So while reading some old threads/blog posts on the black hole firewall paradox, it occurred to me that I had some residual confusion regarding why firewalls (supposedly) form at all. IIUC, the argument is that usually, the vacuum is a highly entangled state, and that disentangling it (in order to be able to save entanglement monotony and preserve unitarity of Hawking radiation) produces a maximally mixed state, which is infinitely hot. Basically, we trace out the vacuum on the other side of some boundary (e.g. the horizon), and what's left will be maximally mixed, i.e. a convex combination of states containing any possible number of modes. But of course, the resulting reduced density matrix would have been the state according to an infalling observer, anyway---even if the entanglement is not broken, if an observer has access to only one part of a system in a maximally entangled state, his local physics are described by a system in a maximally mixed state. In the simplest case, if I have two entangled qubits, and one is behind the horizon, if I jump in, and have no access to the second qubit anymore, the qubit I have with me will be maximally mixed. No? I think the resolution of this must somehow be that I'm mischaracterizing what an observer has access to. I'm imagining something like a 'bubble of vacuum', which is entangled with its exterior, and which, if we don't have access to the exterior, ought to be in a maximally mixed state. But actually, the 'state I have access to' is always given by (in the extreme) my cosmological horizon, or my causal past: whatever's influenced my state here and now. So I don't only have access to my bubble of vacuum, but also to its purification---and thus, observe it in a pure state. It's only when I'd do something to break the entanglement of the bubble with the rest of the vacuum that I would create a ball of fire, with the energy for this presumably coming from whatever disentanglement process I have performed. Does this make sense to anybody?