously 1, not infinity.  (If it's not obvious, bitcoin is simply a large constant times this series, so it too reaches a finite total, even if it had infinite precision)

Also; let's think practically.  When the block reward is 0.000000001, then it will take five billion blocks to make the same number of coins as 1 block does now.   That's 95 thousand years.  For all intents and purposes, when the current BTC resolution is used up (i.e. 8 decimal places) there are effectively no more bitcoins being generated.

If more deciamal places keep getting added in an infinite amount of time an infinite (edit): closer amount of bitcoins to 21 million would be created.  What my main point is there is no need to worry about the 21million bitcoin limit being reached.  No bitcoin will just resemble the gold or a very rare earth mineral market.  So no need to worry about bitcoins stopping being created and a hyper-deflationary spiral.

what?

You said some things that don't have any thing to do with the protocol. I fail to see the importance of decimal places.

Is easy. reward is 50BTC, 25BTC, 12.5BTC, 6.25BTC...., not exactly 21 million total but very close.

The only way this will not happen is if bitcoin dies or community decides to change the rules (not gonna happen with this rule because it can lead to abandon of this currency)

I don't think there is anything in the bitcoin protocol that states on exactly 21Million bitcoin will ever be made.  No just that the block reward will keep half'ing.  So if bitcoins value goes up and more decimal places are and keep getting added as the value of bitcoin goes up then the block reward can keep getting paid.  Even if the block reward gets really small as long as the value of bitcoin keeps rising and more decimal places are added then a valuable block reward can keep being paid.  At some point in the future the amount of newly created bitcoins entering the market will be similar to the amount of gold entering the gold market.  As where we are at now with such a high block reward is like a gold-rush.