I think this is more relevant for ~devs than for the "non-dev" territories, simply because this is only meant for solution space discussion, absolutely not usage, as the warning implies.
Speaking of dev perspective, there's a question I'm starting to ask myself more often based on this and other recent PWA and web library implementations/demos with advanced encryption:
is webassembly inevitable?
It's still inferior to the Web Crypto API (SubtleCrypto) within the scope that provides, and of course very inferior to decoupled cryptography, like letting your keyring or yubikey gen/derive/sign. But as long as sandboxing isn't totally broken (last time was in 2025) and when combined with a service worker... it could become the only way to cope with the extremely slow process on the side of the browser implementations.
I have switched wasm off on my phone though; also because per Google, all the V8 sandboxing bugs are AST/JIT optimizer logic errors, not things like read after free that you'd catch with memory tagging.
I think this is more relevant for ~devs than for the "non-dev" territories, simply because this is only meant for solution space discussion, absolutely not usage, as the warning implies.
Speaking of dev perspective, there's a question I'm starting to ask myself more often based on this and other recent PWA and web library implementations/demos with advanced encryption:
is webassembly inevitable?
It's still inferior to the Web Crypto API (SubtleCrypto) within the scope that provides, and of course very inferior to decoupled cryptography, like letting your keyring or yubikey gen/derive/sign. But as long as sandboxing isn't totally broken (last time was in 2025) and when combined with a service worker... it could become the only way to cope with the extremely slow process on the side of the browser implementations.
I have switched wasm off on my phone though; also because per Google, all the V8 sandboxing bugs are AST/JIT optimizer logic errors, not things like read after free that you'd catch with memory tagging.