Which is the point. I just dont see a significant bottleneck anywhere.  Assembly cant be a (long term) bottleneck, if you consider the industry assembles roughly 1 million PC's per day and since this number is taking a nose dive, all these assembly houses are desperate for new business.

Chip production is not a bottleneck either. One single 28nm wafer holds enough chip candidates for somewhere around 50 - 100 TH. TSMC alone processes tens of thousands of such wafers per day, and they too are desperate for new production volume as they are operating well under 70% of their capacity.

The only thing really holding production back is probably a combination of strategic planning (its better to sell miners further in the future as their ROI is less predictable then) and risk adverseness of asic vendors, who are not ordering more than they can afford to lose, in case their chip doesnt work, bitcoin takes a nosedive, some competitor produces so much it destroys the value of their product,  or whatever.  


In 12 months growth will have tapered off. It wont even take that long. Not because I think production would be a problem, but simply because mining would no longer be economically feasible unless bitcon price goes x10 or more again. At this point however, I dont think anyone really loses sleep whether difficulty will be 500B or 5T next Christmas.  What really matters is the next 3 or 4 months.

You are right its not like bacteria and it wont continue doubling forever. But do consider this; during the previous difficulty interval, over 3PH was added in 10 days, during a time "no one" was shipping or deploying high performance miners. Well, clearly, someone must have, but it wasnt KnC, nor HF, CT, Bitmine, AM, BA or BFL. Unless it was some dark horse in this race, that was the result of smaller vendors like Bitmain or some bitfury resellers. If they can deliver 300TH per day, just how much do you think all the companies i just mentioned combined can ship/deploy once they are going in full production ?

FWIW, Josh from BFL is on record saying his company will easily be able to assemble and ship over 1000 Monarchs per day. Thats over 10PH per month. And thats only a single company. Next year we will have over 10 companies shipping latest generation miners. At least one of which is using a seriously large factory to assemble their units:
https://hashfast.com/pictures-of-the-ciara-assembly-floor/

576500 Square feet. Makes BFL look like a garage shop really.

I dont see whats highly improbable. We know for fact Silk Road is being investigated, and its almost certain some other bitcoin scheme's/scams are being investigated as well. Its unthinkable authorities are not at least trying to track associated bitcoin transactions. Its also very likely at least some SR dealers or users would have used BF/WE (as well as other exchanges and possibly security exchanges). For authorities to gain access to BF/WE's IP logs, email addresses and other data, they would have needed a 2703(d) order. I would go as far as saying it would be very surprising if this had NOT happened. Since a 2703 order is subject to a gag order, and a gag order actually would have made some sense, especially if a "trap and trace" was put in place, Ukyo would not be allowed to disclose this.

As for the inability to pay out, here is one of many possibilities; while BF/WE may have been targeted initially just to help catch SR users, it would quickly have become clear to authorities that the exchange itself (not too mention, all the asset issuers) is/are violating all kinds of laws, and once they collected the data they needed for the main investigation, it was shut down, possibly with servers/assets/wallets seized. Danny might be helping in providing a legal way to refund customers without violating money transmission laws, as his company might have the required licenses. The fact that most bitcoins are unavailable might be related to the distinction between BF and WE, with funds supposedly belonging to one or the other being seized. You might guess BF funds were still available as in the past security exchanges have been allowed to close to down properly, see BTCT.

Foot note: Im not risking a single satoshi in this, and Im not exactly afraid to call something a scam when I think it is one. Its just that in his case there is a much more likely explanation. One that in fact, Ive been predicting for quite some time.

In the above example, it was the PSU that caught fire. An Antec, and a fairly highend model at that;  Im sure they have the required certifications.

ANyway, Im just pointing out miners are not without risk.  Electrical fires are nasty, take whatever sensible precautions you can. At the very least invest a few euro's in a smoke detector.

Ask Twitter. Since gag orders and national security letters became legal:

http://www.ca4.uscourts.gov/opinions/Published/115151.p.pdf

IM not saying this is what happened here, but its plausible,  particularly if bitfunder/weexchange is being investigated as part of a larger criminal investigation. Think Silk Road, or any of the bitcoin asset scams. Gag orders do apply to this:

http://www.law.cornell.edu/uscode/text/18/2703

And its not exactly far fetched to see how an exchange would fall in that category. Would also explain why Ukyo is not able/allowed to destroy ID verification data etc.

Sushi, we arent all blind. Could you please refrain from constantly posting everything in obnoxious huge font and bright colors?
ty.

This is from the same guy:


Based off the damage to the power supply and nothing else besides the one 12v wire going to miner it seems like something shorted in the power supply and one or more wires overheated and failed, spraying hot molten plastic aka: wire insulation around. The molten insulation then caught some papers and a small cardboard box on fire and a wireless keyboard and my oakleys. 

http://forum.kncminer.com/forum/main-category/hardware/25328-fire-woke-up-to-sound-of-smoke-alarms-and-smell-of-smoke

Doesnt take a lot of imagination to think what could have happened if more flammable objects where nearby or if he hadnt been woken by the smoke alarm.You dont need 6 feet flames to burn your house down, a cigarette or burning cable insulator can do it to.

QFS

We all did. But what you pretend to fail to understand that the only explanation that makes <600W remotely plausible (as per bitmine spec) also implies the machine should be capable of ~2.4 TH in turbo mode. Either you believe that is the case and you think AMT is grossly understating the potential performance, or they are lying through their teeth.  Since you seem to reject the possibility of AMT lying, I'll put you on record as expecting your 1.2 TH to come with ~60 coincraft chips. Prepare to be disappointed.

Merry Christmas all

Ive pretty much solved all trouble with the printer, I tweaked the calibration and software settings,  its no longer leaking, Ive replaced the Z axis rod and printed a motor mount to replace the terrible stock 2 screw alu plate:



(yeah, I didnt have the right bolts and washers, so I used an excess amount of nuts and some hot glue for the microswitch )

Im quite pleased with how well it prints now. In preparation for new years eve, Ive been printing some (hopefully) flyable rockets, like this "firecracker" rocket which took 5 hours to print:





Its almost a pity its supposed to blow up

Electrical equipment is always a fire risk, but Bitcoin miners which are rushed to market with generally little to no QA and testing, and that consume high amounts of power are a bigger risk. Also 3000 watts implies very high amp draws that can heat up cables, (bad) connectors and potentially make them catch fire:
http://forum.kncminer.com/filedata/fetch?id=25647

Now Ive had a 500mW USB charger catch fire on me as well, so its not like this is exclusive to high current bitcoin miners, but  I would certainly rate them a higher risk and take appropriate precautions.

My mind follows perfectly simple logic and Bitmine's spec. But lets have you on record too: do you think AMT's 1.2 TH miner will achieve ~1.2 TH @600W or less?

Lets keep this simple, so maybe you can follow: you are saying these 1.2 TH machines have ~60 coincraft A1s?
If they have substantially less than 60 chips, there is no way for them to hit that power efficiency based on Bitmine's spec, and may I assume you would then agree with me they are deliberately deceiving their customers?

And if you do think they have 60 chips, in what world would they not provide cooling and power to enable the 1.5-2.4 TH their machine should be capable off, as per bitmine's spec?

You are missing the point. Bitmine could miss their targets, and I wouldnt necessarily call them liars if they do. But if Im going to resell bitmine chips or build gear around their chips, and I put numbers on it that are impossible with bitmine's specs, then Im lying. If you like car analogies so much, its like BMW advertising their new electric car will have a range of x miles, and a dealership taking preorders claiming it will do twice that (and claim to take delivery of the cars before even BMW says it will manufacture them).

And no, even 600W is not possible at the wall at nominal speeds if bitmine hit their specs. PSU's, VRM's and the rest do consume power, thats a fact, not an estimate.

IOW, its a lie. More so since the spec sheet states "1,200 GH/s nominal performance "
AT nominal speeds, there is no way they can hit the stated power consumption range. It will use >800W. That it can use less at unspecified speeds is about as relevant as what it consumes when its not hashing or turned off.  Power consumption scales fairly linearly with clockspeed (at any given voltage) so you can put any arbitrary low power estimate there as long as you dont specify the speed. They might as well have called it a 20W miner.


What other vendor sells miners stating only a low power consumption mode without the corresponding speed, while advertising the speed without the corresponding power consumption?


Which does not reflect well on them either. Keeping up the appearance its their own chip.


Its a lie if its deliberate. As for your GPU, you dont connect a GPU to an outlet at the wall. It doesnt have a PSU, so there is no way for the vendor to state those numbers. But at default settings, particularly when running typical workloads, it will use ~250W or less from the PSU, not 500+W. Otherwise that manufacturer is lying.


or two? Clenell was promised delivery first week of January. You wouldnt be implying they might be lying about delivery dates too ? .

Ok, those numbers are new. I got the answer I got when I asked. I guess Bitmine made an estimate since then.

However, first of all , those are number for the chips, not the miner. AMT sells miners, so their power consumption is going to be higher since VRMs, PSU's etc cant be 100% efficient. So there is no way in hell these machines could possibly be expected to consume 0.35 * 1200 = 420W. If you factor in real world inefficiencies, 500W at the wall would be a minor (miner?) miracle.  For AMT to state power consumption as 300W-.. is a lie. They might as well have put 3W or 0.3W there.

Secondly, assuming bitmine hit their specs, to get anywhere close to the stated efficiency would mean AMT is under clocking the chips to their minimum speed and have 60 A1 chips in there per miner. Which also implies at nominal speeds,  their rig should have no problems hashing at 1500 GH and in turbo mode it should achieve 2.4 TH. What are the odds they are selling machines that are perfectly capable (with adequate PSU) of hashing at 2.4TH, but only advertising 1.2 ? Considering they were marketing 40GH bitfury boards as 64GH, I rate those chances slim to non existent.

AMT, feel free to clarify how many A1's will be in that miner.

 

A few months ago, I asked Bitmine what speed their chip would run at in lowest power mode (0.35W/GH at the chip level) since they only advertise hashrate of their chip in nominal mode (0.6 W/GH) and turbo mode (1W/GH) :
http://bitmine.ch/?page_id=863

The answer was that while the chip should work at 0.65V (low power mode) he wasnt sure yet what speed it could achieve at the lowest voltage, and would wait until after receiving silicon and characterizing the chip before specifying. Which is an honest answer btw. Predicting extreme ends of voltage/clock relationship is a bit stabbing in the dark.

For AMT to base themselves on that number, one would have to conclude they do not know yet how many Coincraft A1s will be in their miner, because they wouldn't know how many it takes to achieve 1.2TH. One would also have to conclude the machine should be able to hash at considerably more than 1.2 TH in nominal mode (likely 30+%)  and  >2 TH in turbo mode, but for some reason they wouldnt exploit that.

Thats not what the changed topic says. What it does say is that AMT seems to find it more important what is being said about them than how they conduct their business.

AMT, if you dont like whats being said, all you have to do is change your conduct. If you look for instance in the Bitmain thread, it also starts with some healthy skepticism, but it took only a few pages and weeks for that skepticism to vanish completely and be replaced by mostly happy customers posting their experience. Delivering the goods you promised in a timely fashion while providing at least adequate customer service is all it takes really. Works miles better than paying shills.

I admire your faith. So far AMT haven proven to be misleading or lying about most things; shipping dates, performance claims (which allegedly they will fix by adding more boards, but no word how they will compensate the "many" customers already served), power consumption figures seem extremely improbable, the no VAT in the EU lie, they are paying off shills, and now they pretend that they will receive 28nm chips this week when even their supplier said a while ago it wont happen until next year.

Every new asic vendor is met with a lot of skepticism here, sometimes even too much, but AMT is doing a good job showing just why that is. Trust has to be earned, and so far AMT has earned none in my book.

What makes you say that? Scrolling back to june in this thread, I see people used 25% growth per difficulty adjustment as a guideline:
https://bitcointalk.org/index.php?topic=170332.msg2451810#msg2451810
If that had panned out, we would be around 700M difficulty today. Not a bad estimate, but the reality is nearly 2x higher, and thats despite all the delays from KnC's competitors.

FWIW, I stumbled upon this post, from you:
https://bitcointalk.org/index.php?topic=170332.msg2799559#msg2799559

You did say it was a best case scenario, maybe your worst case scenario was spot on, but in reality difficulty is still almost 20x higher.

The one thing few people did correctly predict was the astronomical rise in BTC price, leading to a distorted perception of profitability.  Difficulty growth so far has been pretty much exactly where most people expected it to be, not far from where most calculators that used exponential extrapolation predicted. Betting that this trend will break in the next few months, when over 10 vendors start shipping their next generation miners, not something I would do.