Is merge mining the best defense against this? What are the downsides to it?
As with merged mining you have potentially access to the hashrate and security of large coins, including Bitcoin, I think it is indeed the "best" or "most secure" PoW alternative for smaller altcoins.
One of the downsides is of course that if you have already a group of miners sustaining the chain, very likely this group will lose its business model if this implies an algorithm change and the hardware of the "traditional" miners isn't optimized for the new algo. The transition from independent mining to merged mining could have thus some potential for attackers, when the merged mining hashrate is still not fully developed. For example, if Monero introduced merged mining with Bitcoin, it would take some time until all Bitcoin miners have upgraded their software, and the "old XMR miners" during that time wouldn't be able to "help" with hashrate. Maybe a hybrid transition with both algos in parallel is possible though, but it could add complexities and perhaps new attack vectors.
And there's the problem that perhaps it isn't worth the hassle for all miners to upgrade to merged mining in the case of a large coin like Bitcoin. It could perhaps even be better for a coin like XMR to introduce merge mining with the Doge/LTC "complex", as for these miners the Monero income would be far more substantial than for Bitcoin miners.
IMO the biggest problem of merged mining is if the "hybrid hashrate" is too low, e.g. if only 10% of BTC miners merge-mine XMR, then a big Bitcoin pool could try to 51% attack the merged-mined coin, like it occurred with CoiledCoin in 2013 or so:
see this Reddit thread. If instead the incentives are high enough to ensure a high "hybrid hashrate" this attack should be very difficult (this could be again easier if the merged-mined coins are all altcoins, see previous paragraph for the example with Doge/LTC).
There are also some issues with MEV, e.g. if you're merge-mining Coin A and Coin B, and there was just a block in Coin B found with a large amount of transaction fees (like the BTC halving block in 2024 where a lot of NFT guys tried to include a Rune or Ordinal transaction), then another miner could try to reorg both chains to collect these fees himself. I'm not expert enough to evaluate that risk in depth though. Sergio Demian Lerner wrote an article about disadvantages of MM
here. It is focused on sidechains though.
Qubic miners are crazy if they think they can control Monero for an extended period of time. They are the largest privacy network in the world.
Currently I agree. There could be a problem however if other groups with other kinds of interests join the Qubic miners. So I would really recommend a "Plan B" to the Monero folks, some nuclear option for a "real" attack.