Starting a new FPGA mining farm/contract! Cognitive Resurrected on[Havelock]

[investr]

If we can, we should add on to the Jan 30 cointerra order.

[1714212199]

1714212199

[theterabyte]

If we can, we should add on to the Jan 30 cointerra order.

I'd agree, but I'm pretty sure (their website is messed up for me now, probably under high traffic) I saw under "shop" that april orders were sold out and they were taking orders for may, implying that we can't just add to any order that would arrive before may.  And there is considerable uncertainty that 2TH/s in may for $6000 is a good deal at this time.

[investr]

So what you're saying is it would be a bad deal.

[lunarboy]

not all COG.F 2 sold??

It is sad we have to speculate on stuff like this.

Why is it not listed somewhere?

Remember initially there were 100 shares of Cog.F2 then Garr posted 500 additional shares, but then people were like "wtf, no motion", so then he paused trading and raised a motion... i forget what eventually happened with it, I think before anything else happened after that, BTCT announced closure and the resulting mess just obscured the details.  There were  suddenly *way* more important things to worry about than how many COG.F2 shares were sold (the effort was to raise a second order from cointerra at the time, if I remember correctly...)

EDIT:  btct.co is still up for legacy data, according to: https://btct.co/security/COG.F2 and https://btct.co/security/COG.F and https://btct.co/security/COGNITIVE there are 100 outstanding shares of COG.F2, 100 outstanding shares of COG.F1, and 10420 shares of COGNITIVE.  This means if these numbers are accurate, 92.9% of all shares have been claimed, and the total after COG.F1 and COG.F2 make will be 14420, which means all the numbers I ran in the past are actually under-estimates because I was assuming 26k shares I think.

Using the current diff of 1790mil, assuming a total hashrate of 28.75TH/s, if we got the machines today, we would project:

1.0/(1.790*1000*1000*1000*(2**32)/(28.75*1000*1000*1000*1000)/(60*60*24))*(25) => 8.077532385980616 BTC/day

This, at 14420 shares and 50% going to the reinvestment fund, means a dividend of 0.00196 per week, not including whatever the scrypt miners get us.  After the next difficulty increase, assuming 30%, the calculation yields 0.00151 BTC/week per share.  If we assume 30% network growth every 2 weeks, this means we would make 0.0085BTC/share over the first 2 months.  Given that there is considerable support at a current share price of 0.06 - 0.08, which would imply a return of 10% in just 2 months, I think COG is a very good value at the current prices and prices as high as 0.12 could be supported.  Once the hardware is received and confirmed hashing, and the majority of the risk is dispelled, I wouldn't be surprised to see 0.2 or 0.25 - but of course, this is not investment advice, just me dicking around with numbers.

Thanks for the analysis, this total number of 14420 shares needs to be confirmed by Garr and written into the Havelock contract.

Also we still need transparency and clarification on the dividends from between the BTC_TC and Havelock move. The fact that nothing has been done with this is laughable. Garrett you simply can't ignore this, it will not go away. For share price to reach its true potential full past accounts MUST be published. We are  working in an environment were trust is paramount, the longer this is left unresolved the more damaging it becomes.

[snowdropfore]

not all COG.F 2 sold??

It is sad we have to speculate on stuff like this.

Why is it not listed somewhere?

Remember initially there were 100 shares of Cog.F2 then Garr posted 500 additional shares, but then people were like "wtf, no motion", so then he paused trading and raised a motion... i forget what eventually happened with it, I think before anything else happened after that, BTCT announced closure and the resulting mess just obscured the details.  There were  suddenly *way* more important things to worry about than how many COG.F2 shares were sold (the effort was to raise a second order from cointerra at the time, if I remember correctly...)

EDIT:  btct.co is still up for legacy data, according to: https://btct.co/security/COG.F2 and https://btct.co/security/COG.F and https://btct.co/security/COGNITIVE there are 100 outstanding shares of COG.F2, 100 outstanding shares of COG.F1, and 10420 shares of COGNITIVE.  This means if these numbers are accurate, 92.9% of all shares have been claimed, and the total after COG.F1 and COG.F2 make will be 14420, which means all the numbers I ran in the past are actually under-estimates because I was assuming 26k shares I think.

Using the current diff of 1790mil, assuming a total hashrate of 28.75TH/s, if we got the machines today, we would project:

1.0/(1.790*1000*1000*1000*(2**32)/(28.75*1000*1000*1000*1000)/(60*60*24))*(25) => 8.077532385980616 BTC/day

This, at 14420 shares and 50% going to the reinvestment fund, means a dividend of 0.00196 per week, not including whatever the scrypt miners get us.  After the next difficulty increase, assuming 30%, the calculation yields 0.00151 BTC/week per share.  If we assume 30% network growth every 2 weeks, this means we would make 0.0085BTC/share over the first 2 months.  Given that there is considerable support at a current share price of 0.06 - 0.08, which would imply a return of 10% in just 2 months, I think COG is a very good value at the current prices and prices as high as 0.12 could be supported.  Once the hardware is received and confirmed hashing, and the majority of the risk is dispelled, I wouldn't be surprised to see 0.2 or 0.25 - but of course, this is not investment advice, just me dicking around with numbers.

thanks for your good job, you are so great!!!

[joele]

@ 30% Difficulty increase
28.75TH/s

Est Date      Difficulty           24 hrs BTC       Total BTC/Diff   Accumulated BTC

13 Jan 2014   1,789,546,951   8.07945404   88.87399447   88.87399447
24 Jan 2014   2,326,411,036   6.21496465   68.36461113   157.23860560
04 Feb 2014   3,024,334,347   4.78074204   52.58816241   209.82676800
15 Feb 2014   3,931,634,651   3.67749387   40.45243262   250.27920062
26 Feb 2014   5,111,125,047   2.82884144   31.11725586   281.39645648
09 Mar 2014   6,644,462,561   2.17603188   23.93635066   305.33280715
20 Mar 2014   8,637,801,329   1.67387068   18.41257743   323.74538458
31 Mar 2014   11,229,141,728   1.28759283   14.16352110   337.90890568
11 Apr 2014   14,597,884,246   0.99045602   10.89501623   348.80392191
22 Apr 2014   18,977,249,520   0.76188925   8.38078172   357.18470363
03 May 2014   24,670,424,376   0.58606865   6.44675517   363.63145880
14 May 2014   32,071,551,689   0.45082204   4.95904244   368.59050124
25 May 2014   41,693,017,195   0.34678618   3.81464803   372.40514926
05 Jun 2014   54,200,922,354   0.26675860   2.93434464   375.33949390
16 Jun 2014   70,461,199,060   0.20519893   2.25718818   377.59668208
27 Jun 2014   91,599,558,778   0.15784533   1.73629860   379.33298068
08 Jul 2014      119,079,426,411   0.12141948   1.33561431   380.66859499
19 Jul 2014      154,803,254,334   0.09339960   1.02739562   381.69599062
30 Jul 2014      201,244,230,634   0.07184585   0.79030432   382.48629494
10 Aug 2014   261,617,499,825   0.05526604   0.60792640   383.09422134
21 Aug 2014   340,102,749,772   0.04251234   0.46763570   383.56185704
01 Sep 2014   442,133,574,704   0.03270180   0.35971977   383.92157680
12 Sep 2014   574,773,647,115   0.02515523   0.27670751   384.19828432
23 Sep 2014   747,205,741,249   0.01935018   0.21285193   384.41113625
04 Oct 2014   971,367,463,624   0.01488475   0.16373226   384.57486850

[btcboston]

Also the manager could just claim he sold as many as he wanted to and collect dividends, or even sell counterfeit shares on the market diluting our profits.

this needs to be recorded somewhere. im honestly shocked this was not done.

And given this company is being run by a proven scammer who refuses to answer any direct questions, I'm certain that this will be happening.

[MilkyLep]

The problem with not knowing how many shares there are is we have no idea how to value the company.

Also the manager could just claim he sold as many as he wanted to and collect dividends, or even sell counterfeit shares on the market diluting our profits.

this needs to be recorded somewhere. im honestly shocked this was not done.

Seriously...after everything that has happened with this "security" the past 2-3 months you find the fact of unaccounted shares surprising?

I should've sold my holdings when I only got 6.65 BTC of my 8.5 BTC avalon chip order back from Garr.

[Tafelpoot]

Not all of those CogF2 600 shares were sold.

100 CogF shares
100 CogF2 shares

That means 4000 incoming Cognitive shares. Resulting in about 2 GH/s/share.

I'd like to add that you should take in account that the ASIC tech is top of the bill. I don't expect speed/power improvements that are as huge as the transition from GPU to ASIC for the coming 5 years. Which means that the hardware could be viable for a much longer time compared to first gen ASICs.

[superresistant]

I'd like to add that you should take in account that the ASIC tech is top of the bill. I don't expect speed/power improvements that are as huge as the transition from GPU to ASIC for the coming 5 years. Which means that the hardware could be viable for a much longer time compared to first gen ASICs.

Good point. First, the difficulty increase will not be linear. Then the diff increase could be much less than expected in the mid/long term.

[kakobrekla]

In this case I'm more shocked at all the buyers of the stack rather than Garr. However I do hope Garr lists it soon. This is information we need.

Hey, you were pumping it, remember?

[theterabyte]

@ 30% Difficulty increase
28.75TH/s

Est Date      Difficulty           24 hrs BTC       Total BTC/Diff   Accumulated BTC

13 Jan 2014   1,789,546,951   8.07945404   88.87399447   88.87399447
24 Jan 2014   2,326,411,036   6.21496465   68.36461113   157.23860560
04 Feb 2014   3,024,334,347   4.78074204   52.58816241   209.82676800
15 Feb 2014   3,931,634,651   3.67749387   40.45243262   250.27920062
26 Feb 2014   5,111,125,047   2.82884144   31.11725586   281.39645648
09 Mar 2014   6,644,462,561   2.17603188   23.93635066   305.33280715
20 Mar 2014   8,637,801,329   1.67387068   18.41257743   323.74538458
31 Mar 2014   11,229,141,728   1.28759283   14.16352110   337.90890568
11 Apr 2014   14,597,884,246   0.99045602   10.89501623   348.80392191
22 Apr 2014   18,977,249,520   0.76188925   8.38078172   357.18470363
03 May 2014   24,670,424,376   0.58606865   6.44675517   363.63145880
14 May 2014   32,071,551,689   0.45082204   4.95904244   368.59050124
25 May 2014   41,693,017,195   0.34678618   3.81464803   372.40514926
05 Jun 2014   54,200,922,354   0.26675860   2.93434464   375.33949390
16 Jun 2014   70,461,199,060   0.20519893   2.25718818   377.59668208
27 Jun 2014   91,599,558,778   0.15784533   1.73629860   379.33298068
08 Jul 2014      119,079,426,411   0.12141948   1.33561431   380.66859499
19 Jul 2014      154,803,254,334   0.09339960   1.02739562   381.69599062
30 Jul 2014      201,244,230,634   0.07184585   0.79030432   382.48629494
10 Aug 2014   261,617,499,825   0.05526604   0.60792640   383.09422134
21 Aug 2014   340,102,749,772   0.04251234   0.46763570   383.56185704
01 Sep 2014   442,133,574,704   0.03270180   0.35971977   383.92157680
12 Sep 2014   574,773,647,115   0.02515523   0.27670751   384.19828432
23 Sep 2014   747,205,741,249   0.01935018   0.21285193   384.41113625
04 Oct 2014   971,367,463,624   0.01488475   0.16373226   384.57486850

These numbers look pretty good, except the following two caveats:

Facts:
Note that these numbers are the total income - only 50% will be distributed as dividends.  That still bodes well however, because you can buy an awful lot of mining hardware for half of 380BTC.  Our order of 28TH/s with cointerra was 680 BTC (according to cognitivemining.com) and occurred in October, when BTC was under $200/BTC.  One has to imagine TH/s is cheaper due to rising diff and one knows factually that BTC has more purchasing power today, meaning COG can continue to grow.

Forward Looking Statements / Opinion:
Also, one would hope, the difficulty increase will not remain 30% per diff cycle for this long, so our miners will maintain some more of their value than this chart might anticipate [opinion].  We are unlikely to see miners at a smaller process than 28nm due to the extreme upfront costs of manufacturing such chips, and the increased complexity / expertise required.  As such, the existing manufacturers of 28nm hardware (hopefully this will grow to include AsicMiner and KNC as well as BFL and Cointerra).  At this point, with only minor improvements available and similar GH/W ratings, competition will drive $/GH down and cut margins to be razor thin.  That is the environment where COG can really be valuable, because it can organize large orders at bulk pricing that individuals cannot.  Individuals will want to own some shares of COG instead of buying a miner because it will be more efficient.  This will become especially true when miners are not losing value so fast (due to network growth leveling off).  This will mean the 50% reinvestment fund alone is enough to keep up with network growth and mining contracts (like COG.F1 and COG.F2) are not necessary (unless an especially good deal on an especially big order could be arranged, or an order from a new player is considered riskier and we want to isolate that risk from all share holders).

To Garrett:
If Garrett gets some help running cognitive, dedicates the proper time investments, and becomes better at communication, the future of COG is bright.  Seriously, Garrett - please do not fuck this up.  All of this pipedream about the future value of COG depends upon a typical miner feeling as comfortable investing in COG as they do buying hardware directly - given the recent history, that is probably NOT true.  People in this community are VERY jumpy about anyone throwing around the "S" word and for good reason - we have to move forward but we also have to acknowledge the past and show good faith.

EDIT: Tafelpoot also made the point before me about growth (*hopefully*) not remaining exponential forever.  Credit where credit is due.

[Garr255]

I've spoken with Ravi, and out CoinTerra Hardware is on track for a late-January delivery!

Wow, so they arrive in approximately two weeks? 

Yessir!

Finally we wil get the machine and start mining. And our f1  f2 will get the dividends too?

That's a great question - historically we have had delays which have been attributed to havelock not moving as fast as we like.  Are there any potential blockers to F1 and F2 converting, or is it self-service and you can just click a button, Garr?

I just have to send Havelock an email and they will handle the conversion promptly.

The current share totals for all assets are:

Cognitive: 10420 shares, 9577 of which have been claimed.
COG.F:     100 Contracts, 96 have been claimed
COG.F2:   100 Contracts, 95 have been claimed

Initially, COG.F2 was set up to sell 600 contracts, but only 100 sold before moving to Havelock, therefore the remaining 500 still "for sale" did not transfer.

Theterabyte, I believe you are correct that hiring some help will be in the best interest of Cognitive. Whether this is for PR purposes, programming infrastructure, or anything in between, I don't know yet. To date, whenever I have needed help with something that pertains to Cognitive, I have paid him or her out of pocket.

I am having some electrical work being done to our current warehouse, converting some 240v lines into 120v, in preparation for our new hardware. This will be completed on Tuesday of next week, the same day our Litecoin rigs will be complete and hashing. I will be posting pictures and potentially video of the completed setup.

With regard to the time necessary to continue operation of Cognitive, I know this has been an issue for the past six months. As some speculated even a year ago, university has been a distractor from Cognitive. To counter this, I am taking a schedule of classes that only occur two days per week, Monday and Wednesday.

With additional focus on Cognitive, and a bit of hired help I am determined we can restore Cognitive to the state it once was.

[damiano]

How often would you be upgraded and ordering more power?  The 28 TH is only good for a few months... March-May is when the dividends really start to fall off.

[Surf Capital Management]

Looks like a $100/share is right around the corner!

Disclosure: I own ~0.7% of the company.

[theterabyte]

Been thinking a lot recently about the network growth thing.  Here are some more (completely out-of-my-ass and opinion based) thoughts:

Imagine in 6 months or so, the difficulty is 54,200,922,354 (30% per period projection).  At that time, 2TH/s will yield about 0.25BTC per div period.  If BTC is still around $1000/BTC, that is $250 every 2 weeks.  That means a $6000 2TH/s machine will cost 12 months to break even (and even more if the difficulty continued growing at 30% per period).

So obviously 2TH/s machines won't sell at $6000 each.  What price would they sell at?  Well, 6 month break-even is the absolute highest most people would consider a "good deal".  If we assume negligible network growth (which is obviously not the case, so we are low-balling it here) 6 months means 12 div periods means $3000.  add in 10% network growth and you are probably looking at more like $2000.  There is also cost-of-power to consider.  With all the sunk costs taken care of, can cointerra sell 2TH/s machines at $2000 and still turn a profit?  Probably yes.  But they won't be making so much money "hand over fist" to make investments in even more sunk costs, so they will make a constant number, rather than an increasing number, of machines, in all likelihood.

When GPUs went crazy exponential, part of why that happened was manufacturers were able to produce a LOT of GPUs and computers.  The infrastructure was pre-existing.  Also, people presumed that their hardware would be able to run for a long time, and if it become uneconomical, could still be used to play games or sell used.  ASICs won't look like that.  They are single purpose, and every company that has produced ASICs has sold out and been limited not by number of orders, but by manufacturing capacity.  110nm ASICs from AsicMiner would already be useless if the value of BTC hadn't increased as it did (thus lowering the relative cost of power).  Even as it is now they are likely to become uneconomical in the near future.  The thing is, theoretically, 28nm should never become uneconomical unless it was to be replaced with an even more efficient design (which for reasons I stated previously I think is very unlikely any time soon).

(calculation:  5W = 330MH/s, at current difficulty you produce 9.271602390864709e-05 BTC per day which at $1000/BTC is ~$0.092 and 5W means 0.005 kW*h, which is 0.12 kW*h per day, which at $0.15 per kW*h costs $0.02 cents per day.  If my numbers are correct a USB block eruptor is currently only producing 4x more money than it consumes, and when BTC was 100$/BTC it would be producing half what it consumes, i.e. no longer economical to run).


(the block eruptor blade is 75W for 10.7GH/s.  At current diff, produces 0.0030 BTC per day, or ~$3.01.  75W = 0.075 kW*h => $0.27 per day cost to run.  If BTC was at $100 per BTC, then a block eruptor blade would net only 3 cents per day and be a single difficulty bump away from being uneconomical)

(the same calculation for cointerra's 2TH/s 1200W device, on the other hand: 0.562BTC/day, or $562 per day at $1000/BTC, 1200W => 1.2 kW*h => $4.32 per day to run.  Even if BTC was still $100/BTC, it'd be producing more than 10x the power it uses, and as things are today, it produces 100x the power it uses).

All of this is evidence that the time very well may come when the network grows linearly, 2-10% per diff cycle, or even less, simply because if the network did keep growing at the rate it is, even the most efficient miners would lose money and people would sell them or turn them off (except those who get free/cheap power), putting natural limits on the network size.  And, once that occurs, existing hardware will last longer and be more profitable to hold on to, while buying new hardware will have risk and little additional reward.  At that time, all those other things I said about why COG > personally buying hardware become true.

(Disclosure: using the post-COG.FX numbers, I own roughly 1.6% of COG)

[killerstorm]

@ 30% Difficulty increase
28.75TH/s

Est Date      Difficulty           24 hrs BTC       Total BTC/Diff   Accumulated BTC

13 Jan 2014   1,789,546,951   8.07945404   88.87399447   88.87399447
24 Jan 2014   2,326,411,036   6.21496465   68.36461113   157.23860560
04 Oct 2014   971,367,463,624   0.01488475   0.16373226   384.57486850

These numbers look pretty good, except the following two caveats:

Facts:
Note that these numbers are the total income - only 50% will be distributed as dividends.  That still bodes well however, because you can buy an awful lot of mining hardware for half of 380BTC.

Ehm. We already missed two weeks (13 Jan an 24 Jan). Which means accumulated BTC is less by 246 BTC, i.e. we only get only 134 BTC.

 Our order of 28TH/s with cointerra was 680 BTC

If this math is right and mining will start in the end of January, we get only 20% of bitcoins back. Which demonstrates that purchase of mining equipment is a hedge rather than investment. (I.e. buying mining devices and not keeping any bitcoins is just dumb.)

How often would you be upgraded and ordering more power?  The 28 TH is only good for a few months... March-May is when the dividends really start to fall off.

Maybe it is better to wait a bit, taking the above into account, and taking into account that multiple companies will be producing 28nm devices in large quantities: that will start a price war.

[theterabyte]

 Our order of 28TH/s with cointerra was 680 BTC

If this math is right and mining will start in the end of January, we get only 20% of bitcoins back. Which demonstrates that purchase of mining equipment is a hedge rather than investment. (I.e. buying mining devices and not keeping any bitcoins is just dumb.)

That's only true because bitcoin went up another 4x in value though.  We really paid 680BTC * ~200$/btc => $136,000.  If we make 134 BTC today, that's pretty close to an equivalent return.  If you want to invest in BTC, just buy BTC - you don't need cognitive for that - if you want to own mining equipment however, COG is a great way to do that.  If BTC goes up in value, BTC might have been a better investment than mining equipment - but that doesn't automatically make it a BAD investment.  Maybe bitcoin stays around $1000 for the next year so a year from now a 6-month-breakeven miner would beat just buying bitcoins - nobody knows the future.  A good investor adds exposure to several investments which are hedged.

How often would you be upgraded and ordering more power?  The 28 TH is only good for a few months... March-May is when the dividends really start to fall off.

Maybe it is better to wait a bit, taking the above into account, and taking into account that multiple companies will be producing 28nm devices in large quantities: that will start a price war.

That is probably true - I think that's the best explanation for our investment in scrypt miners - there wasn't a more logical bitcoin asic to buy at that time.  COG should probably hold on to their BTC for at least a few months now to see which way the price wars go, unless a particularly good deal comes along.

[btcboston]

With regard to the time necessary to continue operation of Cognitive, I know this has been an issue for the past six months. As some speculated even a year ago, university has been a distractor from Cognitive. To counter this, I am taking a schedule of classes that only occur two days per week, Monday and Wednesday.

With additional focus on Cognitive, and a bit of hired help I am determined we can restore Cognitive to the state it once was.

Hi Garrett,
Just wanted to check in and see how things were coming along with getting an explanation for the (lack of) dividends back in October/November.

Thanks in advance for taking the time to respond.

btcboston

[kakobrekla]

Let us just look forward I suppose, we have bigger fish to fry

Adequate.