Topic: Starting a new FPGA mining farm/contract! Cognitive Resurrected on[Havelock] - page 127. (Read 300686 times)

Which ones can you buy apart from BFL? Avalon is sold out, bASIC gone, ASICMINER mine only for their shareholders initially... Any others out there?

The way I see it, BFL FPGAs are the way to go if you can buy them below their trade-in value, and you plan to execute a trade-in. Does this make sense?

As for diversifying our "bets" on these ASIC companies (giving them funds now instead of later), I am growing hesitant of supporting this. It seems like you never know who's going to close up shop and either give you your money back or run away with it :/

A classic dilemma when a disruptive technology comes along. 
a) Wait too long to get on board, and you get left behind (see Blackberry and Nokia re smartphones)
or
b) jump in too soon, and you may get cut by the bleeding edge and die before the technology is ready for prime time. (see Myspace and Freindster)

Unless one thinks ASICs will never arrive, we have to assume FPGAs have a very short useful lifetime remaining.  From what I understand, the ASIC technology is nearly 80x as powerful as the FPGA (832MH  vs 60,000MH in the case of BFL projectons).  This means that the difficulty level will skyrocket as soon as ASICs are deployed, eventually causing a 80X increase, rendering FPGA returns 1/80 of what they are today.  Any new investment in FPGAs needs to pay-back very quickly, perhaps in just 30-90 days, before they are scrap.  Can they be bought very very cheaply now?

You currently have FPGAs making money, and you have a bet placed with BFL.  Seems the only way you might lose is if BFL fails (or is very late) and some other ASIC dominates the market before you can hold of some.  What are the chances of placing a bet on early ASIC hardware with someone else as a hedge?

Sounds like a good idea.  If BFL doesn't deliver in Feb like they said then growing via more FPGA is a good plan, IMO.  It would also be good to diversify FPGA platforms.

I believe that 50% is a good ratio for growth while Cognitive is still relatively small. In the future, when our total hashrate is much larger, we might want to consider paying more of our revenue as dividends


You know, that's definitely something to take into account. So right now we know that FPGA Singles are in the wild and do exist, so how about we buy these, and we hold bitcoins for upgrades until BFL has proven themselves to be true?

When BFL turns out to be a scam bigger than pirate it wont look so good 

does it have to be 50 percent? Seems high to me but I also see whether or not I reinvest my divs as a way of voting on whether I agree with the direction the company is taking.

Ah, I missed that, my bad.

Yes. The reason I specifically bring up PatrickHarnett is because Cognitive is/was invested in it.

Or amazingrando. Or that other guy. Or etc.

Correct. This money will be placed in redundantly backed up and sufficiently secure cold storage. I hope we all learned from the PatrickHarnett incident!

This is a good plan.

Just to make sure, the 50% of the revenue held will be held by Cognitive and not held by a person to earn interest, right?

I would love to grow Cognitive to the point where it will be totally self sustainable and capable of expanding significantly without issuing any additional securities. This, of course, would only be possible via a growth plan.

Currently, we're producing ~10 coins weekly, so we will be growing at the rate of ~5 coins per week. I just went on the BFL order page, and it currently costs BTC9.1123 for a Jalapeno, so we could have adequate funds for expansion after only two weeks of the growth plan! The next step up is the Little SC Single, which currently costs BTC38.2374, so we would have enough funds for this in roughly two months.

Personally, I'd like to see cognitive growing faster than that. This is why I'd love to see large, share issuance induced, expansion until we get to such a point.

And to more directly answer your question, I want Cognitive to be at this point within a year. For the past months our sole reason for this lack of growth activity has been the lack of ASICs in the wild. When they come, I see Cognitive flourishing.

Edit: Actually, a viable option for immediate expansion would be to purchase FPGA Singles, and order ASIC upgrades for them via BFL. I have talked to them and they have said they will allow me to expand my ASIC order that is currently first in the cue, so we would have first priority with these boards, which is largely advantageous especially at the beginning of ASIC mining

Thanks for the replies.

Deprived: I agree that that price for the preferred shares is low. It would definitely be raised before issuing these proposed shares.

Alright, so after digesting these replies along with an email sent from another shareholder, I've decided that (taking into account the current state of ASIC production) the preferred stock is not the way to go for the time being, and the forwards contract would work well after a bit of modification. Preferreds should represent a higher value than common stock, so my initial price for those (1btc for four) was significantly low. So, here's what I'm proposing for a forwards contract:

We can either issue forwards now, or wait until ASICs are available for purchase (not preorder). When we do, they should be offered in bulk, and priced above the current asks for Cognitive. This is to allow for large investments without driving the price up or leaving a bid wall and hoping your orders get filled. Because I agree with you that Cognitive is best as a pure mining asset, the futures coins will go straight into ASIC (pre)orders, and will not be subjected to any kind of arbitrage or speculation.

I think it's best to focus on increasing share value, which is why I support the growth plan along with expansion into the world of ASICS

This is really good analysis.  The preferred shares would only be beneficial to Cognitive it was done on a short term basis.  Also, the income needed to pay these preferred shares need to be  supported by the income of the mining and not dip into the savings fund that will go to purchase new mining rigs.

I suppose eventually Cognitive will get large enough that the retained mining income can support the growth on its own and not have to rely on additional investment that will dilute initial shareholders.  I would like to see a plan on where Garr sees Cognitive in 2-5 years.

How many weeks would it required with the 50% dividend to grow the mining company by buying ASIC from, for example, butterfly labs when they actually exist?

It's impossible to do - as one of the main factors in such a calculation is difficulty - which depends upon the total network hashing power.  That depends on which ASIC manufacturers deliver when, how many they ship, how well (and for how long) they work etc.  All of which are at present unknown.  Earnings after that will obviously drop - but how quickly depends on a bunch of factors which also aren't known (one of the biggest being just how soon the prices of ASICs get slashed - letting in a whole new bunch of miners/mining companies with lower capital costs).

It would be possible to give a range - but that range would have too big a spread to be particularly useful.

Some thoughts:

If the 50% reserve fund proposal is passed, it sounds like the common share dividend will be cut approximately in half.  While I think this may be a good investment in the future value of Cognitive, it may cause the common share price to drop substantially in the short term.  

The convertibility of the preferred shares might put a floor under the common stock, but how much depends on a few factors:
a) how many preferred shares are sold and available for conversion (are there enough to really affect the common share price?)
b) Investor expectations of the future dividends to be earned and paid on the common.  The common stock dividend depends on profitability, while the preferred are at a fixed rate.  The new investment might improve expectations for future profitability.

What funds are used to pay the preferred dividend?

• If preferred stock dividends are paid out of operations, that would further reduce the common dividend.
• If preferred dividends are paid out of capital before ASICs arrive, and is held in cash, then the preferred investors are just getting their money back in weekly installments with no other benefit during the waiting period.

I am not a fan of speculation with reserve funds.  Cognitive is most appealing as a pure mining investment.  

Before ASICs are available, the company would be paying 0.004 BTC/wk (20.8% annual rate) just to have cash on hand.  Perhaps the P shares should be issued only when actual hardware is available to be purchased?  You could ask investors to subscribe now, but not pay until later.

Has anyone done a calculation of estimated post-ASIC earnings per share with new investment (assuming upgrades are delivered) vs with additional hardware?

I don't think you should set a price for these shares.  You should let people bid on the shares and see what the market values this at, sort of like the PPT bonds did on GLBSE.  Everything else seems about the preferred shares seem great to me.


I am not so hot on the future shares right now.  I think having one type of  shares like the preferred or the futures it makes it even more difficult to set an intrinsic value to the COGNITIVE shares.

I'd be interested to see the math that has you value current shares at under .25 each - not disputing it, would just be interested to see it.  As the .P can be converted at will into standard shares they HAVE to be worth at least 4 times as much - putting an upper bound on the value of existing shares of .25 each.

However that only applies IF they're sold on open market.  I presume with the preferred stock you'd offer them to existing shareholders (as of a specified in advance date/time) with availability based on number of shares they currently hold - which would encourage them to buy and try to resell the .P whilst not devaluing existing shares quite so much.  Open market would then stabilise prices at .P being around 4* normal shares at some point above .25 for normals and 1 for .P (would necessarily have to be above this point as noone would buy at 1 with the intent of reselling for less).

Basically a good deal for anyone who would buy max allocation of preferred shares and not so good for those who wouldn't (although price of normal shares may not immediately fall below .25, the fact remains that when .P get converted there's only .25 extra cash in the company per new ordinary share - lowering asset value/share if it's above .25 and not taking it over .25 if it's already below there).

Not so convinced on the forwards - I'm generally not a big fan of companies moving into new areas of business (in this case currency trading).

You also need to fix or remove one of the following two clauses:

• Liquidation priority (In the event of liquidation, you will be reimbursed XX BTC per share.
• May be forcibly converted to Cognitive common stock, at the ratio of 1 preferred share to 4 common share

Together the liquidation priority has no value - as just before liquidation you could forcibly convert them, removing that entitlement.  In principle I think conversion shouldn't be forcible unless at a predefined date or under predetermined circumstances.  Once the ASICs arrive and get going all holders of the .P should WANT to convert as their stop-gap dividend would now be far less than from 4 ordinary shares.

For my money there's not really enough benefit in the .Ps over common stock to justify their existence.  It's about .003-.004 extra dividend per week compared to 4 ordinary - but only for the likely 6 weeks or so until hardware arrives - so a total discount on the price of about 1.5%.  That means purchasing would be almost entirely for the .25/share price and wouldn't even make sense if price of ordinaries fell below about .246 (and price WILL fall if this gets passed - as some people will have to sell their ordinaries to buy .Ps).

I'd have been inclined to price the .P higher and make the dividend significantly higher - so they were a good choice for those who want to invest but don't want to lose out if ASICs get delayed even further (0.4% dividend per week isn't that good if it's being paid from capital AND the capital is tied up in fiat so will lose value fast if BTC rises).  That then puts a bit more risk on ordinary stock - but also doesn't collapse the price so much (I'll have to have a look at the accounts some point and work out my own asset valuation).

To prepare funds for ASIC hardware (to be spent only when a manufacturer proves there is a product, and delivers it) I am looking into issuing preferred stock and/org forwards for Cognitive.

A few features of the preferred stock:

• Set dividend of BTC0.004 weekly per share
• Any-time Convertibility into 4 Cognitive common shares
• Liquidation priority (In the event of liquidation, you will be reimbursed XX BTC per share.
• May be forcibly converted to Cognitive common stock, at the ratio of 1 preferred share to 4 common share
• No voting rights

This would be listed on BTC-TC as COGNITIVE.P for 1btc each, which is an awesome deal.

Another thing that I believe would benefit Cognitive, is a forwards fund.

Forwards features:

• No dividend or voting rights
• Will be converted to 3 cognitive common shares upon arrival of hardware.
• To maximize our purchasing power, I will be able to trade these bitcoins on various exchanges, whether it be buying, selling, or arbitraging.

This would be listed as COGNITIVE.F on the exchange, and sold for 0.5btc each. My ability to move these funds around will likely be beneficial to Cognitive, as I do this with my own funds frequently as it is

Amount of each to be sold will be determined via motions which will be raised ~48 hours from this post. In the meantime, I would like to ask for shareholder input here.

Looking forward to the future of Cognitive,
Garrett