Scientifically valuable alternative for bitcoin?

[jarekduda]

Cryptocurrencies like bitcoin are quickly growing social phenomena, which motivate huge amount of people to spend lots of valuable resources (manpower, electronics, electric energy) to solve complex computational problems ... which are otherwise completely worthless.
As from some perspectives it could be seen as pure waste of valuable resources, I would like to propose a discussion about directing this social phenomena to some socially useful tasks. For example we could change solving currently used problems, like finding a sequence with small hash value, into e.g. finding optimal protein folding, useful in understanding of life and developing new drugs: for given amino acid sequence, find its configuration (angle sequence) minimizing energy. So “mining a coin” for a given protein would be e.g. finding a configuration with essentially lower energy than current best. It would also allow to use (motivate economically) other approaches for “mining” like crystallography. Other kind of problems here could be finding optimal set of parameters for some algorithm...

So there should be some market with database of useful problems which can be attacked this way and can be well specified. Each of them would correspond to a separate currency – their number can be huge. In contrast, the amount of coins could be very limited – e.g. finding the best configuration would freeze the amount of “scicoins” corresponding to this problem. However, obviously they can be still exchanged and their value can still fluctuate – e.g. would probably grow if given protein would turn out very useful. It would also motivate people to focus more on science as a partial replacement for purely egoistic economical interests.

 Some questions:
-   what kind, examples of tasks could be used for such cryptocurrencies? They should be very well defined, cheap to verify, they should have a relatively large family of improvements like better energy minims for protein folding, maybe e.g. better algorithm/parameters for some useful problem ... Maybe it could be used for some more general than purely computational (scientific) tasks?
-   what kind of infrastructure would the best for them? Can be treated like bitcoins, but containing e.g. protein angle sequence instead,
-   how would their economy look like: huge amount of currencies with very limited amount of “coins”?

Generally: how to directly human egoism from focusing on socially useless problems, toward the valuable ones?

ps. A year old article says they spend 15 million dollars a day for electricity only ( http://www.forbes.com/sites/timworstall/2013/12/03/fascinating-number-bitcoin-mining-uses-15-millions-worth-of-electricity-every-day/ ) ... and it will grow - cannot we direct it to something useful?

[Bitobsessed]

Here is just one.  I have not followed all of the alts, but this is not the only one.

"A new type cryptocurrency which is proof-of-work based on searching for prime numbers."
http://primecoin.io/

Here is a pretty good list to look through. (~270+ cryptocurrencies)
https://coinmarketcap.com/all.html

[matt608]

Curecoin, long awaited, launched 2 days ago

https://bitcointalk.org/index.php?topic=603757.0

[Lethn]

Einsteinium is a new coin that's been released but already they've honoured 2 of their epochs and donated to cancer and diabetes research projects, I think it's a pretty clever idea, of course this relies on the developers' honesty but so far they've done really well with their plans, they don't do it through mining but they take a cut of what the miners earn and use that to fund the projects.

http://einsteinium.org/nominations/

http://einsteinium.org/the-fund/

To kick the fund off we created a cryptocurrency, Einsteinium,  it’s decentralised and open source which means no single bank or nation state is in control. Because no single entity control Einsteinium (in facet the network is peer-2-peer) it affords users a certain amount of anonymity, with near to zero transactions fee. For every Einsteinium that is created 2% is paid in to the Fund pool. This money, plus other donations is what drives the Einsteinium Fund.

[jarekduda]

Thanks - while scientific value of primecoin is questionable, the curecoin sounds great.
But if I properly understand, it is not that a folding itself is a part of the coin, but folders are just paid in coins - it this case you still need to waste a lot of resources on classical mining.

Cannot we make folding itself a coin?
What other valuable problems could be directly a part of a coin as a replacement for hashing?

[greenlion]

I think these kinds of ideas centering around having proof-of-work do some kind of useful work are horribly misguided and just a bad idea in general.

Two main reasons, one cryptographic and one economic --

1) Having a non-arbitrary proof-of-work undermines the security of the system by undermining the randomness of successful proofs.

2) Having a side-utility of proof-of-work outside of the functioning of the system itself introduces the risk of economic incentives that could interfere with the incentives internal to the system.

[jarekduda]

greenlion, could you explain your objections from the perspective of e.g. protein folding?:

There is a fixed generally known protein energy formula, given task is defined by a sequence of aminoacids of a difficult to fold protein - its "foldcoin" would contain a sequence of angles determining a configuration of this protein, which should be a local energy minimum.
Checking proof-of-work for given foldcoin would be inserting this configuration into the energy formula and checking if
- it is a local energy minimum (|gradient|<epsilon1, hessian is positively defined),
- it is essentially better than previously best energy (energy < previous_best - epsilon2),
where epsilon1, epsilon2 are some small constants initially fixed for given protein.
Given foldcoin would initially have probably 0 value, which could grow accordingly to general market rules - when people start focusing on given protein and appreciate its difficulty and significance (see http://fold.it/portal/ ).
After finding the global energy minimum, the amount of foldcoins for this protein would freeze - there would remain a limit resource market, like for pieces of art, which prices could then further evolve - for example accordingly to significance of this protein.
For example buying foldcoins for given protein would be a form of investment in working on it - would increase its price, motivating people to work on it, and finally they could be sold.

What is wrong with such picture?

[nextgencoin]

All these threads just attract people posting their garbage. We know what are the half decent ones cause they are in the top 10 of market cap. though that doesn't meant they are all good either.

[jarekduda]

nextgencoin, to mine these "decent" ones, beside hardware and manpower, there is wasted more the a billion dollar worth of electricity a year and it is growing.
We have currently a real growing problem with energy sources, especially the clean ones - if people have to spend precious limited resources this way, let it be more than just finding otherwise useless e.g. sequences with small hash, as it could be replaced with finding something socially/scientifically useful like protein folding - very difficult problems currently being the base of medicine development.

[iLiekPai]

There's Gridcoin which is essentially folding and earning something for it, look it up. 

[MessyCoin]

The electricity-expensive has functions in Bitcoin and its clones are performing valuable functions such as network security and transaction confirmation - they can't simply be replaced by any arbitrary non-related computation such as protein folding. Thus the best ways to achieve better outcomes are in making more economical variants (e.g. Proof Of Stake rather then Proof of Work) and also in funneling the proceeds of these currencies towards more lofty goals.

Another related experimental coin to mention is Emercoin. Emercoin does not have any built-in scientific value to the computations used in mining (AFAIK that's near impossible for a truly decentralized currency), but Emercoin is a Proof Of Stake coin and therefore low-energy when compared to Bitcoin. Emercoin also claims to have a scientific agenda, having been established to sponsor high tech projects such as robotics, autonomous intelligent electric vehicles and also scientific endeavors. For example, Emercoin developers have been rewarding Emercoin to participants in Folding@Home for a while now ( see https://bitcointalk.org/index.php?topic=472142 - I helped translate for them ).

The other examples mentioned are all interesting attempts at finding new opportunities for greater social and scientific value from the electricity spent.

I guess the OP is hoping that some new Proof of Work is possible that doubles as providing a focused social or scientific benefit - but I don't think that's possible. Primecoin is a good example of a best attempt.

[jarekduda]

Thanks iLiekPai,
unfortunately like curecoins it just pays folders with something like bitcoins - which mining still requires huge amount of solving otherwise useless problems ...

My question is if we could directly replace bitcoin mining like finding sequences having small hash, with solving some useful problem like finding protein configurations having small energy?

MessyCoin,
please explain why we cannot replace finding small hash sequences, with finding small energy protein configurations?
I mean local energy minims - there is a complex discrete structure of them.

[MessyCoin]

MessyCoin,
please explain why we cannot replace finding small hash sequences, with finding small energy protein configurations?

Well, I am not expert in cryptography, but basically: it would be like cooking the dinner at the same time as cleaning the oven. You can't achieve both at the same time, because they have separate objectives.

Lets take Bitcoin because it is the main example:

Bitcoin's hashing isn't useless. The hashing performs the purpose of verifying transactions and securing the network. Although the hash function can perhaps be improved and made more economical, the most efficient function for the purpose was sought. Economies may be found but we can not simply replace this with a non-related computation. What are the chances that the best function for network security of a decentralized currency will also happen to be the most efficient for a scientific benefit? I think we'll be much more likely to make a more efficient hashing function so that we have unspent electricity (and with this unspent electricity we can then go apply to the other thing in a separate fashion and seek the most efficient methods for it!). But trying to find a function that provides network security and verification AND scientific value is an unrealistic objective I think as it won't be the best choice of function for either job.

I believe in Bitcoin and others, the Proof Of Work (hashing) is designed specifically so that it is very hard to find a correct hash (crack the code), while being very easy for the rest of the network to verify the result. From what I understand, this is a vital quality of the Proof Of Work, and I'm guessing this is generally not a quality of calculations with more scientific value. But it may be found in some. If we could think of scientific computations that are difficult to find but easy to confirm, then maybe we could be on track to a scientifically beneficial Poof Of Work. I don't know enough about Folding@Home to talk about it but the SETI program (Search for Extra Terrestrial Intelligence) comes to mind as a similar example: It is difficult for the computers to find a "Little Green Man" in the data, but easy to confirm. So it's good so far. But then it has the problem that it is not a decentralized data source. It relies upon SETI to provide the data and therefore won't qualify as the basis for a decentralized trustless internet currency such as Bitcoin.

Perhaps I'm going on a bit here but trying to show that while it's certainly worth thinking of new ideas it's not so easy to find a good answer.

[David Latapie]

Thanks - while scientific value of primecoin is questionable, the curecoin sounds great.

Riecoin came to the same conclusion: finding "raw" primes has little scientific uses. Finding Riemann constellations has more.

it this case you still need to waste a lot of resources on classical mining.

With Riecoin, the mining itself is the calculation, so problem solved.

I don't own and never own a single Riecoin. I just find the idea interesting.

[jarekduda]

David, riecoin and primecoin sounds solving more useful problems than bitcoin, but still are quite artificial - could we use unquestionably useful e.g. protein folding here?

MessyCoin,
Ok, the bitcoin mining process has some cryptographic restrictions for the problem we solve.
However, I think it could be weakened to use a general complex problems like protein folding(?):
Like for bitcoins, there would be a P2P network for given currency (can handle multiple currencies) – each node knows the entire situation.
Every few minutes there is generated hash of chronologically ordered transactions and new coins in the last period (by some percentage of nodes, can be chosen in cyclic way to equally use all nodes) and broadcasted to all nodes. This hash should be identical from all sources – in this case it is acknowledged by all nodes and new period is started being considered. Otherwise, nodes in conflict compare transaction list to resolve the issue. Such p2p network does not generate coins, but is relatively cheap - can be maintained by some tiny transaction fees.
Mining would be finding a protein configuration corresponding to local energy minimum which is essentially better than (known) current best for this protein.
After finding such configuration, the node broadcasts to all nodes the configuration encrypted with his private key, and his public key allowing to verify the claim – then the global energy minimum is correspondingly lowered and this (encrypted configuration, public key) becomes new foldcoin for this protein – with the same market value as its other foldcoins.
 
What is wrong with such general model?

[lopalcar]

I will try to post something when I have some free time
This is a very interesting topic for me and I think many people are missunderstanding you.
In sumary, I thought like you some time ago, but with the evolution in cryptocurrencys I doubt this kind of project could become very usefull, we should build a system for cloud decentraliced computing but not necessary limited to be a PoW algorith which I think it's very difficult to code and very limited in terms of real utility.

[Noruka]

you guys are months behind on these alt coins.....

[jarekduda]

thanks lopalcar - this is exactly my point: to limit the cost of infrastructure (mainly PoW) as much as possible (e.g. introducing some minimal transaction fee), moving the computational cost nearly entirely to the mining process: of finding solutions for some scientifically valuable problem.

[MessyCoin]

What is wrong with such general model?

One thing that comes to mind is that for a cryptocurrency to work, the time between discovery of new solutions to the puzzle (AKA "block time") must come in predictably regular intervals so that confirmation of financial transactions occur in a reliable fashion. In Bitcoin, this is managed by the "difficulty", which is automatically adjusted depending on the amount of computational power present on the network in order to achieve a 10min average between confirmations. The difficulty of the puzzle will increase if there are more miners, and decrease when there are less, in order to try and maintain the 10min average. I do not know how this problem could be approached in terms of protein folding or some other scientific progress - possibly there is some mechanism that could be used, but potential pitfalls exist if there is no way to limit the time between a new discovery.

The source of the algorithmic work is also important. SETI takes its work data from a certain source and with Folding@Home the work is provided by Stanford University I think. This is great, but not suitable data for the basis of a financial system. Bitcoin's work comes from itself, and this provides an invaluable level of security. Perhaps there is some decentralized way to feed new data into the computations. Sadly I am not informed enough on the topic of protein folding to be aware of the possibilities and I will need to read more about protein folding before I can be more confident. Where does protein folding data come from?

Perhaps there is a way to control difficulty with protein folding, and a way to ensure that the data is coming from within the system. But those would be my two concerns.

A great number of similar posts have been made on this forum:

e.g.
https://bitcointalk.org/index.php?topic=6408.0
https://bitcointalk.org/index.php?topic=335.0
https://bitcointalk.org/index.php?topic=305182.0
And many more, all worth reading for ideas on this topic. But I think this is the best one so far

Against Bitcoin, there are many stories that discuss the negatives, such as this "Virtual Bitcoin Mining Is a Real-World Environmental Disaster"http://www.bloomberg.com/news/2013-04-12/virtual-bitcoin-mining-is-a-real-world-environmental-disaster.html
... and many more

I have recently become more interested in Proof Of Stake systems for cryptocurrencies, such as Peercoin and Nxt. With these there is apparently little or no need for mining after the initial distribution. So then we can use our electricity for something else. But distribution is important too and at least we all get a fair shot of finding a block in Bitcoin, if we buy enough hashing power.

I still believe Bitcoin uses less energy than traditional banking systems, for the same value. I came to decide that there probably is not an alternative to Bitcoin's "wastefulness" unless some genius will come forth with a new solution.

Bitcoin does all it is supposed to, and if the solution is a compromise in some important regard, then I don't think it is a true solution.

[sid7039]

Research Support Coin:
https://ripple.com/forum/viewtopic.php?f=3&t=6472