Topic: The state of crypto - The only serious thread on the subforum - page 4. (Read 19463 times)

That's silly. It's like saying that everyone has to do hashes with pencil and paper.

The reason ASICs are a problem is that very few people have the ability to develop and manufacture them so it is a cocentration of power. There is not really a sybil attack, since no one is claiming otherwise. We are perfectly well aware there are only a few ASIC developers and there will probably be fewer in a year or two.

An ASIC is one. One chip, custom designed to only do hashes. That's disproportionate power in its essence.

No the essence of sybil is that you can make one appear to be many (and get more votes or whatever benefit it is from appearing to be many). You can't do that with ASICs. One is just one.

Everyone is using ASICs now so there isn't really anything disproportionate about it.

Actually, I say the reverse is true: ASICs represent one CPU, but have disproportionate voting power, that is the essence of sybil.

With 1C1V, pools represent the aggregation of CPUs, which is the very best case you can hope for - linear scale. Of course, they still do represent a different kind of sybil, which is bad in its own right.

The reason this seems to be true is due to the adjusting difficulty. Difficulty must adjust with increase in mining power, otherwise it becomes easier to sybil attack the network and otherwise the block rate would increase, leading to even more inconsistent confirmation times. The consequence is centralisation, but not as the primary design.

first i've heard of it, quick read still a bit muddy on it but will look further.

you a fan of the DAG thing?

https://bitcointalksearch.org/topic/dagcoin-a-cryptocurrency-without-blocks-1177633

smooth had another link somewherezzzzz....

on a seperate note but inline with the thread title.

whats peoples thoughts on this;

http://cryptovoter.com/

i say litecoin is the best we have, but yeah always looking for better, again i think the whole blockchain idea is a dud (and you guys yourself point out the many flaws).

but i'm very much interested in the dev of a p2p currency, course i have random thoughts and ideas, and hoping others put up ideas to see if something can come out of it.

this section of the forums about alternatives to bitcoin. yet almost no one can get their head outside of the blockchain.

Can you be more specific?

All I can recall reading from you is how LTC is the best and how everything in cryptocurrency (including LTC?) is terrible.

What do you think is a good model?

This thread is going in circles.


anyone willing to think outside the blockchain model for a p2p currency?

Hey, all I did was re-use your own example of a giant, underwater Rube Goldberg machine PoW mining away with an unlicensed nuclear reactor.

Of course it is. As I explained to you, (D)PoS can't form a consensus at all. PoW, whether delegatable or delegated or not, is strictly more powerful. Whether that difference in power is valuable or not is a different question, but you can't correctly argue that the two are equivalent.

Also, as myagui suggested and I clarified (and stated earlier), the "best" place to mine is not unique. There is specialization and different cost functions of scale in different locations and by different operators. Your idea of one miner in the perfect place makes as much sense as the old soviet union which had one enormous pipe factory which was theoretically the most efficient due to central planners' conceptions of economy of scale but in fact turned out to be a disaster because it not only wasn't producing the right pipes but wasn't even the right scale. (A hypothetical statement about the absurdity of unconstrained economies of scale; I don't know that the USSR literally had one pipe factory but I've read real examples of this fallacy actually being put into practice.)

You gave an example of this right in your post (climate). Since weather is not constant, the best place to mine is not constant. Unless you are going to have one miner who moves his gear around chasing the best weather to mine right now, you are going to have multiple miners who operate at different intensities at different times in different places. Likewise cost of electricity, space, construction, maintenance, security, management overhead, bandwidth, etc. There is just an enormous amount of multidimensionality that doesn't resolve to a single point once you get your head out of that asinine one dimensional "lowest energy cost" model.

Oh ffs smooth, it doesn't matter if the mining is occurring at the bottom of the ocean or in outer space.  Since energy price and climate aren't a constant, it still creates a constantly diminishing working set or complete monopoly.  This Rube Goldberg machine you've built on the ocean floor is just an additional bonus to the fact that delegated proof of work is providing no benefit to decentralization compared to DPoS in the first place.

That's not even very much of a dream. Anyone off grid (or even on-grid when the grid has a surplus) with solar will likely have free electricity some of the time. I'm pretty sure that solar is getting cheaper faster than any other method of energy production, so this may be common soon.

The pool sybil is due to pool mining, not ASICs. Even before ASICs, there were pools nearing 50% hash rate, outcries, and pools subsequently "shrinking" (though that may have been a sham).

What ASICs do is concentrate mining (at least indirect control over it) in the hands of ASIC developers as a sort of hydraulic empire. That may break down as hashing ASIC technology becomes commoditized, which seems somewhat inevitable.

The dreamer in me thinks that before either of those exotic scenarios, we'd have free & universal access to electricity, 'cos you know, Sun ... and non-destructive, non-pollutant energy storage methods ...  Anyhow ...

I do agree that PoW has flaws, in my limited understanding of the matter.
My question however, was to how DPoS compares, specifically:

I noticed that the earlier link points to an interesting discussion about DPoS, will read up on it.
Cheers gents!

Well that's a stupid assumption and is the biggest and most challenging to fix design flaw in Bitcoin that I know about.

Space station mining doesn't really make a lot of sense given latency issues. Possibly low earth orbit would work. High earth orbit is probably okay for transactions but not mining. Anything beyond that will probably be separate blockchains. I don't know how Bitcoin maximalists cope with that inevitability.

Dumping heat in space is not easy or cheap either. Space is "cold" but without conduction you are limited to radiative cooling. This is a non-trivial design issue for spacecraft. Under the ocean is likely a better exotic outcome to propose.

Yes, assuming the miner has enough hash power to solo mine, and since neither energy price or climate is a constant, is willing to travel anywhere in the universe where it's cheapest from China, to Antarctica, to possibly outer space to remain profitable.  The system is currently designed to create a constantly diminishing working set, and anyone who has ever mined before (me), knows the succesful miners are the people who constantly parlay their profit into the next mining hardware over and over to infinity (designed to force a small number of monopoly players or complete monopoly).

Assuming transaction fees provided the incentive, the end game mining market would eventually become so asinine, people will say, "awesome, I can build an unregulated nuclear power plant in outer space and utilize the 2.7 kelvin temperatures to lower my cooling costs at the same time forming the perfect monopoly". 

Then we can have space colonies whose only purpose is to tend to the mining equipment where 90% of the hash comes from one miner and the other 10% is some guy that sees Bitcoin PoW as a religous movement.

I have a feeling that since delegated proof of work doesn't provide any decentralization benefit over DPoS, PoW will most likely be abandoned before the first Bitcoin space station is constructed.

PoW doesn't require 1C1V. It only requires that (or more precisely requires that no actor have disproportionate voting power) to converge in the manner stated in the paper. Nevertheless it will still eventually converge with weaker assumptions on rate.

Even if you can sybil at the level of miners or pools, you still can't sybil at the level of hashes.

So the pools in China could fork the chain starting 20 or 30 blocks back and double spend. But it would cost them 20-30 blocks worth of hashes to do it. Trying to do this say a year back, while theoretically possible since they have 50+% of the hash rate, would not be feasible. Even if it were, and they did it, they certainly couldn't keep doing that repeatedly because it would cost infinite energy. Eventually they will run out of money and need to let the chain progress forward.

Thus there will still eventually (and remember, even if the ideal 1C1V case PoW doesn't observe a deadline) be an unambiguous longest chain, and therefore a consensus, although it may be a slow and messy one.

By contrast, in PoS with a sufficient portion of old stake, you can create as many versions of the chain as you want without incurring this kind of cost. DPoS doesn't matter here because you can create new chains with different election results, and the proceed as above. (You can infer a cost if you view PoS as a form of PoW where creating chains is the PoW, but this doesn't help the argument.)

The only defense is "everybody knows which is the right chain" which is not a system reaching consensus, it is people forming it.