Bitcoin Forum

The question of where bitcoins get their value has been chewing at the back of my brain for a while now. Yes, I know speculation and dollars invested gives it value, as do the goods you can buy with them. However, as long as dollars remain a more convenient alternative, I think the general public will be slow to adopt bitcoins. This low adoption rate will likely keep the value of the bitcoin down.

If we can find a "killer app" for bitcoins, a commodity that can only be bought through bitcoins, then we will have effectively ensured demand for bitcoins will remain at least as high as demand for that commodity. I hear the US uses oil for this purpose, though I haven't researched that claim at all. This commodity would, of course, have to be unique to our own "nation;" that is, our bitcoin community.

The collective computing power of our network is phenomenal. There are a massive number of GPUs and GPU clusters grinding away to create new bitcoins every day. But what if it were more profitable to pool together our resources and rent them in exchange for bitcoins? We would be selling bitcoins to people only to have them give those coins back to us later, and at potentially a higher price than we might have earned by mining those bitcoins. After all, we are more efficient than any of the commercial clustering services out there and our margins could potentially be high. How many of you have tried looking for a clustering service only to find they were all priced too high to be profitable for you?

And, of course, if mining ends up being more profitable because of bitcoin deflation as a result, it would be a small matter to devote more of our nodes to mining again until the situation reverses again. At the very least it would establish a way for us to earn some good money with our mining boxes once all the bitcoins have been mined.

As a network we are faster, more efficient, and much more powerful than any other clustering service I've come across. Computing power is a commodity and we have a ton.

What do you all think? Am I on to anything?

What kind of uses do people have for giant GPU clusters?

Protein folding, prime number finding, DNA analysis, data mining. Lots of research applications, from what I understand. Anything involving lots of parallelizable computation would stand to benefit, really.

Who would pay for these research?

I want to echo what Kiba says here.  Protein folding et al. already have a huge volunteer following, and that takes the profitability away.  I'm mulling over other solutions so I'm not a total negative Nan... ::think, think, think::

Universities and biotech startups, I imagine. Not sure. I'd have to do some research to find out. The same people using the Amazon GPU cloud and Sabalcore's clusters, essentially. There are businesses out there already doing this, so it seems there's a market for it. But like I said, all the clusters-for-rent I've come across so far consist mostly of crappy NVIDIA-based boxes. Our network contains nodes of a much higher quality.

http://www.theregister.co.uk/2010/11/18/amazon_cloud_sha_password_hack/

Interesting. Yeah, that's always a possibility when you start renting out computing resources.

Here are the specs on what you get for $2.10 an hour with Amazon:

Cluster GPU Quadruple Extra Large Instance

22 GB of memory
33.5 EC2 Compute Units (2 x Intel Xeon X5570, quad-core “Nehalem” architecture)
2 x NVIDIA Tesla “Fermi” M2050 GPUs
1690 GB of instance storage
64-bit platform
I/O Performance: Very High (10 Gigabit Ethernet)
API name: cg1.4xlarge

This is their most expensive and powerful offering. I had been focusing on the GPU aspect of that particular product. I just noticed it also gets you 67 CPUs.

Here's the pricing schedule for Sabalcore:
http://ryepdx.com/SCI-pricing_schedule_2011c.pdf (http://ryepdx.com/SCI-pricing_schedule_2011c.pdf)

Also their user guide:
http://ryepdx.com/SabalcoreUserGuide.pdf (http://ryepdx.com/SabalcoreUserGuide.pdf)

+1

This is exactly what I've been thinking. Paying in bitcoin for online resources makes alot of sense.

Also, this can also help secure bitcoin against propaganda "whitchunts" by the powers that be. I remember a great presentation at a conference I went to once, about Cute Kitten Activism, this was long before the Facebook revolutions. The presenter described their experiences with a photo posting site and how they knew they were functional when people posted kitten photos, they knew they could scale when people posted porn, and they knew they were in over their heads when activists were using them to organize and to document atrocities in real time. What they found was that the kittens made the site too politically costly for the network to block. Governments have gotten a bit bolder since, but it hasn't worked out well for them.

I like your idea, but we also need a "Cute Kitten" app, something so obviously popular and non-threatening that people will be as eager to join the Bitcoin economy as they are to spend their baby formula money to buy power-ups in a Zynga app.

http://imagepaste.nullnetwork.net/img/1299802637feel_teh_powerz_small.jpg

It is unlikey this would exist without bitcoins.  (I'm wondering its half-life even with anonymous bitcoins).
  http://movies.witcoin.com/p/376/On-demand-movie-rip-and-upload

Don't forget finance industry. There is no limit to the computing resources they can utilize.

Get the finance industry renting time on mining boxes? That would be awesome. I think I would crap myself. One million internets to the person who pulls that off! :-D

+1

I'm not so much into the money laundering and illegal filesharing applications that I've seen going on for that very reason. Those seem like "Anti Cute Kitten" apps. If bitcoins become synonymous with shady dealings, the government will be much more likely to try taking down the network and, failing that, destroy the credibility of our currency with the public so that the coins become virtually useless. While it's nigh impossible to take down a good-sized p2p network, it's quite possible to forestall an idea's acceptance by the mainstream.

Ha. Maybe I should just make a kitten photo site and incorporate bitcoins somehow. :-D

Do it! :)

Maaaybe. First I want to finish the other bitcoin-related web app I'm working on. Then we'll see what I can do about the kittens...

Why not? We only need to market it together with an easily usable framework (like BOINC but with some p2p features maybe?). We have time though, until OpenCL catches up with at least the present state of CUDA. OTOH mining is becoming less profitable, so it should be done before people stops paying attention.  :)

Two words: Online poker.


This was enormous in the US until the Feds stopped people from being able to collect their winnings through credit cards.

Not a bad idea...

Sounds like something that could grow the bitcoin economy until we get enough physical density to start introducing our local supermarkets to it.

http://poker.bitcoinvegas.com/

Looking for investors.

Nice!  Way to be proactive!

I can invest development time! :-)

Killer "app", you say?  ;)

How about a business that delivers fresh, hot nachos to your door? People deliver pizza, so why not nachos? Create an order online, select toppings, check out using bitcoin, receive delivery, deliverer's tip jar address is on the box. Now, for a name... NachoMama or NachoBits or something like that.

The finance industry uses a lot of parallel calculations, which GPUs are suited for, in its programs.

you have a very good point, ryepdx:


right now (according to Bitcoin Watch) our network is hashing at about 141,891 TeraFLOP/s.  the main page for the BOINC Project says that their collective network (all projects) comes out to about 5,334.78 TeraFLOP/s.

a staggering difference...

my only question is this:  if we are going to rent out our TeraFLOPs to somebody in exchange for Bitcoin, where are they going to get the Bitcoin to pay us?  we won't be hashing up any new ones...

maybe a few of us could stay on the Bitcoin network, just to keep it going?  hmmm... difficulty drops to about 10,000... three or four blocks a day...

in the interest of solidarity - I VOLUNTEER!!!  ;D

I mentioned in some other threads this would be a way to back a CPU-based currency by something natural.   If you could do this right, it would be far more valuable than bitcoin anyway.  Consider the fact that if you were to try mining using, say, EC2 GPU instances, you would be paying .80 for every 0.01 you made.    In other words, attention people with GPUs: people are willing to pay 80x what you're getting by mining bitcoins to use your GPU.

The main problem with this is security: not security of GPU-owners by job submitters(that is easy, and most of work is already done by people working on WebCL).  Rather, the problem is it seems hard to come up with a scheme where GPU-owners can't cheat and do bad computation/not do computation at all.  This kind of security can only be done with a TPM or crypto-card proving remote attestation.