While ROI is calculated against capital investment, it does not assume the liquidation of capital. Now you could, of course, depreciate your capital, but that becomes rather insignificant with ROI in the mid triple digits.


While this is true, the question was weather we had reached parity NOW, and we have not. I don't see any way for the mining market to rationalize significantly expanding given the current legal and market uncertainty unless there is > 300% ROI on the day the hardware is ordered.

The problem is that we no longer have a competition of miners, but a competition of pools, lone miners will have no noticeable effect on transaction selection for the foreseeable future. The limited number of transactions pickers in the current mining eco-system, combined with the current  complete lack of scalability of transaction verification means that we will see a strong incentive towards fee laden transactions as the market matures. This creates the very real possibility of the mining pools arbitrarily setting rules that bypass the protocol to impose higher fees on the market.

At current we have a 1MB block size limit which is a kludge to prevent transaction spam, even if that limit is not lifted, and the average fee was 0.005BTC per transaction for full blocks a miner would still make 25BTC per block, which is enough to allow miner market equilibrium. This has the twin technical problems that there may not be enough transactions, or they can't be verified fast enough to fill the block. The first problem will fix itself if the market takes off, the second is far from intractable.

I bought 1Ghash that pulls 900W for $800.

go to http://bitcoinx.com/profit/index.php and plug in those numbers for a 12mo period. The calculator counts your hardware as a cost, but it's really more of a capital investment, and should not be counted against ROI.

Now 500% ROI is pretty sweet, but this is a risky investment. My miners would go off tomorrow if my countries government decided that BTC was a criminal enterprise, and I would be left sitting on unproductive capital. I would not add capital to this market with ROI under 300%, and I don't see how any business person could rationalize such an investment. Perhaps I would change my tune in 12-18mo, but not in the short run.

Not even close.

If current conditions were stable ( they are not, but we are talking about the current conditions) then a miner could expect a 500% ROI on reasonably cost efficient rigs, and not particularly cheap electricity.

The value of bitcoins is currently higher than their cost of production, and if nothing changed in the market it would stay that way for a decade, at which point the block reward would have dwindled to near the price in kWh.

There Fixed it.

The bitcoin infrastructure is not a product, if people come in and try to treat it as such, they are bound to be disappointed. If you want to build and market a compliant client, please by all means, if you can find a market niche meet it. As it stands bitcoind is FLOSS, in FLOSS land meritocracy matters, and you don't get to make demands if you don't contribute usable code, otherwise you are asking the devs to work for you for free, which is not a very compelling argument.

If you do this you dispense with the benefits of having a diversified market and all the network effects that carries.

If I said I wanted a tor trading site that only dealt in 18th century french coins, if another site existed that already allowed me to do that, would you think I was being silly?

My main objection is that your stated concern is impossible to rectify, if you run from all possible guilt by association you will never stop running.

Not to rain on the parade, but you realize that the arduino in this vid does essentially Zero processing right?

Even if you don't do the ECDSA (which even on bare AVR is probably not possible due to memory restrictions) you still need to securely connect to something that does, which means running at least SSL(also probably impossible), and having a net connection etc.

Android is a much better platform, and thanks to the economy of scale if the cell phone market, not prohibitively expensive in an embedded application. You can pick up an unlocked older model unit to root for < $200, and once you have a working prototype, you can have a cell enabled embedded bitcoin module built for the application.

The problem is that that excuse can be used for anything. The obvious solution is to just use silkroad.

Using a marketplace, even if other people do illegal things there, does not make you a criminal. I buy things downtown all the time, some people buy drugs there, if proximity to crime was a crime, you are already guilty, get over it.

For the hyper paranoid: If buying/selling gold becomes criminal, then at least you are doing business somewhere that takes security seriously.

Well, yeah, I guess I'll just wait for the DEA to disband, that sounds feasible.

I'm not making a moral argument, just a legal one.

Not to disrupt all the pro/anti-raw milk talk, but what about silkroad?

I mean seriously, nothing is stopping people from using silkroad just like e-bay to sell completely legal shit, but for BTC.

That would be silkroad's best defense, that it has a compelling legitimate function, it would also help bitcoin PR if the big bad evil was selling slightly used tickle-me-elmo along side the hash and meth.

This.

At the moment, in the interest of spam reduction the network has a policy of not infoming the rest of the network in the event of a double spend, I think this is a bad idea long term, we need a forward on blind request and a double spend warning message that will allow everybody to see bad transactions.

Meaning shut down the internet? or simply added enough hashing power to dominate the block-chain?

In case 1, it only takes a couple tor tunnels to stitch the whole thing back together.

In case 2 the Gov nodes are ignored as soon as they break the protocol rules and cause a fork, or their work is accepted as the longest block-chain if it is valid.


Each piece will attempt to talk to the others, none of the valid nodes will have an invalid block-chain, and the longest chain will win, and then all the transactions not in that chain will be added in subsequent blocks.


Without physically partitioning ALL data networks at the same division points, splitting at the network level will not work. If this happens bitcoin is the least of our worries.

splitting the block-chain without changing the rules will result in a quick merge, with the slight possibility of a double spend of the gov's own money, but they have to have 50% hash power to pull that trick off.

I got 3.

I'm on auto-notify for both cheap models of 5830 from newegg, didn't get an e-mail, but I refresh a search screen a couple of times a day to check on stock. I got lucky this time.

I would be interested to see their reasoning on that, my calculations put the sweet spot at 20%, not 50%. In my calculations adding hash power not only devalues your existing investment, but increases the difficulty within 2 weeks. Unless your hardware is both cheaper and more energy efficient than everybody else's, it makes very little sense to expand beyond 1/5 of the mining pool.

I submit that it is both. The "right" way would be to modify the pool management daemon to not think in terms of username/password, but, as discussed, this has not been done. The daemon has instead been fooled into passing the address as username, a great worse-is-better solution, that solves the problem, without breaking anything, a brilliant kludge that does not need to be fixed.

Since the out-of-band data pathway has already been exploited, why stop there? We have one obvious piece of data that could be passed as well by using the same trick.


I don't really see a problem here if the value is sensibly constrained, perhaps more than my original suggestion, say 2 digits from 0.1BTC to 9.9BTC. Use the last value that came with a share (so the trickster must actually pay you in shares for the privilege of changing your threshold), I would be interested in a scenario in which such a prank would matter to the recipient.

Eligius is already committed to the elegant kludge of Address as username, I'm suggesting leveraging this out-of-band data space for an additional purpose, which could be taken to arbitrary extremes, but I don't think my suggestion does that.

In addition I would argue that the signmessage solution is likely more problematic, as it requires processing resources to verify the threshold setting requests. This is an unnecessary liability, since threshold really does not need to be secure, as long as it is defined within reasonable limits. It opens up an obvious DDoS vulnerability for the pool server(s) on top of all the potential security problems with signmessage itself, as has already been discussed elsewhere.

Okay, how about changing how the username is parsed? If a standard format string is added, say -1000 (dash+4 digits, mBTC) to the end of the username, just use everything before the dash as an account address, and set the payout if the number is in range.

this

Since the pw is noise at Eligius, just reuse the space for pay threshold, say as an integer in mBTC.

Pros: better for the bitcoin network to not have a single large pool - this is a known protocol vulnerability, we should diversify the pools.
Pro/Con: size

All pools mathematically average towards the same rate of return. Size provides a level of stability, with DB winning almost half the block you will see LOTS of VERY small credits. On the other hand Eligius will pay you out larger shares, but only a few times a day, if Eligius has a string of bad luck it will matter more, since they expect to win fewer blocks a day the difference of only a few makes a big dent.

If you leave your miner running all week and check it once on the weekend, you will probably not notice a difference in payout.

If you are a console jockey who is constantly hitting refresh to see how your miners are doing, you will likely find a larger pool more edifying.

I duel mine ( two miners running per core ) on Eligius and BTCGuild. I have mined both on slushes pool, and on btcmine and was happy with them. When I decided to duel mine I picked one decent size and one smallish pool both of which had very low fees in order to avoid downtime, and help stem the DB tide.

If you look at http://bitcoin.sipa.be/speed-lin-2k.png it seems that hash rate has just set itself into a straight line.

Usually when we see a massive price jump vs USD we also see a jump in hash rate a short time after, and then a jump in difficulty to compensate.

Have we reached the wall in hash rate growth due to availability of cost effective hardware?