Nice to see you acknowledge the issue at last, and  Im really looking forward to finding out how you intend to pull that off, because I cant see way that wouldnt undermine your own profitability significantly as long as you maintain a sole supplier status, and I cant see a way at all if/when an asic competitor emerges. in the former case I can only see artificially limiting your own sales as a solution, in the latter case youd have to compensate your customers for your competitors sales.. not very likely..

yoyo implies 2 directions. I agree, that wont happen. But if you are implying you will hold prices steady even as difficulty skyrockets, I think potential buyers would like to hear that said explicitly, because I dont believe it for a second.


Yes, but that offer will hardly cost you anything. Mind you, I think its a good idea, and Ill credit you for it,  but its nothing like the tradeoff you will have to make once difficulty begins skyrocketing.

No, it will not be the same.  CPUs may have a low variable cost just like your asic, but the market price of CPUs was never determined by the bitcoin difficulty. Intel wasnt going to lower prices because bitcoin miners where no longer buying, nor could AMD charge 100x their production cost because bitcoin mining made it profitable. It was the opposite, the price of cpus (and later gpus and fpgas) was and still is set by factors unrelated to bitcoin, so that price/MH resulted in a (more or less predictable) bitcoin difficulty for a given exchange rate.

Like GPUs and FGPAs before them, given the inherent performance and performance/W advantages, ASIC prices will soon determine future difficulty. But unlike cpus, gpus or fgpas, future difficulty will also determine future ASIC pricing. That makes it 100% different.

Have to say, Im perplexed and mildly amused by the irony; you seem to have no problems keeping the biggest pool up with basically zero down time, extremely low stales etc, yet calculating the difference between two times on a webpage is a problem that takes months and countless attempts to fix? I would assume you store those date/times as unixtime in the database, and its literally as simple as subtracting 2 numbers and applying strftime to the result 

And how could you? And not only that, you would need to know how many will be sold in the  future. Thats what I said all along, without BFL telling you everything they wouldnt want you to know, you can not make an informed decision and buying their product at *any* price is purely a gamble; one that will almost certainly lose you money if BFL pursues maximum revenue. Its quite the opposite of BFL and miners interests  being perfectly aligned, they are diametrically opposed.


And what makes you think if they announce a 1 GH unit next month, that they will not up that to 5 or 50GH later on? I mean, they could just lower the price as difficulty increases, but if they are in it for the money, the smart thing to do is fuse off miners in those chips and start by selling truckloads of Asics that are roughly comparable to singles in performance/$ (but far superior in performance/W) and steadily increase the hashrate of these units,  instead of lowering the price. It amounts to the same thing, both lower performance/$ but its actually easier to pull off considering the non asic costs of every unit. Oh and then they could even sell you an "upgrade", because they care so much about their customers.

It may seem weird if you didnt think about it, but if the chips cost on the order of $10 to produce, it makes perfect sense to throttle them initially and slowly unleash their true potential. Then you make your estimates to make your "informed" purchase, and you will get burned terribly.

Thank god, someone who understands!
Well, the ones that dont and go for broke buying those asics at whatever price BFL decides upon, I guess will understand too - eventually.

Ill probably be dead by then. And I guess bitcoin will be too.
But the eventual lack of inflation is not what I see as its strength, I see it as a (long term) problem. Id have preferred a small but predictable inflation rate, but as it is, its not exactly useless, I see bitcoins biggest appeal as a transaction facilitator, to replace or complement paypal, western union etc, but not a store of wealth that will replace fiat currency. Or gold for that matter.

Anyway, thats another topic, just wanted to refute the idea that "supporters of bitcoin" could not possibly share Krugmans idea's.  Economically Im much closer to Krugman than Ron Paul.

Excellent, your order will be shipped in 4-6 weeks. Your order number is #101.  
Yes we already sold 100 units this month (80 to gigavps who will receive them tomorrow). We will auction off 100 more units on ebay every month, these chips cost $10 to produce, so why not.

Care to redo your math?

Like I said, i dont see a solution that wouldnt hurt BFL. They could make a promise to not sell more than x GH per y months, or some other form of guarantee thats linked to difficulty and that would allow miners to have some certainty of their ROI. But then what happens when a competitor appears, who is not bound to those limitations?

Another possible solution is not selling the hardware but leasing it at a variable price, ideally linked to difficulty; I guess that would need some DRM solution, not entirely unlike largecoin. But then BFL is essentially taking all the risks and just giving a part of the mining reward away, and BFL might as well mine themselves.


Which would also hurt BFLs potential profitability, probably a lot, and only delay the problem. It would be good for early adopters, but at some point even that low $/GH will begin to become too high as difficulty rises, and BFL will have to drop prices or lose sales.


If they have a full wafer maskset, there is no capacity limit. None. They could literally make millions of these chips. And it could cost them less than what only a few wafers would be worth at todays price/GH.


Agreed.

@SgtSpike,
Clearly Im not getting my point across. Lets try a different approach.

Assume I have 28nm fab in my mothers basement and I just finished my asic miner. I put it up for sale, long before BFL; 100 GH, 1W, delivery in 3-6 weeks. You can order today.

Will you order one if I asked $100K ? $10K? $5K? $1K?

And how will you know? You expect BFL to tell you how many chips they will sell in what period or at what (future) price? Without that information you can not make an informed decision.

No, you still dont get it. This is not typical at all. I cant honestly think of a market with similar dynamics.


Logical fallacy. You dont make a distinction between customers that already handed a check and have a miner thats yet to break even on its cost,  and potential new customers. The former ones want a low difficulty, and the only way to keep difficulty low is if those asics remain expensive/difficulty. The latter ones only want a good price/difficulty (which will increase difficulty). So you cant please both. It cant be both cheap and expensive for a given difficulty.  Now guess which type of customer BFL will care about most?

If your think thats such a problem today (I actually dont), just imagine what it will be like when difficulty starts doubling every 6 months, or faster.

Does that quote really make sense to you? If you truly think any ASIC providers' interest align completely with a miner's interests, and they will "do everything in their power to maximize miners revenue" rather than their own, then you are not the intelligent guy I took you for.

Well, I guess in a way they do have the same interest: their own bottom line.  But after the first asics will begin having an effect on difficulty, the asic provider and the miner will have diametrically opposed interests: one will want to keep selling as many as possible virtually free to produce chips to pay back his NRE, thereby constantly dropping prices to adjust for, and as  a result, increasing difficulty - potentially by orders of magnitude; while the miner would like to achieve an ROI on his investment, and wont be able to do so if difficulty goes up any faster than what he had planned for.

Ive asked the question before: what payback time would you look for to buy an ASIC? And what happens if BFL then sells ASICs twice as fast as you expected?


I got no problems with BFL. The issue is not BFL, its with asics in general; it will be the same with any company that would enter this market first and effectively obtains a dial that can turn up difficulty at will and must turn this dial to make a profit. I can think of some ways they could mitigate the miner risk, but none that wouldnt collide with their own bottom line.

Considering BFLs history, Im not expecting products in customer hands from them this year.  It took what,  almost 6? months  for the singles to really arrive, and those were "just" standard, off the shelve FPGAs. Product cycle for a custom asic is quite a different thing. So between now and then, a lot could happen.  Who knows, maybe even Largecoin will turn out be legit  

That remains to be seen.

And why is that?
You could argue the exact opposite, if you support bitcoin now (and not just in 30 or whatever years), clearly you are in favor of inflation and printing of money.

Wouldnt be so sure. Have you tried selling $1M worth of bitcoins lately?


Lets wait and see what the announcement says. I agree it seems likely they dont have a (full wafer) maskset ready now, but its certainly not unthinkable and I havent seen them take pre orders or announce availability yet, so lets not start calling them scammers again

Also, a maskset for an older processes its more affordable then you may think, and if even that was outside of their financial possibilities, its possible they went for a shared maskset and will produce the chips using a stepper. That increases the per chip cost significantly but could be a valid tradeoff for a first bitcoin asic.

But thats not how it works. Whoever invested in an ASIC will likely have spent well over $1M, and can produce all the GH and TH he wants  for negligible extra cost. Now you can mine with that for 6 or 12 months to try to break even (dont forget you push difficulty up and if you sell all your coins, you likely push BTC rate down) but if anyone else comes along with an asic during that year, you are screwed.

OTOH, selling the equipment to miners who do math like you do, will probably net them a very tidy ROI.

hmm. where? I only see a deposit link and it doesnt even let me deposit, let alone withdraw

I agree with the conclusion, not entirely with the reasoning. By selling hardware instead of mining yourself, you do get a lot of advantages:
- You offload the future difficulty/price risk to miners (but I guess that is more or less your point, and I agree too many miners will misjudge it and end therefore all may well end up losing money).
- You get far better cashflow position. Thats pretty important considering the investments involved.
- You dont compete with yourself as you develop newer iterations.
- You dont undermine the very currency your whole business is dependent on.

So no, I can see the sense of selling hardware even if mining yourself might be appear to be more profitable. In fact I think Vladimir is making a huge mistake by trying to develop his own hardware to mine with, rather than sell. But none of that changes the fact that if BFL prices their units at market price per GH to maximize their revenue, mining will become a waste of money instead of a somewhat risky investment.