Quote from: xulescu on March 21, 2014, 05:02:28 AM

Quote from: TERA on March 21, 2014, 03:31:01 AM

Quote from: Twilight_Sparkle on March 21, 2014, 03:09:02 AM

Why is it so unbelievable? As long as bitcoin gains mainstream adoption, 1mil (of purchasing power) per bitcoin is inevitable.

1. The ~10 minute confirmation time
2. The reliance on set encryption protocols which could be broken in the future (EC and SHA256). For something to become as big as a world currency, it needs to be trusted that the technology would last hundreds of years.
3. The blockchain size.
4. The amount of resources that the client uses on a device.
5. Other scalability issues, such as the amount of transactions which can be handled per minute.
6. Irreversibility of transactions.
7. The extreme ease of the average person losing coins due one of many vectors of mistakes and hardware failures, or hackers.
8. The technical difficulty for the average user trying to even use it.
9. Bitcoin commerce and usage has mostly formed within centralized entities even though it is supposed to be decentralized.
10. There are still humans who ultimately control Bitcoin and who must be trusted by its users, even though the system is supposed to be decentralized.
11. Volatility - it is not an attractive store of value for someone who just wants to use it as a store of value or a currency and isn't into investing or speculating.
12. The present distribution of wealth.
13. The reliance on the internet in order to function.
14. Also, what happens if the internet is split apart due to government firewalls.
15. Governments, corporations, and hackers could ultimate control and restrict Bitcoin usage if they really wanted. And mass adoption will not be reached when users have to do it illegally, run tor, and run checks on every client they download to make sure it isn't a hacked client. Even tor traffic or bitcoin traffic over tor could be blocked too by smart routers.
16. Some of the worlds most powerful contries have already banned or restricted it.

** add more here ** - anyone else is welcome

Your concerns suggest that no crypto-currency will ever make it. I agree with some points and disagree with some:

1. This is a problem for micropayments but any global decentralized system has a fuzzy notion of synchronization that requires a 30-120 second lag at this point. For example, compare how many orphaned blocks BTC creates to how many LTC creates. Maybe 10 minutes is too much, but 2 minutes is too little. I don't see changing much in the future barring A Breakthrough (TM) because as the internet grows, lag-per-bounce slowly creeps down (with physical limits given by switch reaction time and speed of light / EM field in copper wires) while the number of bounces should creep up (logarithmically in the size of the network maybe? I'm assuming that the internet is roughly tree-like). As reserve / settlement medium, this is not a problem.

2. This is a BIG problem that will not go away. First remember that computing is a very young field. EC may have a backdoor already in the "hand picked" nonstandard constants and whether true hashing function (one-way's) exists is still an open question (more so, it MIGHT be equivalent to P vs NP, that most likely we don't yet have the logic and mathematics required to probe). No current tech will last hundreds of years (probably not even 50 years for the most basic software tech, much less for more high level software tech -- and see how x86 is becoming obsolete already after ~30 years). Actually, think about ANY tech that we're still using after 100 years. The lightbulb? Filament bulbs are already obsolete. The radio? I believe (barring global cataclysms) that tech evolves exponentially. You can extend Moore's law back to ~1880 if you accept mechanical computation as well. I know trends hold until they don't, but Moore's law held a few paradigm shifts already (springs, hydraulics, vacuum lamps, transistors, IC's, networks) and the physical limits for computation are still unimaginably out of reach. True, we still don't understand computation much better than the Truly Illuminated People like Cantor, Hilbert, Godel with his friends, etc. But none of them realised using computation is a much, much easier thing than understanding it. We have hundreds of Turing-complete models nowadays and people already started working on (1) super-Turing computation (statistics / machine learning in the limit is the best model so far) (2) more efficient hardware for other Turing-complete models (no computer actually is a Turing machine and von Neumann RAM machines we all use are starting to not scale well with code complexity) (3) quantum computing which is (in its generality) actually waaay into the future now. Tech evolves fast and that won't change any time soon. Kurzweil and Kaczynski got that right, even if they're in general completely mental.

3. This is not really a problem. On one hand, not everyone needs to use a full client. I would trust something like a transparent nonprofit with relaying the blockchain correctly. On the other hand, trust issues are much more problematic than the blockchain (DNS, CA, my software, my compiler and OS if I write my own software, my hardware, etc). It would be great if this could be bootstrapped, but most likely it wil not happen for the foreseeable future. Maybe NameCoin will catch up, but likely not. CA's are going to be a problem until everyone creates a WoT (not gonna happen as long as Johnny Can't Encrypt). People do not understand that privacy, trust and security cannot be achieved with plug-and-play methods and actually require a lifestyle change -- at the very least, constant active defense. And mobile devices are much worse in this respect than good ol' desktops and laptops.

Also, storage per dollar will most likely still grow exponentially for the foreseeable future.

4. This is not really a problem. A RasPi can easily be a nonmining node.

5. Same with 1. Micropayments vs reserve/settlement.

A worthy mention here is the electricity amount needed to run the network. This is the main reason I do not believe BTC "will be it". Even conservative estimates now are getting out of hand. It's worthy to compare this to the maintenence cost for the USD as reserve and I think BTC scales terribly in this respect. I'm not saying LTC is the answer, but is the right idea. Bind PoW to more kinds of resources (cycles, RAM, maybe even bandwidth). Make it so that it's not much more expensive in total for a million small peers than one large peer with a million peers' power. Cycles does not cut it because the economies of scale are for large clusters (this could change with ubiquitous computing but that's still far into the future). RAM is better because it doesn't use much power, you just need to have enough. On the other hand, you need to adjust RAM requirements as well as difficulty, because RAM also gets cheaper exponentially so if you don't, in a number of years it becomes a complete non-limitation. AFAIK, no alt scales RAM requirements. Another option is PoS but PPC didn't get it right (and if I knew what is "right" I would do it myself).

6. Not really a problem. Most of the world did not forget cash yet, which is just as irreversible.

7. This is a problem and refers to 3 above (Johnny Can't Encrypt). On the other hand, people are willing to trust one entity to do it for them. I know that If You Are Not The Only One To Hold The Keys You Don't Have It, but the same can be said about cash vs digital bank IOUs. People don't care and a transparent insured service could do the trick.

8. Same with 7. On the other hand, the average user slowly becomes more savvy. Think about the history of e-mail or internet, cars, microwave ovens etc.

9. They are required in the early stages to provide nucleation. If/When the system matures a bit, the low barriers of entry mean a more free market so this problem slowly solves itself. Banking now is ridiculous everywhere (why can't I open a bank with my own rules, clearly stating that I'm not insured and be transparent about my finances and activity? consumer banking SHOULD be completely transparent, there's no "know-how" there).

10. This is a problem. The Foundation is the first example that comes to mind. Until/Unless bitcoind becomes as audited as, say, debian, this will still be a problem.

11. This is a problem and I don't know what to say about it. I don't believe volatility will get lower on its own as the network gets larger. The network would need ample cushioning with derivatives to achieve this, but who is going to do it if there is no One True Price? In the absence of exchanges, there is no One True Price for BTC. Maybe what we need is embrace price discrimination. Maybe full reserve and savings buffers means everyone can accept more short-term volatility if long-term the price tends to move slowly. But then, derivatives has the same leveraging effects as fractional reserves. I believe a general solution to this would be a true paradigm shift from growth to stability. So probably not happening in the next century, if ever. Maybe when we join the swarm/grid things will change.

12. A problem for the world at large that is the same problem for BTC. I don't think BTC is worse. Maybe LTC but I doubt that too. The distribution of wealth will always be exponential, only question is, how skewed.

13. Obvious problem, but then again, so is 90% of modern world. Fuck it, even driving a car will probably need a permanent internet connection in the next 10-20 years IMHO. All problems of the internet/computing world in point 3 above affect cryptos too. Namely, not all internet is decentralized, not everything in the internet system scales well, not all internet is trustless (DNS, CA, infrastructure -- including protocols like IP -- are the largest problems).

14. This is not a problem. Firewalls can be pierced/avoided and a physical separation will never happen. Even if it does, someone will plant their own fiber in the oceans and reconnect off the record. Even if this doesn't happen, a few node satellites could solve it "forever". Either way, infrastructure cannot be decentralized because the internet infrastructure is not decentralized. Grid projects could solve this, but none is practical yet, and their adoption is a bigger problem than crypto adoption. Ubiquitous computing will probably solve this problem, but that's still SciFi. This is the internet's problem.

15. Same as 3, 7, 13, 14. Internet's problem. Also, a complete effective ban will never work (undecidability, etc etc).

16. This is not a problem. On one hand, China also banned capital outflows and how is that working with UnionPay? Technically speaking, every single one of us does a number of illegal things every day. On the other hand, if governments manage to control THIS, then BTC is clearly noone's largest problem. Citizen drone DEAD00BEEF00-1, move along.

IMHO:

A. The largest problem is the reliance on the internet. But this is a problem that affects (or can affect) everything in our lives to such a great extent that I would not consider it a BTC problem. We NEED a truly decentralized and scalable internet for many, many reasons not related to BTC. If one gets going, BTC is plug-and-play in the new network. This is why I do not consider it BTC's problem. At best, it's all crypto's problem. At worst, it's every one's and every thing's problem.

B. BTC itself's largest problem is electricity consumption. Second is EC. Third is that if BTC becomes reserve currency, it will get dilluted with BTC IOUs just like gold was (and vulnerable to the same things). Fourth, BTC is not actually good for microtransactions and a credit network like Ripple ( But Not Ripple , maybe ZeroReserve?) is way better.

So yeah, I don't think BTC Is It (TM). If it reaches 1M USD, then it's most definitely NOT 2014-USD. But it's a first step in the right direction. The next right direction IMO is toward efficiency: to realize that the blockchain is not DISTRIBUTED but simply just REPLICATED. In terms of efficiency, that is a great difference. Don't forget BTC is and has always been and will always be an experiment. It's not even beta, it's alpha at best.

And finally, I fully believe that a system that is entirely non-hierarchical CANNOT scale. You need structure to manage size. You need dynamic structure to manage dynamic size. Some "centralization" is inevitable in any large enough system.

This is an extremely well-articulated and thought-provoking response. My hats off to you. You've given me much to contemplate and I envy anyone who has the chance to sit down with you and pick your brains further on this matter.