[Archive] BFL trolling museum

[kano]

Of course if the bfl Asic became too much of a problem one could modify bitcoin protocol just enough so that the ASIC doesn't work or is severely hampered. They are very specific. A fpga can be reprogrammed. An ASIC would be nothing more than a coffee warmer then, literally.

Modification of the protocol requires that the majority of hashing power approve of the change. What percentage of hashing power is now tied to people with a vested interested in keeping BFL going, and what percentage will that be once this HW is in the wild?
If BFL sells 11 MiniRig SCs, even if no one else drops out BFL SC hardware will be 1/2 the network and there's no way the protocol will be changed to obsolete the ASICs.

... as I said before ... unless there is a reason why it needs to be/must be changed ...
https://bitcointalk.org/index.php?topic=87934.msg967195#msg967195

[RoloTonyBrownTown]

Wow, just finally finished getting 8 GPU's to play nice together, including (5) 5870's that have personality's of their own. Ranging from champ to the diva.
Multiple power supplies, some custom power wiring.

Hours of fun and frustration (I would say more fun, but still..), to get a little over 3 GH.

BFL is basically going to replace my system with a USB  "coffee warmer" @ 3.5 GH for $149.00

::This is where the slam head on desk smiley would go::::

I guess it is just a matter of when to dump the GPU's before everybody else floods the market and I cannot get rid of them.

(I will keep the champ however..452 MH, 1.08 volts, under 200 watts @ 42 degrees  )

Now THAT is the REAL question in this FUD thread.

Anyone gullible enough to think BFL comes out with ASIC in near future is LOL.

12 months AT LEAST if it is real and this is not a fake "stop other FPGA folks dead" announcement.

Oh fuck off Bulanula.  You and the rest of the brain deads said the exact same thing about the FPGAs when BFL announced them.  And you all looked like idiots in the end didn't you.

I'd have thought you lot would have learnt by now but I obviously underestimated your level of retardation.

[abeaulieu]

My best guess is that trade in policy will only be unit for unit within a product category.
Meaning you can only trade in 1 "old" single for 1 "new" single, 1 "old" minirig for 1 "new" minirig. And jalapenos will be sold without any trade ins. This way they will still be getting $$$ flow with every sale or trade in. Mark my words

Something interesting to consider.  I'll hold off purchasing another single until the trade-in in is clarified.

They announced that the trade-ins will be dollar-for-dollar (can't find the post atm). So 1 Bitforce Single =/= 1 Bitforce Single SC. In fact it's more of a 2 to 1 deal. (FPGA version = $600, ASIC version = $1300)

[cunicula]

Well, if this is true. The CEO of BFL (or the VCs who may be able to replace him) can own the network at trifling cost and there is jack shit anyone can do about it. You deluded idiots and your precious decentralization. What a fucking joke.

Well, yes, sort of. This has always been true. Didn't have to be BFL of course, but it was obvious for a long time that owning the network was within the grasp of anyone with a decent amount of money who could get ASICs designed and manufactured.

However, as several people have pointed out already, what you need to bear in mind is that even if this is true we don't know how many THash/s worth of chips BFL would need to sell in order to reach this kind of pricing. We can guess, though, and it's not likely to be a small amount. Once you factor in that and the resulting increase in difficulty those prices will probably look rather less attractive.

Read more carefully. The most profitable approach is to sell units en masse while quietly building up a larger cache of personal units. The fact that they profit from selling units to customers in no way implies that they cannot profit from turning these units into paperweights after the market is saturated. That is what I would do.

[Splirow]

WOW!

[rjk]

So I wonder how easily this could be reconfigured into a "fuck rainbow tables, crack passwords in realtime" type appliance. 1 bitcoin terahash per second is 2 regular SHA256 terahashes per second, which is a lot...

[MrTeal]

So I wonder how easily this could be reconfigured into a "fuck rainbow tables, crack passwords in realtime" type appliance. 1 bitcoin terahash per second is 2 regular SHA256 terahashes per second, which is a lot...

I'm going to guess not useful at all, unless the passwords happen to be hashed with SHA2.

[Inspector 2211]

So I wonder how easily this could be reconfigured into a "fuck rainbow tables, crack passwords in realtime" type appliance. 1 bitcoin terahash per second is 2 regular SHA256 terahashes per second, which is a lot...

I'm going to guess not useful at all, unless the passwords happen to be hashed with SHA2.

twice.

[MrTeal]

So I wonder how easily this could be reconfigured into a "fuck rainbow tables, crack passwords in realtime" type appliance. 1 bitcoin terahash per second is 2 regular SHA256 terahashes per second, which is a lot...

I'm going to guess not useful at all, unless the passwords happen to be hashed with SHA2.

twice.

If the PW was just hashed once, you could always just hash it again to get a double hashed version of the PW to feed the chip. You lose half the efficiency compared to if you weren't double hashing, but it would still be alot better than any other system.

[1l1l11ll1l]

Why are some people seemingly under the impression that ASICs are 'evil' or 'a problem'? ASICs are not the boogeyman. They will be used for mining, just like FPGAs are this year, just like GPUs were last year, and just like CPUs were 2 years ago.

They are a natural evolution of an efficient solution for the proof-of-work problem that the bitcoin protocol presents. The fact that someone has the cajones to dump serious capital into development of something like this is evidence that bitcoin is thriving, in fact, growing. This is a good thing for all of us.

Yes, the large-scale deployment of ASICs will make GPU (and to a lesser extent, FPGA) mining obsolete. It will drive difficulty up to levels never seen before. That is, mining with GPUs and FPGAs will no longer be profitable. So what? Just like the arrival of GPU mining en-mass heralded the demise of CPU miners. So what? We either chose to embrace ASICs as miners, or we take our profits and GTFO of the mining scene.

What may happen is that when ASICs are deployed en-mass, the 'average miner' will have a farm that has 20x the hashrate of what they currently have. So instead of a 5Ghps miner being the 'average', it will be a 100Ghps miner. If that miner has $4k worth of GPU/FPGAs now, he can choose to 'upgrade' and buy $4k worth of ASICs. The difficulty will be 20x what it is now, so 30M instead of 1.5M. Bitcoins will still be $6 (they will still be generated at the same rate so supply remains the same).

And this all means that mining profits for the previously mentioned 'average miner' with $4k worth of 100Ghps ASICs will remain the same as they are now. If he is making 3BTC a day with his $4k 5Ghps rig now, he will continue to make 3BTC a day with his $4k 100Ghps ASIC rig. Nothing changes.

This is 'evil' or 'a problem' how? It is only a problem for those who are currently invested in GPUs and FPGAs and who don't want to change. Change or die. That's how it's always been. It was the case with the CPU-to-GPU transition, and we are seeing the beginnings of it now with the GPU-to-FPGA transistion, and we have already been warned that an FPGA-to-ASIC transition is looming. No one is forcing us to continue mining, but many of us will.

Plain and simple... Well put.

[demkd]

Butterfly Labs (BF Labs Inc.), a market leader in microprocessor design...

really? wow 

[jgarzik]

Amusing.

[crazyates]

So I wonder how easily this could be reconfigured into a "fuck rainbow tables, crack passwords in realtime" type appliance. 1 bitcoin terahash per second is 2 regular SHA256 terahashes per second, which is a lot...

https://bitcointalk.org/index.php?topic=87476.msg968659#msg968659

[Unacceptable]

Of course if the bfl Asic became too much of a problem one could modify bitcoin protocol just enough so that the ASIC doesn't work or is severely hampered. They are very specific. A fpga can be reprogrammed. An ASIC would be nothing more than a coffee warmer then, literally.

Why are some people seemingly under the impression that ASICs are 'evil' or 'a problem'? ASICs are not the boogeyman. They will be used for mining, just like FPGAs are this year, just like GPUs were last year, and just like CPUs were 2 years ago.

They are a natural evolution of an efficient solution for the proof-of-work problem that the bitcoin protocol presents. The fact that someone has the cajones to dump serious capital into development of something like this is evidence that bitcoin is thriving, in fact, growing. This is a good thing for all of us.

Yes, the large-scale deployment of ASICs will make GPU (and to a lesser extent, FPGA) mining obsolete. It will drive difficulty up to levels never seen before. That is, mining with GPUs and FPGAs will no longer be profitable. So what? Just like the arrival of GPU mining en-mass heralded the demise of CPU miners. So what? We either chose to embrace ASICs as miners, or we take our profits and GTFO of the mining scene.

What may happen is that when ASICs are deployed en-mass, the 'average miner' will have a farm that has 20x the hashrate of what they currently have. So instead of a 5Ghps miner being the 'average', it will be a 100Ghps miner. If that miner has $4k worth of GPU/FPGAs now, he can choose to 'upgrade' and buy $4k worth of ASICs. The difficulty will be 20x what it is now, so 30M instead of 1.5M. Bitcoins will still be $6 (they will still be generated at the same rate so supply remains the same).

And this all means that mining profits for the previously mentioned 'average miner' with $4k worth of 100Ghps ASICs will remain the same as they are now. If he is making 3BTC a day with his $4k 5Ghps rig now, he will continue to make 3BTC a day with his $4k 100Ghps ASIC rig. Nothing changes.

This is 'evil' or 'a problem' how? It is only a problem for those who are currently invested in GPUs and FPGAs and who don't want to change. Change or die. That's how it's always been. It was the case with the CPU-to-GPU transition, and we are seeing the beginnings of it now with the GPU-to-FPGA transistion, and we have already been warned that an FPGA-to-ASIC transition is looming. No one is forcing us to continue mining, but many of us will.

+1,Very well said 

Just the resale value may not be there if Bitcoin tanks 

[Boognish]

I, for one, welcome our new BFL overlords.

[MrTeal]

For those gamblers here, a couple bets for you.
1. Will any BFL SC products be in customers hands by the end of the year?
http://betsofbitco.in/item?id=446

2. Will the BFL SC meet it specs (38GH/s Min, less than $1300)
http://betsofbitco.in/item?id=447

[nrd525]

If it takes 2 weeks for the difficulty to adjust, could BFL mine all the remaining bitcoins before it adjusts?  Or perhaps a smaller number - like 1 million?

If it is possible, with a low marginal cost of production it might also be profitable.  (If it didn't kill confidence).

[-ck]

If it takes 2 weeks for the difficulty to adjust, could BFL mine all the remaining bitcoins before it adjusts?  Or perhaps a smaller number - like 1 million?

If it is possible, with a low marginal cost of production it might also be profitable.  (If it didn't kill confidence).

No, the difficulty adjusts every 2016 blocks, regardless of whether it took 6 months or 6 minutes.

[the joint]

If it takes 2 weeks for the difficulty to adjust, could BFL mine all the remaining bitcoins before it adjusts?  Or perhaps a smaller number - like 1 million?

If it is possible, with a low marginal cost of production it might also be profitable.  (If it didn't kill confidence).

No, the difficulty adjusts every 2016 blocks, regardless of whether it took 6 months or 6 minutes.

I anticipate that the time it takes to reach the 12.5 BTC block reward will be a few months less than the time it has taken to approach the 25 BTC block reward.  It's going to be hilarious to see the difficulty and frequency of block generation increase when those 1 T/hash models start reaching consumers.  We might even have have a few periods where it takes as little as 3 days to see a difficulty change.

One thing is for sure...whoever the lucky few are to receive the 1 T/hash model before the rest of the pack are going to have one hell of a good month.

[Garr255]

If it takes 2 weeks for the difficulty to adjust, could BFL mine all the remaining bitcoins before it adjusts?  Or perhaps a smaller number - like 1 million?

If it is possible, with a low marginal cost of production it might also be profitable.  (If it didn't kill confidence).

No, the difficulty adjusts every 2016 blocks, regardless of whether it took 6 months or 6 minutes.

I anticipate that the time it takes to reach the 12.5 BTC block reward will be a few months less than the time it has taken to approach the 25 BTC block reward.  It's going to be hilarious to see the difficulty and frequency of block generation increase when those 1 T/hash models start reaching consumers.  We might even have have a few periods where it takes as little as 3 days to see a difficulty change.

One thing is for sure...whoever the lucky few are to receive the 1 T/hash model before the rest of the pack are going to have one hell of a good month.

If the first ASIC provider decides to mine for himself for a while, there will be very little benefit for the end user.