> 2017-18 when it became apparent that the bitcoin network was unsuitable
That's an extraordinary statement. The scaling aspects of Bitcoin was apparent to everyone involved in the project from the start, including Satoshi.
The limits of an architecture where everyone processes everyone else's transactions should be obvious to anyone reading the whitepaper. It is also the focus of the very first email replies. Few respondents chose to focus on what is possible instead of what is impossible.
Bitcoin doesn't need "everyone processing everyone else's transactions", only full nodes need to check transactions for validity. Bitcoin's whitepaper proposes SPV wallets, which are meant for end users and require minimum disk space and network bandwidth, and no need to run a full node.
The answers Satoshi gave were convincing, which is why the project was able to grow in the first place, and why so many of the people involved from the start wanted it to become a global payments solution.
Computers are fast, Bitcoin payments are simple, and we can get data on roughly how common payments are. Even with the hardware of 2011 it wasn't implausible to keep up with PayPal with optimised software, and with multiple-machine nodes you could even keep up with the bigger credit card processors. That was in 2011. Since then we have things like AMD Rome and 128 core machines, NVRAM that's only 10x slower than DRAM and so on. Satoshi bet on hardware improvements and his bets were solid.
Meanwhile, there is no requirement in the architecture to never truncate the on-disk copy of the block chain. Some people should keep it for historical reasons and to preserve the low-trust nature of the system, but there's no requirement for every full node to keep it, and no requirement that everyone use a full node.
Bitcoin and scaling is unfortunately a topic that became horribly corrupted around 2015 by ideology, full blown information warfare and an overtly political manipulation of the community. As a consequence there are a lot of people who think, incorrectly, that they understand the design of Bitcoin better than Satoshi did, and who think it needed to be completely replaced with the "Lightning network". The reason this accompanied massive levels of censorship, DDoS attacks, intimidation and outright extortion is because the arguments were bad and the proposed alternatives, even worse. The original Bitcoin design was fine for what it needed and represented some pragmatic engineering that balanced complexity against scaling and decentralization. What Blockstream forced on the community was in contrast, significantly worse.
Blockchains does not require append-only. There are pruning techniques like Automatic Transaction Rebroadcasting coming out on high-throughput blockchains like Saito:
These techniques will eventually get ported over to Bitcoin or whatever survives long-enough to need it. We'll get 20, 50 and 100-year rebroadcasting chains. Fees will be higher if the data needs to be stored longer. Spectrum of uses from investment to cash with different degrees of value-persistence.
Historically, there has been a number of ideas proposed. Many different variants of subchains and sidechains, different types of payment channels, ecash systems and so on.
The idea that someone discovered these in 2017 seems oblivious to what actually happened. Support for payment channels with nLockTime was likely influential for implementing transactions with smart contracts in the first place. It was there from the first Satoshi release, and payment channels was described already in 2013. It probably wasn't first on the list of things to fix as long as transactions were heavily subsidized by inflation.
That's an extraordinary statement. The scaling aspects of Bitcoin was apparent to everyone involved in the project from the start, including Satoshi.
The limits of an architecture where everyone processes everyone else's transactions should be obvious to anyone reading the whitepaper. It is also the focus of the very first email replies. Few respondents chose to focus on what is possible instead of what is impossible.