I never said that's what would happen (I said even if that happens hypothetically). I said if difficulty increases very fast as new ASICs come online, it shuts out the little guy. I don't think the Jalapeno will be that great since the delta between it and what I would imaging the pool rates will be much greater than it is today between our GPUs and the pool rates (in other words it may end up making way less $ in proportion than today's GPUs do despite its 3.5GH/sec rate).
That was the only conclusion I could draw from your statement that
If difficulty skyrockets 1000x, what you have long-term is 0.0001% of top guys holding 99.9999% of BTC.
and
Imagine the top 50 accounts holding 95% of BTC where the value of BTC has reached $10k, or something. OK, let's say the top 10 accounts have 1 million BTC each.
45% of BTC has already been mined. What you described is impossible unless your new ASIC miners buy up all the existing BTC on the open market, which anyone could do anyway miner or not.
Actually it has not. If you look at the 12-15 months graph, it's all over the place, as high as $33 per BTC. Contrary to your belief, there is no magical formula that determines a correlation between difficulty and BTC value. There are many other macro-economic factors involved. In fact, even the most knowledgeable people on the topic can't exactly pinpoint why Bitcoin value has increased in the last 1.5 months to $9. There are various explanations, with difficulty being just 1 of many factors, speculation being another, lack of derivatives to short the currency, etc.:
http://www.bitcoinmoney.com/post/26295113993/june-2012-results
Prior to it's hacking, Bitcoinica was an option to short the currency, so means of doing that has been available.
Obviously price isn't going to track hashing rate exactly, but the profitability of GPU miners has been a huge factor in determining price and difficulty.
After the artificial speculative bubble last summer, the price/hashrate relationship has been pretty stable. I'm not at all claiming that price is dependent on hashrate, I actually argued against it. I'm just giving a reason why it's been pretty stable the last 12 months.
Right now the Single is $599 and can do 832 Megahash/sec which is very close to an overclocked HD7970 that costs $400-450.
The current MiniRig costs $15k and can do 25 GH/sec (or 30x faster than the Single).
In the future:
- Jalapeno = 3.5 GH/sec - $149
- Single = 40 GH/sec (11x faster) - $1299
- MiniRig = 1000 GH/sec (285x faster) - $30k (<Everyone who owns the MiniRig can trade to this one and get $15k rebate).
Using a Single/GPU mining today isn't as bad. The performance delta between the next 2 levels is much larger than it is today (esp. the single at $1300 is as fast as the $15k MiniRig). Basically, if these numbers are true, unless you get the Single, you can pretty much forget about making anything worth talking about. That Jalapeno's 3.5GH/sec against the entire network will be much much worse than a 600-800 Mhash system today is when compared against the network. This is because the new added units will increase the difficulty more relative to Jalapeno, esp. since the top rigs will be 285x faster.
Unless you have a crystal ball, the bolded part is purely speculation. The top unit might be 285x faster than a Jalapeno, but that doesn't say anything about what the total hashrate will be. Keep in mind that the Minirig SC is a $30k unit; to increase the current 14Th/s network rate 10x would require $3.8M worth of Minirigs even if they can meet their 1TH/s spec. At 140TH/s, a single $150 Jalapeno would make up 0.0025% of the network. That's basically the same percentage as a single 6950 right now. The 6950 won't keep your coffee warm without sparks, either.
Yes, but if the entry level point to have any decent mining rig moves from $400 GPU to $1300 Single, that essentially would mean mining will become concentrated in the hands of the more affluent individuals. In other words, beyond that point the upkeep would be very expensive unless you get in with 2-3 of these Singles right away and constantly keep upgrading. It's too hard to say right now what will happen but unlike a GPU that can actually be used for games, these $1300 ASICs are not very useful outside of bitcoin hashing.
Are you willing to drop $1300 on one of those just to try it out?
I considered a Jalapeno, but decided against it given the sketchiness of the situation and BFL's track record. I do think it's a little unfortunate since it will keep new miners from entering the market (I started with a single 6870 that I bought to game with), but I don't think it's anything like the doom and gloom scenario you've listed. Again, in the next year and a half the number of BTC in circulation will only increase by 25%, and I would actually be surprised if BFL delivered ASICs by the end of October (or ever). ASIC will come eventually, but by that time half the BTC that will ever exist will already have been mined and the rate of new generation will be much slower, so it's value as a currency shouldn't change much unless no new users enter the market and attrition moves old players out.