Something I'd like to check here is did LZ fail to see Elf's good moves or see but misevaluate them?

Next nice move of Elf, this cutting across knight's move in response to white j16. It's the natural shape weakness so I think plenty of strong humans could find it, but the timing is smart. Letting LeelaZero #150 analyse for ages she considers k15 practically the only move with 120k playouts and 28.4%, e17 gets 5 (not 5k). Once played LZ quickly realises it's a good move, black win up to 35.3%, a 7% swing which is pretty huge for 1 move. A severe blindspot, looks like LZ needs to be more exploratory.

I think such blind spots are a tricky problem also involving the search code of the bot. Even in training it will only learn moves that its search can find within the low selfplay visit limit (in each position it is trained towards search results). So if there is a blind spot in the net, that can only get fixed if the bot doesn't rely too much on its policy during selfplay search - otherwise it will just keep reinforcing the oversight.

IIUC, Monte Carlo bots are guaranteed to find the correct play in infinite time. I would be surprised if the Zero bots don't meet the same criterion. OC, that does not mean that they cannot develop blind spots that will last for millennia.

I pointed out the possibility of self reinforcing blind spots long ago. In fact, I suggested the possibility that AlphaZero Chess reached a self play plateau that could not beat a version of Stockfish which used an opening book and endgame table base. I am unaware of anybody else suggesting that.

People used to be pretty worried about self-play plateaus until AlphaGo Zero came along and slaughtered all the top pros. But they may have been too quick to stop being concerned about it.