4 ms·

I don't quite understand. So they say: > We believe that the next wave of computing innovation will be driven by crypto. But then they also say that [0]: >

by qayxc 5y ago

I don't quite understand. So they say: > We believe that the next wave of computing innovation will be driven by crypto. But then they also say that [0]: > Crypto is purely a software movement and doesn’t depend on a hardware buildout, in contrast to, say, the internet, which required laying cables and building cell towers That doesn't make any sense, though. Crypto heavily depends on specialised hardware (ASICs, GPUs, SSDs, ...) and existing network infrastructure (have fun trying to establish crypto in underdeveloped countries). I get that products are just based on software and that infrastructure is provided by 3rd parties, but so far very little has come out of crypto in terms of "innovation in computing". Just a random pick of their investments, ARwave [1], reveals actual regression, not innovation: > The maintainers of the network are able to filter and screen the transaction in whichever manner they choose, checking against known illicit material, scanning the data with computer vision software, et cetera. So basically "store data forever, decentralized, no more 404s, no stealth edits, blah", but also "we decide what's actually going to be stored and reserve the right to censor your data however we see fit". Honestly, I don't see the innovation in that. Most of the other investments are either NFT market places or crypto-based trading platforms. Again, not really an innovation in computing. Maybe I'm just too illiterate when it comes to decoding marketing speak. [0] https://a16z.com/crypto/#vertical-landing-investment-thesis https://a16z.com/crypto/#vertical-landing-investment-thesis [1] https://www.arweave.org/technology#papers https://www.arweave.org/technology#papers [2] https://www.diem.com/en-us/ https://www.diem.com/en-us/

testific8 5y ago
I agree with most of what you've said, but some cryptocurrencies use proof-of-work systems (like RandomX) that truly are ASIC/FPGA/GPU-proof by design.
megameter 5y ago
The 90's dot com gold rush is the rhyme to the crypto gold rush. Lots of talk about ecommerce, revolutionary tech, etc. The 2000's saw the sobered outcomes: ads, ads, ads. All the apps were just conduits to adtech - very wide marketing funnels. Nearly all the actual services had been demonstrated before in a non-monetized, pre-Web context. A16Z is correct in saying that the dot com boom had a major build-out aspect. It needed broadband deployment and intensive consumer marketing. Cryptocurrency has so far demonstrated itself to be more like a bottom-feeder on hardware resources; hardware not being used for something obviously more valuable gets repurposed into mining of some kind, and energy negative uses have so far been accommodated through various kinds of theft of either the energy itself(in industrialized mining) or the compute time(in malware). In this light, the tendency towards specialized compute is not shown to be universal, but part of a set of interlocking arms races - securing systems and energy resources better, finding proofs of work with broader applicability across different kinds of hardware, and creating demand for hardware that is more efficient at those proofs that exist. Cryptocurrency in application is likely to be a one-note phenomenon in the same way that the internet boom became an ad boom - the same services with a little speculative cherry on top. It has a big inherent application simply by being decentralized and tradeable. It can be worse on other dimensions and still succeed by being really good at that one thing, financializing the world in greater degrees, and pushing the arms race forward.

by qayxc 5y ago

testific8 5y ago

I agree with most of what you've said, but some cryptocurrencies use proof-of-work systems (like RandomX) that truly are ASIC/FPGA/GPU-proof by design.

megameter 5y ago

The 90's dot com gold rush is the rhyme to the crypto gold rush. Lots of talk about ecommerce, revolutionary tech, etc. The 2000's saw the sobered outcomes: ads, ads, ads. All the apps were just conduits to adtech - very wide marketing funnels. Nearly all the actual services had been demonstrated before in a non-monetized, pre-Web context. A16Z is correct in saying that the dot com boom had a major build-out aspect. It needed broadband deployment and intensive consumer marketing. Cryptocurrency has so far demonstrated itself to be more like a bottom-feeder on hardware resources; hardware not being used for something obviously more valuable gets repurposed into mining of some kind, and energy negative uses have so far been accommodated through various kinds of theft of either the energy itself(in industrialized mining) or the compute time(in malware). In this light, the tendency towards specialized compute is not shown to be universal, but part of a set of interlocking arms races - securing systems and energy resources better, finding proofs of work with broader applicability across different kinds of hardware, and creating demand for hardware that is more efficient at those proofs that exist. Cryptocurrency in application is likely to be a one-note phenomenon in the same way that the internet boom became an ad boom - the same services with a little speculative cherry on top. It has a big inherent application simply by being decentralized and tradeable. It can be worse on other dimensions and still succeed by being really good at that one thing, financializing the world in greater degrees, and pushing the arms race forward.