Isn't it possible to make changes to the infrastructure of bitcoin or eth? I know the decentralisation is pretty strong, but I wonder if Etherum can be changed in the future to work around f.ex. scalability issues? Sorry if my question is stupid, I'm just pretty new to bitcoin
I'm not a computer scientist so use appropriate amounts of salt with my comments. From what I understand there is a fundamentally different design at the base of bitcoin and bitshares (and other graphene based blockchains).
The difference seems to be mainly whereas bitcoin stores and processes operations on the network, bitshares deals with messages (the results of operations). The latter allows for the massive optimizations and speed (not using the inefficient blockchain for the slow stuff, but only for the absolutely essentials). It also allows for more complex stuff like order matching on the decentralized exchange, something that is apparently virtually impossible to do on both bitcoin and ethereum.
Ethereum uses a similar design with bitcoin,
but with a
lot more operations and complexity. While there probably are ways to massively improve bandwidth and tx/s, from what I understand most of those can be applied to bitcoin and because of it's simplicity it would always have higher bandwidth then ethereum using the same techniques.
What ethereum can't do however is migrating to the message-system graphene uses without breaking off completely from their current design (unlikely since all the current customers would have to rewrite all their current apps and software) and that seems to be one of it's achilles-heels. One of the reasons that there is still not a bitshares-level app on ethereum is this design, despite all the devs clamoring that the first thing they'd write would be a dex and there being an opensource working one right here to copy. Also even if the bandwidth issue would be solved, it would still be held back by it's blocktimes and having to do linear operations. Graphene based chains seem to be capable to do multiple operations during a single block on top of the low 3 second blocktimes.
I only skimmed through the article referenced at the top, but superficially he does seem to have a point. Despite all the smugness I'm seeing in the Ethereum-community about the issues with bitcoin, I'm stupefied by the observation that most not realizing they have all the exact same issues, if not even worse in their case. (centralization, governance issues, waaaay to much power in the hands of a few, immediate bandwidth problems) For example, delayed transactions on the "simple" bitcoin network does not necessarily destroy the value propositions of the bitcoin tokens (although it is not a very good user experience), I'm pretty sure that the same situation on ethereum would have more serious consequences.
The hype-machine trying to convince people that ethereum can do and replace everything seems to be more than a little misleading. To be fair though, I've seen people like Zamfir issue public warnings, about not running mission critical code on the network, because of it's experimental state.