Time for a bit of haphazard ranting.
I just finished (insofar as it can ever be finished)
the ASIC page at the Bitcoin wiki. Besides being a fair bit of work unto itself, some of the things found along the way and effectively over time since the last wordy post make me want to write a few things here; it's as good a place as any.
More people and/or companies should edit the wiki.There's a place for everything, and while the wiki is terrible for discussion, it's pretty good for basic findings that tend not to change very often. This StickMiner thread is effectively one of those things, and I'll be moving it over to the wiki in due time. Don't worry, I'll still be making update posts here as well and it'll be a while before the main content is moved over anyway. But it does mean that this list will no longer be at the sole whim of yours truly.
But there are many more threads that are effectively posts of information, rather than pure topics of discussion. Canaan Creative / Avalon project are, so far, the only manufacturer who have embraced the Bitcoin wiki as effectively the place to go for information about their own hardware - chips, miners, software, people to contact, etc.
In the course of making the ASIC page, I wound up creating pages for most of the manufacturers. I certainly invite them to flesh out their own pages - it's not wikipedia, original research is perfectly acceptable
Scams are everywhere.Holy bejeebus are scams everywhere in Bitcoin ASICs. Not so much any more, but certainly late 2013, early 2014, coinciding with rising Bitcoin exchange rates and people eager to throw money at everything, there was a great boom in scams. In the wiki I largely refer to them as being 'dishonest', but if you read between the lines on the references (yes, practically everything has source references), 'scam' should be the easy conclusion for most of them.
Now I've had to deal with scammy things before. Personally - heck, I might be in the process of being scammed right now. That's okay, I know what amount of money I can stand to lose - but also again in terms of editing the wiki. Specifically, two curious cases of manufacturers claiming to have certain hardware, (pre-)selling it, but offering no proof whatsoever of that hardware's existence.
In reaction to that, I ended up penning the
Why was my hardware removed Talk section, which outlines a few things on how a manufacturer can better prove themselves legit. From providing pictures to providing videos. However, I aslo concluded that "Ultimately, you may have to have a third party review one of your miners in order to build your credibility." Poring over pages and pages of information about ASICs really proved that point;
- You can post a picture of a miner, but if you Photoshop an existing miner, you're gonna have a bad day, ASICRigs
- You can post a picture of a miner, but if you badly Photoshop an external drive enclosure you're gonna have a bad day, Phoenix Technologies
- You can post a picture of a board, but if you badly Photoshop a motherboard you're gonna have a bad day, CedarTec
- You can post a picture of a chip, but if you badly Photoshop away the markings you're gonna have a bad day, labcoin
- You can post a picture of a mining screen, but if you badly fake the numbers you're gonna have a bad day, HashCoins
- You can post a video of a supposed mining rig mining away, but if you screw up the rest of it, nobody's going to believe that video.
But here's the thing - those were badly done (even though some took a while to be caught and some people had already lost money at that point) - what if they were well done, or there's simply no
good information either way?
Personally, I'd say hold off - let somebody else be the person risking the money on getting a potential product sooner than you, be happily proven wrong.
Thus is the case with, for example, HashCoins. There's no actual product being shown other than a (render of) a case), and it's known that they have been less than forthcoming with information in the past - be it with regard to a miner that was
supposedly their own board design (incidentally, an explanation of the difference was
for sale for $100,000 but bears a striking resemblance to reference HashFast boards, or
faking hashing details (yes, two for the price of one), or claiming custom chip and miner solutions for Scrypt miners when it's
just another Alcheminer clone.
HashRa isn't in a particularly better position, emulating Butterfly Labs' early days of '2 weeks'. Granted, Butterfly Labs eventually delivered. Too little, too late, but delivered.
And thus I haphazardly jump back a bit and say, what good is a review if the reviewer is bound by certain conditions for the review? In this particular case,
Dogie was asked to obscure the markings on the chips used in a miner. While the markings would be immaterial to the purposes of his guides, the fact that this was done under an NDA may lead one to wonder what else isn't entirely on the up-and-up (drag
the VAT thing into this thread and see your post deleted, thanks).
NDAs suck.. mostlyWhether it's the aforementioned NDA, or the
NDA that covered Dogie/Spondoolies-Tech, KnCMiner's Jupiter datasheet being behind an NDA, or just about any other NDA - they all suck. They're overly broad in scope and more often than not introduce a very lopsided relationship where the party who drafted the NDA holds almost all of the power.
Just as with preorders, the incentive is often to be privy to information that others will not learn or only learn much later. But unless that significantly benefits you personally, odds are it'll just haunt you further down the road. I've rarely signed a voluntary NDA myself for this very reason, but I fully acknowledge that other times there's a good incentive - or it's just not particularly voluntary (e.g. required as part of job)
PatentsJust as a quick aside - as it fits in with the topic of NDAs - remember when
HashFast released their Golden Nonce protocol specification as open source? Well maybe you shouldn't click that, because they neglected to mention that
HashFast has patent applications on parts of that thing (page 6). Who knows what else in Bitcoin hardware is patented - which is a shame, as a lot of Bitcoin technology development came from the early days of open source everything (often by necessity) and collaboration.
'Free' reviewsHaphazard jumping back a bit again and speaking of incentives, there is of course no shortage of reviewers; as long as they get the gear for free. A good example,
SFARDS tape-out announcement thread. There's about a dozen people in there already who will happily do a review if they're sent a sample. You know, for free.
Now, I get it - there's something to be said for both approaches. If you send a product for free, the reviewer may be biased to write a more positive review (not to mention that they can then go on and sell the product - decent profit to be made there on high value items even when offset against time+equipment; low value items not so much, to the point where a reviewer may opt not to perform a review, even if it's sent for free (alas, NF6, a guide was not meant to be)). If instead the reviewer has bought the product, they might still be biased because they don't want to admit that they bought a dude.
Of course there is a middle ground - sending a product but expecting it back, for example, is becoming more common for many hardware review sites - but I certainly lean more toward buy-and-review than review-for-free-and-sell-for-profit.
Home mining - Part 1Hey speaking of SFARDS, they already mentioned that they'd
try to target various levels of mining equipment, which may mean a home miner of sort.
Now, this caught my eye because:
- BitFury has exited the home miner market, including chip sales, at this time (yes, the same BitFury whose chips are found in most StickMiners and oodles of other miners), while they had their share of setbacks.
- Spondoolies-Tech would like to cater to home miners, but doesn't know if the market will exist (Miner Edge, using SP-Tech chips, squarely target datacenters, but perhaps there will be more information from SP-Tech at Inside Bitcoins NY 2015)
- ASICMiner is in organizational turmoil as friedcat appears to have disappeared. Which sucks for this thread a bit because Block Erupter USB - but I genuinely hope he's okay, personally, regardless of the reason for his absence.
- KnCMiner halted sales a long time ago (and whether one blames the user for not realizing consumer sales means having to know one's way around an oscilloscope or blames KnCMiner for mishandling consumer sales is a bit beside the point) and even took down their Forums (which I mention for a later haphazard jump). Of course they're also facing potential lawsuits (probably in part leading to the decision to no longer sell - though with their hash rate on the network, selling hardware only made sense if it was obsolete to their purposes anyway)
- CoinTerra went bankrupt (chapter 7)
- HashFast went bankrupt (forced chapter 7, modified to chapter 11)
- Black Arrow is under legal fire.
- Bitmine and Innosilicon are probably still trying to figure out their IP mess
Which, if you look at the ASICs page in the wiki, doesn't leave a whole lot of companies that make the chips, let alone that put miners together. Well, let me clarify - there's plenty of other companies in the past, but most of them are either no longer around or simply have no plans for new chips.
Sure
BitFury do have some "portable Bitcoin processor" planned, and
previously secretive 21e6-turned-secretive 21 Inc. have "a suite of consumer products that integrate with bitcoin’s core technology called the “blockchain.”" planned as well. But there's good chances that just means
Trezor clones like Black Arrow is selling or other such non-mining products. They don't even have to be physical products, with all the "Bitcoin 2.0" sidechain action going on.
Intermission: The internet never forgets? Hah!Speaking of companies that are no longer around - what about sites that are no longer around?
I mean, here you are, getting a bunch of good information on ASICs, sticking them in a temporary document for later perusal in detail, then later comes and *poof* the information is no longer there.
Ah, but "the internet never forgets", right? "The internet routes around censorship". Hogwash, I say. Unless you saved the data yourself, odds are it will disappear at some point in the future.
There are several ways that I have found that the information is effectively lost:
These last ones are particularly egregious, as
current robots.txt files can make archived pages at archive.org from years ago inaccessible.
With that in mind, almost everything in the ASICs page and every. single. link. in this rant post - and the post itself - is archived at
archive.today. archive.today 1. ignores robots.txt as it's not a spider and 2. captures everything on the page (if possible) at the moment of capture, instead of delaying resource capturing to a later date (and thus the many missing images in archive.org captures). It's also a relatively young service with no clear funding model - so I'm under no illusion that it will stick around - thus the links are also archived locally.
Home mining - Part 2But what of those that are still around and do have plans for new chips? Well, there's SFARDS as mentioned.
Bitmain should have a new chip out half-way through this year,
as should Canaan Creative (Avalon project).
From Bitman we've already had the Antminer U1/U2, and the Antminer U3, perhaps an Antminer U4 is in the line-up next to the Antminer S6 (presumable name).
From Canaan Creative we already know that their
new Avalon chip won't be suitable for a StickMiner, but as mentioned in that link they do have a small miner planned, at least.
But
Canaan Creative have also mentioned that their main business is large deals, as have
Spondoolies-Tech with regard to, effectively, cloud mining. Hey, at least they had a good
April Fool's joke 
Of course we can't, or at least, shouldn't blame them. Home mining is a dwindling market, and sales to home miners often don't make sense - a lot of potential profit instead goes into shipping, support, RMA, etc. All things that can be much better dealt with in bulk.
Not group buys, but bulk.Bulk almost commands centralization. We've already seen this in the huge mines in China, Iceland, northernly European countries (I know, I know... Iceland is technically part of Europe, too).
Without getting into whether that is a good or a bad thing (my opinion: a bad thing) on a deep level, let's instead tackle whether home mining was even meant to last.
The current system where every user is a network node is not the intended configuration for large scale. That would be like every Usenet user runs their own NNTP server. The design supports letting users just be users. The more burden it is to run a node, the fewer nodes there will be. Those few nodes will be big server farms. The rest will be client nodes that only do transactions and don't generate.
Well that settles that, then. Okay, kidding aside, ASICs probably expedited the process that was already foreseen on a level that was possibly not foreseen, but ultimately home mining was not meant to last. In fact, it can't last, as
the way Bitcoin mining works (its Proof of Work algorithm) always favors larger miners disproportionally due to economies of scale.
On the one hand that means pools. Pools aren't bad - even though some would try to convince you otherwise and even
attempt to make pools disappear by making them untrusted by default (and miners at pools just as untrusted) - if nothing else, they're the only way that home miners can still meaningfully (for some definitions of meaningful) participate in mining at all, next to playing the
solo Bitcoin mining lottery.
I'm certainly not saying that the PoW algorithm should be changed - though
it absolutely could be if it's deemed necesary, and
there are interesting alternative proof-of-X models - but as long as it is what it is, the realities of mining are what they are.
I
am saying that "1 person 1 vote" certainly doesn't apply in mining of just about any coin, least of all Bitcoin - and believing otherwise is, in my absolutely not so humble opinion, delusional.
In fact, the Bitcoin paper states this:
The proof-of-work also solves the problem of determining representation in majority decision making. If the majority were based on one-IP-address-one-vote, it could be subverted by anyone able to allocate many IPs. Proof-of-work is essentially one-CPU-one-vote
The subversion by allocating many IPs is instead a subversion to allocating many CPUs. Thus even a CPU-only coin is still fully at the whim of people who can afford to buy faster CPUs, or simply more CPUs. The only way to get a "1 person 1 vote" style setup is by issuing black box devices impervious to attack (down to decapping the chip yielding little more than a tangled mess that takes significant resources to untangle, at which point new devices could be issued) to registered individuals. But that presents problems in the 'impervious to attack' part -
The Trezor got hacked via power analysis, for example (by johoe, who has
previously uncovered nasties in Bitcoin transactions as well - props) - and of course in anonymity. Plus it still doesn't entirely eliminate the issue - as Alice could sell her device to Bob and now Bob has two votes. For that matter, Bob doesn't need Alice to sell it to him, he could just take it surreptitiously / forcefully, which would require further security measures to mitigate (and I'm not a fan of biometrics - that's a rant for a different forum).
Should the algorithm ever be changed, I do believe it should incorporate a mathematical curiosity (hard to analyze, easy to duplicate). Not including this in the beginning for simplicity (less things that could be broken) was a good idea, but it would be a missed opportunity not to do so at a later stage.
One other approach would appear to be in not giving out rewards at all, but then where's the incentive? Would Bitcoin have grown as quickly as it did, if there wasn't the potential for 'mad profits'? I doubt it, and PrimeCoin gives some credence to this notion, as
most of the difficult submissions to the project that this mines have come from PrimeCoin miners, rather than altruistic citizen scientists.
For now there's no reason to expect a PoW change anyway, so let's haphazardly jump back again.
Home minersThis is 'ers', the other was 'ing'. So even though a lot of companies are looking toward business-to-business type solutions, does that mean home mining is dead or dying?
Well, no. Miners - the completed hardware - are made by effectively two companies: the chip developers themselves, and systems integrators. So even if Bob Corp Int'l decides to only make their miners for datacenters, as long as Bob is willing to sell some chips to Alice, LLC and provide the necessary technical data, then Alice can put that chip on a board of her own design and sell that to whoever wants to buy one.
We have seen this in the past with BitFury, but also with Avalon, Spondoolies-Tech, Black Arrow, Innosilicon/Bitmine, etc. Sometimes this requires signing NDAs, and sometimes you might have to badger people at the chip company, but as long as you can get the chips you can conceivably make miners with them. This StickMiner thread is obvious evidence of this at generally the smaller integrators, but also consider companies like
TechnoBit (Avalon, Black Arrow, Spondoolies-Tech, ASICMiner, Innosilicon/Bitmine, Bitfury chip - Say what you will of their customer support, but if they can get the chip, they will at least attempt a miner.),
Pepper Mining (ButterFly Labs, HashFast chips),
DrillBit System (Avalon, BitFury chips - sadly disappeared),
GekkoScience (Bitmain chips).
If you build it they will come? If you provide the chips, they will build it.
Which brings me back to GekkoScience's thread -
GekkoScience BM1384 Project Development Discussion - which is, in my opinion, one of the more interesting threads on the Forum right now, certainly when compared to the threads currently seeing page after page of trolling and memes.
Not only are GekkoScience building new miners, integrating Bitmain's BM1384 chip (again, kudos to Bitmain for providing them), but you've got
valkir and vs3 (nanofury) popping in,
goodney (twin chip fury) getting ready to adjust his (canceled) BE200 StickMiner,
The WASP project (various, unfulfilled, miners) and
Philipma1957 (assembly and sales) offering resources, and a whole bunch more. If intron and c-scape joined in as well, it would be one veritable cluster of home mining people).
So while home mining may be diminishing, it's not dead yet - and neither is some of that early collaborative spirit.
Thus ends the rant, and on to happier things, like the StickMiners thread update 
