There's a lot of issues with sidechains. First and foremost being the same issues we see with multi coin wallets

So okay, instead you have thousands of sidechains that all have the same code base thereby remove the need to worry about this security and maintenance issue. Cool. (this is not the case right now obviously as we have liquid and RSK and all these different sidechains of completely different code bases. Basically this is a fungibility issue)

So then, what does that wallet look like? Do I get "sidechain #1 Bitcoin, sidechain #2 Bitcoin, sidechain #3 Bitcoin" in this long list of sidechains that I'm accepting and should be treating as all being the same? Should I be treating them all the same? If I can, then maybe its all just grouped up together in one entry as "Sidechain Bitcoin". Alright, so under that model, how easily can I be tricked into accepting "counterfeit Bitcoin". As you know, we don't have counterfeit Bitcoin because when you run a node, your node would reject any block that would break the rules. So then, in order to protect myself, I need to run a node for every single sidechain that I want to accept, which in this scenario is thousands. Is thousands of sidechains a reasonable expectation? If its only hundreds, that's still hundreds of nodes I need to run to protect myself.

No I don't consider a neutrino connection to someone's BCrash node to be an acceptable solution.

And then if a wallet comes out that includes a malicious sidechain I could be easily tricked because I'm expecting these sidechains to be fungible. But alright alright, again we say each sidechain should have exactly the same codebase, but still be different enough to not merge onto one chain and we can do hashes on the codebase for every chain added to a wallet and only xyz files should be different.

So then, how good is the security of the sidechain? We'll just say we don't trust it until it gets rolled up in a Bitcoin mainchain tx to glaze over this part. If we're going to wait that long maybe that solves our need to run hundreds or thousands of nodes (unless its the drivechain implementation. People forget that miners are paid security and if you pay them to act up they will)

But yeah, the actual prereqs I kinda wrote about in "Bitcoin Adoption into the Future - Build it Right and Grow from There" which was basically protection against

Fractional Reserve/Rehypothecation Outright theft Censorship

That was problems with custodians though. Sidechains introduce new issues. Lightning works because you have this proof that you can always hold onto as a real guarantee that you can take back to the main chain to prove and get a justice transaction for and data propagation only needs to go to relevant parties, and like, maybe LN Bitcoin from someone's node out there is fake or something, but that doesn't matter because in order to send me Bitcoin, they have to update my channel with my channel partner and our Bitcoin is real. If my channel partner opened a channel with some malicious node and accepted that as real Bitcoin, that's their fault and their problem when it comes to resolve their channel with their channel partner and I don't have to worry about that. Its so fast exactly because I don't need the whole world to know about every single transaction until it comes time to settle.

Man its way easier to express why custodians suck than it is to talk about sidechains. A lot more technical minutia, but I hope I explained my position well enough. I hate sidechains, and think they're a silly solution to scaling, and I will remain skeptical and pessimistic about them until I am proven completely and absolutely wrong and after they're normalized I will probably take dedicating time to doing security research about them way more seriously (unless they don't do the every sidechain codebase is the same idea I came up with in which case I'll just continue to complain)