They're blaming them the same way they blamed them last time - their buggy client is crashing by running out of ram due to an x-thin block bug. They claim that core and/or blockstream is DDoSing them intentionally although they can't actually find evidence of it. Even if it IS someone DDoSing them, it's a given with there being so much value in bitcoin that attacks will be a routine part of operation and clients should be hardened against all forms of attack. Additionally this is the same bug they had last time; they didn't even fix it, and that makes it the 3rd critical bug they've had in the x-thin code. The x-thin code is one of their only unique features in the BU client and one of their claimed workarounds for transmitting large blocks across the network - but now they're recommending people disable it as a workaround. Add to that the thought that parallel block validation will help somehow - and that will use even more ram so regularly running out of ram seems intrinsic to their design at the moment...

Once more the same BU morons will just put it down to yet another simple bug that's easily fixed and nothing to worry about blah blah blah... even though the thought of a 2030 billion dollar industry that was this unreliable would be terrifying. The only nodes that have stayed up are those on mega servers with >64GB ram. Of course mining pools that are advertising BU support in their coinbase are NOT affected since they don't trust the BU client software; they're just using the core software with a customised coinbase signature - they'd never trust their mining operations to the buggy BU client.

Note I'd already created as similar thread here but since I made it self moderated it didn't attract the attention of the full time anti-core brigade.
https://bitcointalk.org/index.php?topic=1908442.0