mike at powx.org [ARCHIVE] on Nostr: 📅 Original date posted:2021-05-18 📝 Original message:That’s a fair point ...
đź“… Original date posted:2021-05-18
📝 Original message:That’s a fair point about patents. However, note that we were careful about this. oPoW only uses SHA3 (can be replaced with SHA256 in principle as well) and low precision linear matrix multiplication. A whole industry is trying to accelerate 8-bit linear matrix mults for AI so there is already a massive incentive (and has been for decades).
See companies like Mythic, Groq, Tesla (FSD computer), google TPU and so on for electronic versions of this. Several of the optical ones are mentioned in the BIP (e.g. Lightmatter)
Sent from my iPhone
> On May 18, 2021, at 6:59 AM, ZmnSCPxj <ZmnSCPxj at protonmail.com> wrote:
>
> Good morning devrandom,
>
>> On Mon, May 17, 2021 at 11:47 PM ZmnSCPxj:
>>
>>> When considering any new proof-of-foo, it is best to consider all effects until you reach the base physics of the arrow of time, at which point you will realize it is ultimately just another proof-of-work anyway.
>>
>> Let's not simplify away economic considerations, such as externalities. The whole debate about the current PoW is about negative externalities related to energy production.
>>
>> Depending on the details, CAPEX (R&D, real-estate, construction, production) may have less externalities, and if that's the case, we should be interested in adopting a PoW that is intensive in these types of CAPEX.
>
> Then let us also not forget another important externality: possible optimizations of a new PoW algorithm that risk being put into some kind of exclusive patent.
>
> I think with high probability that SHA256d as used by Bitcoin will no longer have an optimization as large in effect as ASICBOOST in the future, simply because there is a huge incentive to find such optimizations and Bitcoin has been using SHA256d for 12 years already, and we have already found ASICBOOST (and while patented, as I understand it the patent owner has promised not to enforce the patent --- my understanding may be wrong).
>
> Any alternative PoW algorithm risks an ASICBOOST-like optimization that is currently unknown, but which will be discovered (and possibly patented by an owner that *will* enforce the patent, thus putting the entire ecosystem at direct conflict with legacy government structures) once there is a good incentive (i.e. use in Bitcoin) for it.
>
> Regards,
> ZmnSCPxj
📝 Original message:That’s a fair point about patents. However, note that we were careful about this. oPoW only uses SHA3 (can be replaced with SHA256 in principle as well) and low precision linear matrix multiplication. A whole industry is trying to accelerate 8-bit linear matrix mults for AI so there is already a massive incentive (and has been for decades).
See companies like Mythic, Groq, Tesla (FSD computer), google TPU and so on for electronic versions of this. Several of the optical ones are mentioned in the BIP (e.g. Lightmatter)
Sent from my iPhone
> On May 18, 2021, at 6:59 AM, ZmnSCPxj <ZmnSCPxj at protonmail.com> wrote:
>
> Good morning devrandom,
>
>> On Mon, May 17, 2021 at 11:47 PM ZmnSCPxj:
>>
>>> When considering any new proof-of-foo, it is best to consider all effects until you reach the base physics of the arrow of time, at which point you will realize it is ultimately just another proof-of-work anyway.
>>
>> Let's not simplify away economic considerations, such as externalities. The whole debate about the current PoW is about negative externalities related to energy production.
>>
>> Depending on the details, CAPEX (R&D, real-estate, construction, production) may have less externalities, and if that's the case, we should be interested in adopting a PoW that is intensive in these types of CAPEX.
>
> Then let us also not forget another important externality: possible optimizations of a new PoW algorithm that risk being put into some kind of exclusive patent.
>
> I think with high probability that SHA256d as used by Bitcoin will no longer have an optimization as large in effect as ASICBOOST in the future, simply because there is a huge incentive to find such optimizations and Bitcoin has been using SHA256d for 12 years already, and we have already found ASICBOOST (and while patented, as I understand it the patent owner has promised not to enforce the patent --- my understanding may be wrong).
>
> Any alternative PoW algorithm risks an ASICBOOST-like optimization that is currently unknown, but which will be discovered (and possibly patented by an owner that *will* enforce the patent, thus putting the entire ecosystem at direct conflict with legacy government structures) once there is a good incentive (i.e. use in Bitcoin) for it.
>
> Regards,
> ZmnSCPxj