but what about the wallets that is using these weak RNG or even bad entropy back in the day? So Should people we worry more about that than quantum threats?
Old wallets with weak RNG are likely already emptied by now. Note that whether the RNG is solid is a matter of secure implementation rather than software age. The original Bitcoin wallet / Bitcoin Core never had this issue. At least one Android wallet did back in 2013ish (these are the old wallets that are likely already emptied by now). I'm not aware of any cases since then, but it can always happen with whatever new wallet hits the market, especially if the dev team is inexperienced.
As I see the situation with quantum calculations, the main concerns are:
1) **P2PK** (Pay-to-PubKey) early Satoshi addresses with much BTC
2) REUSED addresses with some BTC amount (which have already exposed their public key
2) Mempool transactions (with visible public keys)
As for the First, a consensus or a Fork Or something like that is needed to block transactions from those addresses
As for the Second, mass education for BTC holders is needed to move BTC from reused addresses to New ones.
As for the Third, I don't know.. Does anybody know a solution?
Quantum resistant candidates for replacing ECDSA exist, however IIRC the issues lies with both performance and signature size. So worst case Bitcoin will switch to one of those, once the threat timeline becomes clearer; best case someone comes up with a better signature scheme in the meantime.