waxwing @waxwing@x0f.org

@TallTim basically it appeared to be saying that a guy who has just reported himself hacked through sim swapping for $30M is the guy who ran bitcoinbuilder and that site itself was hacked years ago. Etc.

Other oldbies I'm sure remember the delights of "bitcoin builder" with bob the builder on the home page. Crazy times :)

(personally wasn't ever on gox but several online friends had my sympathies, and ... man was that a fascinating process to see happen).

This is, to say the least, interesting:

https://old.reddit.com/r/btc/comments/f7lbae/30m_bch_sim_hack/ficelop/

@TallTim it's not clawback, it's just atomicity - the individual sub-actions within the contract can't happen separately. So borrow-do stuff-pay back either all happens or none of it does.

@TallTim Was going to respond but forgot: the concept of 'flash loan' isn't as dumb as it sounds. They use the atomicity of the transaction to ensure repayment thus obviating the need for collateral.
It seems this mechanism per se wasn't the cause of the failure, not that I'm claiming it is or isn't watertight.

@6102 The diagrams in this thread are great. Recommended. 👍

@6102 It seems to be. (I guess you mean the example code in particular? You could take a read of his long post about the 1.0.0 version linked on the main python-bitcointx repo readme for more info).

I slightly understated it with "evaluating"; I've already refactored the JM bitcoin backend with it in a local branch, but still quite a bit of work to do all the jmclient code, but so far it looks very good to me.

I noticed he even has some interesting code examples for things in Elements:

https://github.com/Simplexum/python-elementstx/blob/master/examples/unblind.py

requires python3.6 (we already require python3 anyway), and I may have to preserve some custom code for tweak_multiplying points and a few other things.

The best thing about python-bitcoinlib and this derivative is you get more or less full bitcoin script support, including VerifyScript() to check a transaction signing actually is correct. (2/2)

Peter Todd's python-bitcoinlib doesn't get much attention nowadays (Peter himself moved away from it a couple years back, preferring Rust iirc for future projects, while a few other coders are still using it, or want to); this guy seems to have done some very serious work with a fork with extra features:

https://github.com/Simplexum/python-bitcointx

I'm evaluating it now as a swap-out backend for JM and I like what I see so far. Supports segwit, bip32, possibly psbt, and allows me to ditch a lot of old code. (1/2)

Taproot security in the generic group model (h/t Tim Ruffing):

https://github.com/LLFourn/taproot-ggm/blob/master/main.pdf

I'm not sure yet from a brief look, but this may address concerns I had earlier about the security reduction (or lack of) for the script spending path.
If it can be reduced to hash collisions then we're fine, since we have that assumption already.

#bitcoin #crypto #taproot

... sound cool in theory, they are even *more* problematic with cryptocurrency than they are with stocks, due to huge volatility on short time frames.

Basically I think the issue here is an assumption that a market is available for an instrument, is not ever a really safe assumption. Markets literally disappear in a flash, as the smart "attacker" figured out here.
(Disclaimer: I am not even close to an expert in this dex/eth/defi stuff, pinch of salt to be liiberally applied).

... and I think that's the key element of what's at issue. In this very detailed breakdown:

https://medium.com/@peckshield/bzx-hack-full-disclosure-with-detailed-profit-analysis-e6b1fa9b18fc

they seem to claim that it is really an error in the bZx smart contract design that the in-default-ness of the borrow is not checked at a certain step. I'm not qualified to comment on that specifically, but my intuition is that ultimately this is about a deep issue in markets - *there is no ACTUAL price*.

Price is entirely contextual. And while circuit breakers, etc, (2/3)

Have people been trying to follow this bZx exploit (x2) story? It's indeed really confusing/complicated but it does seem very interesting.
One key starting point is the idea of a "flash loan" - a loan granted during the execution of a smart contract, that gets paid back within that contract. Thus in theory avoiding the need for collateral.
Then there is the price oracle aspect, which is what gets stressed by the exploit (not the oft considered "malicious oracle" but just lack of liquidity) (1/2)

wt actual f (click video)

https://www.figure.com/blockchain/

😂