Bitcoin Forum

First Law: An ASIC manufacturer will only deliver an ASIC when it's sale price is less than the gross value of Bitcoins it can be expected to mine.

i.e ASIC manufacturers will do what is in their best economic interest.

Would anyone like to offer a reasonable and succinct refutation of this law? No prizes from tangled specuation about how "everyone will drop bitcoin" and
such witchcraft.

How can the manufacturer know the value of BTC the device will mine without knowing the future network difficulty and Bitcoin exchange rate?

The term "expected" accounts for this. The manufacturer will only deliver when the expected return is below the agreed sale price.

What's in it for the manufacturers to delay? You say it's in their best economic interest, but why/how? Unless they are mining on their own with the hardware (which there is no evidence that they are), their only best interest is to sell as many for as much as possible. Delaying sales doesn't do anything for them, in fact it reduces the number of sales they can make.

1) If one of the ASIC manufacture sold devices a month earlier than everyone else they have a huge competitive advantage over their competitors, and that advantage has significant value. What do they gain by giving it up and delaying?

2) Some ASICs are not for sale, but for lease. They will mine on their own hardware and thus get it at cost. What benefit does that company have at delaying their own production and profit?

I really don't understand what point you're trying to make.  An ASIC manufacturer would deliver an ASIC regardless of how many Bitcoins it can mine.  That's the product they sell, thus they want to deliver said product to maintain their reputation and continue selling more of them.

And if you're trying to say that ASIC manufacturers would mine on them instead of shipping them until they are no longer profitable, then you've completely botched your statement up.  Currently, as it reads, you are saying that ASIC manufacturers will only deliver ASICs when they are profitable to the person they are delivering it to.  Also, who in their right mind would buy an ASIC that isn't expected to profit?

The First Law does not specify who is doing the mining and does not refer to profitability. It remains correct.

Based on what?

Based upon the assumption that the manufacturer will always do what is in its best economic interest.

100% true. If doing something wasn't in their economic interest they would go out of business...

*Mind Blown*

I'm sorry, you are wrong. FAQ as I recall it:
Q: Will you mine with the devices we pre ordered with our money?
BFL: No, we are hardware guys, we really don't care.

Anyway Avalon will deliver, they have a small batch and they will run to get the huge number of orders if they come first.

As I pointed out HERE (https://bitcointalk.org/index.php?topic=134926.msg1439767#msg1439767) and HERE (https://bitcointalk.org/index.php?topic=134926.msg1439975#msg1439975) in the other thread he has been trolling, Bonker really has no idea what he's talking about. I really hope anyone who is new here and reading this thread to look past his FUD and see that while some of his principles are true, they've been twisted and manipulated to the point of not even being applicable anymore.

You seem to like to following me around the forum, besmirching my reputation. But you seem unable to provide any good arguements, only a confused tangle of illl-thought-out drivel.

I readily give credit to anyone that I feel makes a valid counter-arguement. If you're just annoyed that I don't give you any such credit, then I apologise. However,
I don't feel you contributed much of value in our past discussions.

Please read my first post in this thread and respond accordingly.

Likely because your reputation deserves besmirching.  Plenty of counter-arguments have been made, but you simply ignore them.

I frequent all of these threads in the Mining section and all of it's sub-forums. I'm sorry to burst your bubble of self-importance, but I'm not following you. You just happen to be posting the same flawed hypothesis everywhere. Spamming something doesn't make it true.

As for my lack of a "valid count-argument", it was you who stopped posting in the other thread when I pointed out that you really don't know the difference between a difficulty recalc and a reward half. I pointed out a flaw in your reasoning due to a mis-information, and rather than respond, you come here to criticize my contribution to this oh-so stimulation discussion. Why don't you try posting a valid argument to my post in the other thread, and then we'll talk?

It appears that the discussion is over (no refutation of counter-arguments or clarification of the central argument by the OP), and it is clear that the First Law of ASICs does not stand on any sound position.

Bonker, for future reference you can practice getting better at arguing a position by reading Paul Graham's essay:
http://www.paulgraham.com/disagree.html

Because of the risk involved, selling ASICs is much less volatile than building a server farm with the hope of recouping cost.

Also, b/c they don't have the capital to develop and build the asics on their own, they needed that preorder money in order to get the ball rolling.

After the preorders are shipped, you could make a logical argument that they would be better off keeping the asics.

*snip*

Heh.  You seem to be doing fine at that all on your own.

I'm going to mark today on my calendar.  I've seen your picture around the internet for years.  Never thought I'd actually meet you.

I see where you are trying to go with this statement, but I must say that it a very strange statement. Personally, I believe that:

"An ASIC manufacturer will deliver an ASIC when it's available and for a price equal to the average customer's expected bitcoin revenue for the number of months that the average customer would consider reasonable for paying off their equipment."

For example if an ASIC manufacturer has the technology in hand and their average customer expects to make 10 BTC/month and is willing to accept a 1 year payoff on the equipment purchase, then it would be reasonable for the ASIC manufacturer to sell the product to the public for 120 BTC.

With this logic, it is perfectly reasonable for the ASIC provider to sell the ASIC instead of mining with it themselves. Why bother mining yourself when you can capture most of the expected revenue upfront and let someone else do all the work and take all the risk?

This is reasonable, especially in the (still hypothetical) situation where NRE cost has been recovered, and further fab runs can be done cheaply. Bonker?

I reject that law for 2 reasons.  First, you spelled "its" as "it's" and secondly, everyone knows the first law of ASICs is don't talk about ASICs.

Also wtf is a gross value of bitcoins?  That sentence doesn't even make sense.  It doesn't have a gross value when you mine it.  It has a gross value when you sell it.  Also, you'd be more concerned about a net value after the cost of hardware, maintenance, electricity, the building it's in, and cooling costs.  Oh, and all manufacturers claim they won't mine with their hardware (even though 6000GH/s appeared and disappeared off the bitcoin charts over the course of a few weeks, lol).  Also, the devices have no useable life expectation since they're all first generation devices.  That means nobody knows how many they'll mine total.  Also, bitcoin is unstable and can fluctuate in price or crash horribly or be blocked by governments suddenly or have an exchange get hacked so mining 10,000 ASICs instead of selling them is a stupid long term business investment.  Selling ASICs to people and pocketing their cash immediately and permanently is a smart investment by comparison because that's basically a "sure thing."

The first law of asics is that an asic may not harm a human, or through inaction allow harm to befall a human.

That law is for general purpose machines, in the supposedly "ultimate" robot story they actually dropped those laws when they went to special purpose machine such as robot earthworms and robot fly-catching birds and such, as they simply are not necessary for such special-purpose devices. So they are more like laws for CPUs than for ASICs.

-MarkM-

I think OP is completely fuckin' right.

1. Design new efficient Mining Hardware
2. Take money for pre-orders
3. Build hardware using pre-order money
4. "burn in" or "test" new hardware while there's no competition and difficulty is steady
5. Delay delay delay...
6. Once competing ASICs start to ship, ship your promised hardware out
7. Sit on your virtual pile of coins and watch the difficulty skyrocket while taking more orders

Makes perfect business sense, and anyone saying "they wouldn't do that" is suffering from a terrible bout of wishful thinking. "That" is exactly what companies try to do, make maximum profit.

Would be disappointed if all the major competitors weren't in on it too.

Except some (dare I say most?) companies are actually ethical and don't lie to their customers.

You should read about BFL's history of shipping products and their CEO's history :P

Their history of shipping products wasn't unethical - it just wasn't well estimated/planned/executed.  Incompetence, if nothing else, but not unethical.

Regarding their not-CEO employee, people CAN turn over a new leaf and change.

Okay...at this very moment in time, their actual FAQ page live on their site right now states:
We cannot disclose the number of orders we currently have.  Shipping of new units is still anticipated to being late November or early December. We are unable to predict accurate wait times until shipping begins.

You were saying?

P.S. I still can't find that hilarious product FAQ where the question was basically "You delayed your FPGA products.  How should we believe these will come on time?"
And their answer was something like "it's unpredictable, who knows, supplies, ship time, blah blah blah, in other words we're lying about the shipping date [paraphrased]"

^^^^^^^^
I like this guy! One of the few here that understands how business works in the real world.

All the major manufacturers swore up and down that they wouldn't mine on the main net with hardware they personally owned.  They'd all use testing nets and synthetic tests and whatever.  That's obviously not true, it's just a question of how not true.  There's still no explaining why 4000GH/s appeared and disappeared from the network rather abruptly over the last few weeks though so I'd say someone somewhere probably fired up about that many worth of them and profited from it.

But you have to keep in mind that there is no true measure for Total network hashing power, its all extrapolated from the time it takes to calculate a block, which is a value that can swing anywhere from 1 minute to 1 hour (and even more in extreme cases), so variations like that are nothing out of the ordinary, even if total network hashing power was constant.

+1 A 15% variance in the overall hashrate is NOT anyone bringing ASICs on and then off again. It's just natural luck.

It was like 6 weeks or something.  That's one hell of a bad luck streak.  Cue the stat geeks coming in and stating the probability that everyone sucked at mining by a factor of 4000/25000 for 6 weeks or whatever :P  I bet it's one in a trillion.  I've seen individual blocks take 3 hours to solve sometimes.  That just means everyone is super unlucky that time.  But every 10 minutes for 6 weeks having a variance suggesting 4000GH/s that aren't really there?  That's impossible.

If I made a machine, that 'you' paid me to build, that can make me millions of dollars with the only effort needed is plugging in a machine and letting it do its thing, while claiming its not ready yet, I would be stupid not to.

And I think any company is stupid not to.
If you think that the ASIC companies wont turn the machines on, to put it blunt, you're an idiot.

Why would you sell a bank before you made money?
Why would you ship a product that you MADE profit on, but would make more profit in 2 weeks turning it on than you made shipping it?



Honestly a better approach would be refund everyone's money +10% interest, that's what I would do.

It would take them 7-8 months to make the same amount of money as everyone's preorders.  Obviously, everyone would cancel their orders by then, and they would have a net gain of zero vs conducting themselves as a legitimate company.  And that's only if Bitcoin price held (which it would likely not once people figured out what they were doing).

How many times does this have to be rehashed?

2^n ..where 'n' is the number of users who haven't been on the forum in 72 hours.

7-8 months to make as much as the pre-orders, but not 7-8 months to purchase the machine itself at their price.

Your assumption assumes the following:

• The manufacturer is actually able to map out all possible consequences of its own actions and outside events to determine what is in its best economic interest
• There is nothing else influencing the behaviour of the manufacturer

In short, you're assuming that the manufacturer functions like a superintelligent machine with no values or principles other than "optimise for maximum profits". In practice, this is not true, which is why almost every "law of X" turns out to be a "rough guideline of X" at best.

What's your point?