by Tyler Cowen
on November 12, 2013 at 10:31 am
in Economics, Web/Tech |
The words on the website say it all:
Monetize without ads
Let your visitors help you mine Bitcoins
Monetize without ads
Let your visitors help you mine Bitcoins
The pointer is from the excellent Ashok Rao.
Currencies are fine, new currencies are fine, but I really dislike a currency that uses pointless but mathematically hard hashes as “production.” Too bad they couldn’t come up with something that found new cancer drugs or whatever as you crunched “units.”
Solve cancer –> 3 bitcoins!
I take that point, and I certainly didn’t say “solve cancer” … but you know the whole bitcoin thing is just a conversion of electricity to currency, increasingly with specialized hardware.
It’s no stranger than digging up and refining a metal with very few practical uses and using that as currency either. Or Keynes’ idea of burying bottles of dollars and mining them.
You could say the same thing about gold mining. It uses real resources but doesn’t actually produce anything that helps satisfy human wants and needs (excepting of course the small amount that is used for industrial purposes).
Yeeaaaah, but it isn’t gold miners have the option, to use their “work” to do a secondary benefit. We have lots of distributed computation projects in science. So my complaint really is that “bitcoin” isn’t just rank at folding@home or whatever.
I dunno. I balk at the infomercials that talk gold has never been worth nothing — but I bet in the middle of an African famine it can be pretty worth nothing.
Still, the industrial applications aren’t nothing — medical supplies use it quite a bit and for pretty good ends.
And it’s pretty. That’s more than you can say for bitcoin.
Bitcoin is really nothing but tallying, right? There’s no there there? Not that any of our other forms of currency are much different, but why not work back instead of further down that road?
“Still, the industrial applications aren’t nothing — medical supplies use it quite a bit and for pretty good ends.”
This is an argument agains the monetization of gold, not for it. The monetary demand for gold drives up the price tremendously. If only industrial demand existed gold would easily be under $100/oz. For example Palladium, a precious metal where most of the demand is industrial, is about twenty times rarer than gold. Yet it’s cost is less than $800/oz.
It’s true the monetary demand for gold increases mining, but gold supply is highly inelastic to price (hence the reason why it’s a good monetary commodity). By far the industrial use of gold greatly suffers from gold’s monetary status.
Bitcoin mining isn’t useless. It’s used to verify all transactions and prevent double-spending.
Is it really 1:1? That might change my mind, but I thought that a lot of the mining results were just thrown away, or compared and thrown away.
Verifying bitcoin transactions is computationally complex, which is why the system provides an incentive to do so. Almost all of the work that goes into mining is spent on verify previous transactions and preventing owners of bitcoins from spending them twice.
Here’s the relevant paper, if you want to dig into the details: http://bitcoin.org/bitcoin.pdf
A lot of the mining results are thrown away when they’re compared. It’s a function of how fast new hashes are discovered, and occasionally there’s contention. But many economic processes involve throwing away a lot of the output (Intel throws away something like 25% of the chips it makes due to defects). Just because some process involves throw-aways doesn’t mean that it’s inefficient, since there’s often a tradeoff between lower error rates but slower production.
Same thing with Bitcoin block chains, which can be set to have faster discovery rates, which will lead to more conflicts. Or to have lower discovery rates, but will lead to lumpier payoffs for the miners. That poses more risk for small miners and disincentives mining, which would endanger the security of the block chain.
I think the reason for making it a completely useless problem was to ensure that miners were completely isolated from conflicting interests. For example, if they were finding molecule folding, a large molecule folding miner could move into the area and take over the mining.
You would also need the useful problem to be unpredictable ahead of time. If there were a list of, say, molecules to find to win a blockchain, someone could work ahead on the list for next week’s molecule. (Or if a consortium decided which molecules to fold, they could pick on they’ve already done work on.) Bitcoin depends on people not being able to work mining the block N+1 until block N is discovered, so no one gets a head start on anything.
Maybe I am being convinced. (Thus demonstrating my flexibility and convincibility.) Not that this means I loooove bitcoin.
This is borderline unethical–you’re taking money from the consumer in the form of higher electricity bills without their permission.
On top of that, the payoff is pretty crummy: based on their stated hash rates, if you have a thousand people sitting on your web page for an entire day, you’ll still make less than a penny. So if your site has a million page views per day, each ninety seconds long, your yearly profit for those several hundred million page views is… a bit under three dollars. I suspect one can do better with AdWords.
Can you do both?
People were talking about flash miners years ago, before GPUs totally took over, and even then they were considered impractical.
It’s not just taking the electricity, but taking it in a very inefficient way. A dollar of electricity used for mining in this fashion will not net you anywhere near a dollar. Only ASICs bring back more coin than the electricity they produced. So along with taking some of the user’s money, you are giving some free donations to their power company.
So if you get a million concurrent users you can make 0.0008 bitcoins per hour at current conditions!
oh wait, that’s like $0.28, how much does my server farm cost?
Unfortunately CPU mining is impractical and heading towards infeasible, the time for this business model has passed.
But… but… but… who’s going to harvest the crops in my FarmVille?
(Disclaimer: I’m a Bitcoin zealot who normally lurks, but felt compelled to comment today)
This idea has been proposed frequently, but part of the reason it doesn’t take off is that it’s cost-ineffective. Since Bitcoin mining relies on integer-crunching performance, the only way to honestly mine profitably is to buy an application-specific integrated circuit (ASIC) designed exclusively for Bitcoin and other compatible SHA256 cryptocurrencies. CPUs are designed for general-purpose floating point calculations, so the cost of energy outweighs the payoff. Matt (above) is right – this service is basically taking advantage of your visitors because anyone informed would object to such a waste. It would be more cost-effective to mine Litecoins (which are designed for CPU mining) and then sell them for Bitcoins.
Someone usually asks here – “why doesn’t Bitcoin use CPU mining like Litecoin does?” – it’s because CPU mining is heavily influenced by virus-infected botnets.
John Persona’s idea for a “useful” coin is also a frequent proposal. It can’t be done with cancer drugs, protein folding, SETI, etc because the goal must be defined by the p2p protocol, not a specific participant. If you rely on someone to issue specific real-world challenges like “fold this protein”, then there’s no telling if that person already got a head start on the solution. However, prime numbers ARE useful for humanity and can be defined entirely within the p2p protocol, so you might be interested in Primecoin. Primecoin uses CPU mining and has a lower hash rate so it’s less secure than Bitcoin, but it’s reasonably secure enough and >does< produce a public good.
Wow, seems just as informed as any of Tyler_Cowen’s other posts, especially on Bitcoin.
You can’t make significant money from passing off Bitcoin mining to clients. The typical CPU mines at something like 30 kilohashes/sec. According to this chart, it now takes bout 86400 gigahashes to make $0.20 USD, or 432 terahashes/USD.
So, to get just $20 this way (far to low to monetize a site), you would need about 9 x 10^15 hashes. Assuming assuming each user comes to your site for 30 minutes (giving you 54 gigahashes), you would need 170,000 users.
But thanks for your brilliant insights, Tyler_Cowen.
I assume you’re old enough to ignore the meaning of MMORPG.
Wow, seems just as informed as any of Tyler_Cowen’s other posts, especially on Bitcoin. …But thanks for your brilliant insights, Tyler_Cowen. –
Tyler, didn’t post any insights. He merely gave a link to another site in order to facilitate an on-line discussion. Which it did and which I, for one, found interesting.
On a practical note, if you don’t like the blog go somewhere else. I really don’t understand the mentality of people, who continuously waste their time posting about how ‘bad’ a site is. If it were really that bad, you wouldn’t be reading, let alone posting.
No one here is suggesting that you can make significant money by doing this.
Not with the site being down half the time and most of the requests failing. Ashok, your mavenry is too powerful.
No, but they did suggest you can at least monetize your site this way.
Comments on this entry are closed.
Previous post: “Nobody dislikes currency inflation more than strippers”
Next post: Assorted links
Email Tyler Cowen
Follow Tyler on Twitter
Email Alex Tabarrok
Follow Alex on Twitter
Subscribe in a reader
Follow Us on Twitter
Marginal Revolution on Twitter Counter.com
Get smart with the Thesis WordPress Theme from DIYthemes.