Topic: CPU friendly Altcoin in development - page 2. (Read 8218 times)

I think so aswell. A machine if not now, eventually will be able to do just about anything that we can come up with.

Well that's in the definition of algorithm "An algorithm is an effective method expressed as a finite list of well-defined instructions for calculating a function."

If you take something that it's hard to define a algorithm for ...
http://en.wikipedia.org/wiki/List_of_NP-complete_problems

Then humans may perform more effectively at the task than computers, particularly if learning and experience develops an intuition for the task.

I don't believe it's possible to write an algorithm that a machine can't do but a human can.  Is there any evidence to the contrary?

Okay serious suggestion, but only half formed idea..

Entropy, proof of work based somehow on production of entropy, thinking cellphones have built in hardware RNGs, then require some XORing with a software pseudorandom and hashing to add a bit of CPU make work...

but checking work may be difficult.

However, it could somehow improve encryption of transactions with having huge entropy pool to work with.

Thinking entropy sources are kinda slow and throwing more CPUs or GPUs into a system would not necessarily make it faster, would be a per system advantage.

Not sure of total resistance to custom hardware, but at the moment tending to think about the same cost for about the same speed, but power efficiency would be an advantage.

Anyway, for anyone thinking they want to get the world on crypto coins, something where cellphones might be equal footing with PCs would be good, where they've gone straight to cellphones, Africa etc.

But as I said, very half formed idea, needs massaging into something workable.

I consider it miners getting rewarded/paid for securing the network and processing transactions.

I think you guys are too concerned about fair distribution when what matters is free distribution.

With a fair distribution, everyone gets an equal share of the worlds coin. Even if you could do it, 99.99% immediately cash it in for something more familiar.

To bootstrap a cryptocoin economy, what matters is free distribution. That just means that those who want it can get it. With a genuine CPU currency, anyone with electricity, cpu and a network connection can trade some electric for coin.  

aerobicoin for webcam, hash encoded as dance and yoga positions.... followed by the copycat DDRcoin and wiifitcoin.

If you can figure out how to set up decentralized proof of work where human input is required, you will be a millionaire.  Everyone will switch to your coin.

Just 5 or 10 minutes human input when a block is found. For example, the block hash could be used as
RNG seed for something like 1 small level of a very complex roguelike game.
Keystrokes to solve it are recorded and saved in the block chain as additional PoW.
Hard for computers, incomprehensible for 99% of all people (mechanical turk resistant),
and takes only a millisecond or so to check validity.

But I guess it would be difficult to implement in a way to make savescumming+bruteforcing inefficient for bot(net)s and
to require some human attention without making grueling working hours the best strategie.

Mark I agree that bitcoin's distribution mechanism is incredibly terrible.  I mean 99.99% of the world's population does not own any bitcoins and the distribution is halfway done!  Unfortunately, no one has been able to think of a better decentralized solution, and Satoshi was a genius to even be able to think of the one we have now.

Giving coins to everyone with a forum account or email address, while open to manipulation, still gives a WAY better initial distribution than mining does.  Ripple does this.  Unfortunately this method sabotages the ability to use proof-of-work to secure the network since the miners won't have any incentive to mine on a new network with few transactions if they don't get a block reward, so proof-of-work must be abandoned in favor of a centralized solution.

Howabout reverse difficulty with reward doubling, time limited and proof of stake (So it's not last minute gazumped) gold at the end of the rainbow coin or something

Oh, didn't know that, that's quite an interesting experiment then.

I've speculated about a protocol that might use the Peer-2-Peer network itself as a place of 'work'.  Block solving is done by daisy chaining together a number of signatures, the resulting hash then determines (randomly) what public keys are eligible to sign again.  The chain of signatures just grows as it bounces from node too node until it legally can be signed by its original creator to complete a chain-link, this makes the total number of signatures random.  The obvious weakness it that private keys are free so a huge number can be created to churn through, so an additional PoS signature might be needed with the chain-link requiring a cumulative total.  Also a small nominal coin cost to keep an address active would also be a good idea as it may make the mass address attack unaffordable, and it would likely have a stabilizing effect on valuation too.

I've also begun wondering if a more distributed chain is also something we want.  Rather then one monolithic chain, ever account is its own chain with each transaction being it's own block that requires both parties to sign, thus each transaction is a point were two chains share a block.  The only way to double spend is thus to create two forks of your private chain with two different people and preventing them from learning of each other.  By doing this daisy-chain proof you force the disclosure of all transactions to the network at large and eliminate mining-in-secret which is the object of any double spend.

No need for a clone. Take a closer look at Infinitecoin. The name is a con to make people think it is inflationary . . but everything will be distributed within 6 months.

I thought about making my own clone, where all coins are given out after 1 year or so, calling it Finitecoin. 

The model I favor is rapid distribution followed by a 2% inflation. I'll have more about this soon.

I do not think it is practical, given the number of CPUs in existence and the projected numbers of things-that-can-do-a-CPU's-job going forward, to expect, aim for or hope for one person with one CPU, or even one person with one normal everyday everyone has one billion-trillion-CPU-equipped-handheld-device, to be able to earn any really useful fraction of the common weal / common wealth in any reasonable span of time by mere application of that device to well known familiar task(s) everyone knows full well can be done by such a device and has access to the program/app for doing.

There are many examples in history, even merely in the history of the internet.

Every fully automatic method of raking in money over the internet from home using your PC died a horrible death once someone made available a Windows binary GUI app for doing it.

Every single one, I am pretty sure?

Can anyone come up with one that didn't?

Even CPU-mining Bitcoin, which is one of the latest such things, died a horrible death once compiled GUIs for Windows became available to do it.

This whole thread is motivated by the same quest that motivated the spam ten million email address apps for Windows, the spam the newsgroups apps for Windows, the build a pyramid of marketers marketing for you apps for online marketers and so on and so on.

There are too many people, with too many PCs, dividing up pies that like ponzis and pyramids can only grow so large before the piece of pie each PC gets is too trivial to be worth doing.

Such things always (any counterexamples, anyone?) seem to get to a point where even just trying to keep up with trying all kinds of new twists on the same old ideas, join new launches of the same thing renamed/rebranded and so on and so on ends up making you less money per hour that you'd make simply solving captchas for pay or other "mechanical turk" type tasks.

I strongly suspect that even the time put into these discussions only seems potentially likely to pay off if you plan on being an "early adopter", one of the top of the pyramid / early entrant to the ponzi people who runs around making money off the suckers who buy into the scheme or buy the Windows app not realising the very fact that anyone with Windows can run it already means the thing the app does is obsolete. Selling the Windows app is pretty much the last resort, the last way to try to eke out a few last bucks from a wonderfully lucrative exploit that unfortunately too many people are exploiting for it to be worthwhile anymore for any one of them.

I realise I seem to be arguing that no one would sell shovels unless the shovels would dig them less gold using them directly or by means of employees than they make just selling the damn shovels. You sell the shovels once using them is no longer lucrative. Hmm. So I can see that I still need more insight.

-MarkM-

Markm, of course we could MASSIVELY improve upon bitcoin and get everyone to switch to a new coin if we designed one that was:
a) more efficient while still providing the same level of security
b) equally distributed to everyone on the planet in a decentralized way

The thing is, we don't know how to do either of those things yet.  If ripple is proven to work (unlikely), that could solve A.  B is virtually guaranteed to be solved within a few decades, but is impossible now.  Governments are reasonably well equipped to do an initial distribution to a population with minimal gaming of the giveaway mechanism, but it would never be decentralized.

By the time these things are solved, bitcoin's network effect is likely to be so large that even a more efficient and fair coin will not be able to displace it.

Btw, you cannot use events in the block chain to distribute an alt-coin because that would result in an even narrower initial distribution than bitcoin had.

Okay but what about in 25 years time, 100 years, 100,000 years, etc?

Do we have to keep minting forever? it seems like part of the appeal of Bitcoin has always been that after the relatively brief initial 140 years or whatever during which initial distribution takes place we do eventually reach a point where that is all, we have reached the cap, the only way to get some now is to go into business earning them (including going into the transaction-processing business which is just one of many possible financial-services businesses one could choose to get into) or just outright buy some.

Transaction fees presumably do ensure that there will always be some for sale, even if they are being sold only for the specific service of processing transactions; but can a currency be a success, even only as a store of value, if that is the only product or service they ever get used to buy? If not then there will be other products or services one can buy the coins with, so transaction processing surely should have no special desirability it should become merely yet another product or service people can choose to go into business as a provider of.

So really unless your model is that a cap is bad, minting forever is needed, then this whole get the coins minted thing should probably be done rapidly, so that as soon as possible the world can reach the presumable long term equilibrium stage in which minting is long over, part of prehistory, and we get to see the real thing really running not just yet another "get in early to get rich" pyramid / ponzi / tulip-mania / IPO type project?

It does not seem unreasonable to question the division of motives, as in how much if any of the motivation is actually the goal of the long term stable state versus how much is the desire to launch yet another "get rich in 150 years or less" scheme...

-MarkM-

Sure because it works.  POW has been proven to work.  POS is more efficient and is "close" to working but requires significant centralized control (checkpointing) something which may be improved upon in future versions.  You keep talking about Ripple as if it works. It doesn't or at least it hasn't been shown/proven to work in an open network where some nodes are malicious.   

If it works then great but why not cross that bridge when (never) you get to it.  My belief is that Ripple will never work in an open psuedo-anonymous network.  So anyone planning to go that route is planning to build a closed source, central authority coin.  The good new is it will be very efficient.