I've been reading anonymint's writings for the past week or so, which also prompted me to dive into some other rabbit holes.
I'm more convinced now of the dangers of segwit. Don't mistake that for being a promotion of bcash.
Could you provide me with a link where I can read about that? I remember anonymint's post, but I did not pay enough attention and now I can not find it.
Geeze, guys. We've been discussing these very same aspects of segwit since years. Have you had your fingers in your ears and blinders on up 'til now?
Agreed. It's unbelievable how people are coming NOW with points that were debated for years and were (even if slightly) a concern until a few months ago. And now that those are completely debunked they act as if they just discovered em....
Except they haven't been debunked. Well, other than with a handwavey 'we don't believe this to be a significant exposure'.
I don't believe it to be a significant exposure.... But last time we discussed this subject I was a bit more verbose than just handwaving:
(snipped for relevance)
I don't see exchanges, devs and users complaining how flawed Segwit it, nobody lost money ore saw critical errors. I only see that bullshit in the Bcash camp.
Can you provide me technical arguments/proof why Segwit is 'flawed' ore show it at the Github?
Yes. Fungibility.
What's the issue with fungibility that's so specific to segwit?
As I posted between there and here, Segwit creates three classes of Bitcoins. Each with distinctly different exposure to security vulnerabilities.
1) Those that are completely free of any Segwit taint all the way back to their constituent coinbase transactions;
2) Those that are not currently output from a Segwit transaction, but have Segwit taint between here and their constituent coinbase transactions; and
3) Those that are the output of a Segwit transaction.
Reliance on miners not to revert to 'anyonecanspend' - an incentive for which only increases over time.
That is, reliance on miners not to try a 51% attack. Does this imply the chain without segwit is invulnerable to 51% attacks?
No. But without Segwit, all miners were able to do with a 51% attack is roll back transactions. They were unable to steal funds. If miners choose to revert to considering Segwit transactions as anyonecanspend transactions, then they can claim every one of the outputs of all those anyonecanspend transactions for themselves. As over time, transactions tainted by Segwit is a monotonically increasing count, the incentive to roll back to Satoshi rules is ever-increasing. And the funny thing is that this would arguably not be stealing. After all, Segwit is said to be compatible, right? All the miners would be doing to claim these funds is to revert to the previous rule set. That's compatible.
The risk you are mentioning here is somewhat "real" but it is decreasing as more funds get moved to segwit addresses. It was a real concern in the beginning as it would have been a (remote) possibility that miners would do as you say. In fact, even if I believed Segwit was a favourable upgrade I was very cautious to NOT move my funds to Segwit addresses until many months later.
Currently, with so much funds already moved into Segwit addresses IF miners decided to stole any Segwit address this is what would happen:
- A fork between the consensus chain and the stolen one.
- A drop in price to almost zero on the stolen one.
- A considerable drop in price on the consensus chain too.
... And basically a major drop in ALL cryptocurrency market as confidence and trust on the main cryptocurrency and its foundations would be severely harmed.
Or maybe it would just be that noone would give any value to the forked (stolen funds) chain and the crisis would not be as deep... but, in any case, whomever carried out that sort of attack (necessarily a main player in crypto mining) would end in a useless (no value) chain and be economically harmed in every way on its core business.
Also take into account that exchanges can be hit by a 51% attack in the sense of double spending BUT they are not vulnerable to this type of attack. Why? Because their wallet software would NOT acknowledge the "anyonecanspend" spoofed tx's no matter how many hashrate decides to support the rogue chain. (If I am wrong in this point please enlighten me).
So, still a 51% "double spending" attack -while hugely improbable in Bitcoin for many reasons- is a (orders of magnitude) bigger "vulnerability" than a segwit "anyonecanspend" attack right now.