Is it too late to propose to freeze Bitcoin protocol as it is today (no SegWit, no bigger block)?
Hear me out: assume Bitcoin is frozen as it is now: no more forks, no more change in consensus rules. Call it Bitcoin 1 (BC1). So, how to evolve and scale Bitcoin?
Use a very conservative sidechain, call it Bitcoin 2 (BC2), tightly coupled, block per block, with Bitcoin 1 chain. No fancy experimental changes there: fix the bare minimum to support more general two-way sidechains, linear scalability transaction verification time and transaction malleability. Can be one way only if that makes it simple enough: coins sent to BC2 can never come back to BC1.
Miners choose to mine a BC2 block alongside a BC1 block, and it can be done for free once a BC1 block is found. There may be BC1 blocks in the chain whose corresponding BC2 blocks are missing (and that is not a big deal), but there will be an economic incentive to mine a BC2 block, since transaction fees on that chain will be collected.
The economic adoption is completely voluntary and independent of anyone else, doesn't require the strong consensus of SegWit. If it is feasible to create a simple enough two way sidechain over BC1, good. If not, economic value of BC1 will always be at least that of BC2, but never lower. BC2 may have a lower price than BC1 at first, but since it will have every feature of BC1, plus the possibility of massive scalability and instantaneous transactions with a lightning network built over BC2 (a practical economical advantage), it will most likely value over what BC1 could ever be alone, raising BC1 price with it.
If these predictions are correct, fewer and fewer people will transact BC1, eventually with blocks mined empty except for the coinbase, which the miner may choose to send directly to BC2. If I turn out to be wrong, and people prefer to stick to the old BC1 for some reason, BC2 becomes just another speculative altcoin. Thus, everyone migrating should be fully conscious about the risk of immediate devaluation.