41–50 of 59 posts
Earlier quoted context omitted.
I suspect they'd be designing for datacenters rather than consumers. That's essentially what they already do with bitcoin rigs too. It's a rented mining cloud.
> It's a rented mining cloud. That's not all it is. They do sell actual hardware as well, which you then run yourself.
Earlier quoted context omitted.
No. They don't actually control that much hash rate. Either the article is confusing the sales of their mining units to other miners or is outright wrong. If they really controlled that much of the hash rate, they would have activated emergent consensus to increase the max block size. They would have gotten around core and all their toxic censorship and bitcoin cash would never have needed to create a fork to get rea…
It's not wrong, look at the wording: > effectively providing 70% of all the processing power on the network. They are providing the processing power by selling mining chips. They are not operating those mining chips, so they don't control the processing power. They only provide it.
They can't generate infinite bitcoin units anyway but if they controlled 70% of the mining power they could conceivably execute double spends, although even this would be unlikely.
In any event the post above mine asked a yes or no question and the answer is no.
The article implies that providing mining cpus is much more of a problem than it is since they don't actually control that much mining hash power.
Good. Nvidia stuff 1080ti cost too much for a graduated student. AMD is working on their own stuff supposely later this quarter we'll get tensorflow and other support on their card. Google's TPU is not for consumer, only for cloud. If bitmain can sell ASIC for consumer with reasonable price and support tensorflow, pytorch, etc... then it'll be good for the deep learning community.
Earlier quoted context omitted.
Kinda. It's not a major plot point that's being given away It's more like "there's an Alien in Nostromo"
Errrr... I think it's a bigger give away than that. I came to the three body knowing almost nothing of what it was about... the information in the article would have removed much of the surprise.
Good. Nvidia stuff 1080ti cost too much for a graduated student. AMD is working on their own stuff supposely later this quarter we'll get tensorflow and other support on their card. Google's TPU is not for consumer, only for cloud. If bitmain can sell ASIC for consumer with reasonable price and support tensorflow, pytorch, etc... then it'll be good for the deep learning community.
Earlier quoted context omitted.
In this case, no. They produce ~70% of the hardware used by other miners. If they kept the hardware inhouse it would be another story all together. If history is an indication, selling the "shovels" to others during this digital gold rush is more lucrative for Bitmain than mining would be.
Also, even if you controlled 99.99% of the hashing power, you still couldn't "generate bitcoin until everything crashes." You can't give yourself an arbitrary sum.
Earlier quoted context omitted.
Also, even if you controlled 99.99% of the hashing power, you still couldn't "generate bitcoin until everything crashes." You can't give yourself an arbitrary sum.
with 51%, you just give yourself the entire sum of network rewards not an arbitrary sum.
The issue with ASIC is that when there is new AI algorithms that gets developed, they may not be even able to use those. General purpose hardware will be able to. You are stuck and need to upgrade.
Further speedups can be gained by going with lower precision ops (eg: float16). A chip specialized for deep learning could support f16 only, which would allow it to have more compute power using less area and less power.
So, in short, there are ways to make ASICs that are more efficient for deep learning while still giving the ASICs flexibility. ASIC doesn't mean there is no programmability.