Any idea what this is referring to?
schemes that make malleability irrelevant are subject to dangerous signature replay attacks if not handled very carefully
Is he saying that implementing
BIP 62 opens up a new known attack vector?
What I meant was the idea that what goes into transaction should be "open to the user".
Imagine you had a database and added to the ability to store arbitrary information into each row, this is why rational databases exist which require you to define the type of data you want to store before you do add that information. The game of whack-a-mole is because even when they remove malleability for necessary transaction data it still doesn't prevent that attack because each entry has "scrap space" after that.
My suggestion is to abandon that concept because it's not a sane approach to storing data but a software engineering nightmare.
Sorry, but I still don't get it. If BIP62 was implemented, what new attack vector does it open up? What's this "scrap space" you mention? BIP62 appears to shut down all the different ways to maleate a transaction and specifically addresses "Superfluous scriptSig operations" in step 6, which is the closest I can find to anything that might be considered "scrap space".
And you guys have the nerve to call other crypocurrencies "shitcoins".
Well, they are mostly just clones of bitcoin anyway, and so have exactly the same issue unless they fixed it themselves. It's not like copying the bitcoin source and changing a few numbers fixes anything.
This is great news. It exposes the vulnerabilities and weaknesses of bitcoin and allows for better cryptocurrencies, like Litecoin, to grow.
How was this fixed in Litecoin? Do you have a link to the pull request please?
So my reason is to protect your life savings from this ponzi scheme called bitcoin

I want to prove that decentralized trustless system can not exists in long term.
It either transforms to centralized system or loses its security.
But this attack proves no such thing.
The ongoing active mutation of transactions made me wonder, whether targeted mutation could be leveraged - by miners or nodes - to facilitate the process:
Yes, and that sounds like a good solution. Miners could mutate all transactions into their 'canonical' state before mining them. That way well behaved wallets aren't affected, and wallets creating weird transactions still have their transactions mined, but with a different txid.