From my perspective (going back a long ways) real sidechains such as liquid ARE 'bitcoin'.  What they peg out is protected by all of the proof of work of Bitcoin proper.  They can also leverage a lot from basic operations within Bitcoin proper without fucking it up so much.  I always viewed sidechains (even before they were developed) as the escape valve for entities who wanted so leverage real Bitcoin (instead of some shitcoin), but enhance it's capabilities.

So, I don't think it exactly appropriate to say 'forget sidechains'.  (Sidechains in my mind are basically a viable shitcoin PLUS some credibility via a BTC peg.)

Now it must be said that I basically ignored Bitcoin developments since the blocksize wars a decade ago and have not been very interested or in-the-loop.  As back then, I feel OK as long as blocksize growth is _predictable_.  As I understand things this OP_RETURN change does not in and of itself increase growth rates (though it looks like a stepping-stone to me).  It mostly just makes it more practical to buy 10 minutes +/- of the Bitcoin world's time in one transaction.  Same thing could be done, I guess, with the existing OP_RETURN, but it would be a little bit more of a pain-in-the-ass perhaps?

I don't see a technical reason to give the spammers what they want on a silver platter in part because with the sensible rates which have been more-or-less preserved since the infamous 1MB commit, we really don't need ultimate efficiency in mempool or pruning.  (Moore's law is BS, but hardware capabilities are improving.)

As with 10 years ago, I would argue that it is worthwhile to educate people that if shitcoiners attack and transactions are delayed for hours/days, it's doesn't mean that the sky is falling.  They spammers will get tired of spending millions/day in order to choke things up.  Just set on-chain transaction expectations accordingly and use other (much better) sidechain options for day-to-day transactions.

On out-of-band miner buying, seems to me that two could play at that game.  The 'two' could be legitimate/efficient sidechain operators who could only have to cover the difference between that natural block reward and what the spammers are offering in order to keep their business model operative.