A few days ago, Stephen Hawking wrote a paper arguing that firewalls are pure crap (so far so good) because they violate CPT, much like black hole remnants (it's no good anymore: these objects are wrong but not because of CPT violation).
The world's most famous theoretical physicist along with his high-tech friend
But he also says that event horizons violate CPT as well, they can't exist (see the media echo chamber that puts Hawking into the black holes don't exist category), and the right resolution is that the surfaces of black holes are chaotic like turbulence and predictions are as impossible in practice but possible in principle as (long-term) predictions in meteorology. Well, I would say that "surfaces this chaotic" would act as firewalls. Moreover, event horizons that are as perfect as in classical theory violate CPT but event horizons which allow the Hawking radiation (unlikely but in principle allowed violation of the causality) are compatible with CPT!
So I don't really believe that Hawking's criticism of event horizons is kosher.
Hours ago, Fox News and Nude Socialist have upgraded this Hawking story to a higher level:
Black hole theory my 'biggest blunder,' Stephen Hawking says (Fox)Wow.
Stephen Hawking's new theory offers black hole escape (NS)
While I have sympathized with Hawking's unambiguous indication that the firewall theory is wrong, his claims that the black holes don't exist at all – and he is saying that because a black hole is defined by its having an event horizon, so if no objects possess event horizons, it means that there are no black holes – is a little bit too much of a good thing.
As you know, Albert Einstein has used the term "my biggest blunder" for the concept of the cosmological constant, an additional term (they would translate the term as the "cosmological member") that he inserted into his pristine Einstein's equations of general relativity in order to supplement the Universe with a repulsive force that protects the Universe from collapsing (the collapse is what the attractive forces between galaxies in the Universe want to achieve). In reality, the cosmological constant is indeed positive and of the same order that Einstein was imagining. However, it's actually slightly larger (by a factor of order 2) so it's enough not only for stabilizing the Universe but it makes the expansion of the Cosmos accelerate.
(The stabilization of the attractive and repulsive forces that was envisioned by Einstein wouldn't work anyway because the balance would be unstable: a tiny deviation from the "perfect balance" in either direction would be exponentially growing in time.)
The great physicists' criticism of their famous theories as the "biggest blunder" may look analogous – they have the right to criticize their ideas, it's sort of "cool", and both of their criticisms end up being wrong – but there's still a difference. The cosmological constant was something like Einstein's tenth most important contribution to physics, maybe the twentieth. However, the insights on black hole thermodynamics are (by far) Stephen Hawking's #1 contribution to science so Hawking's present criticism is "bolder".
I like his focus on symmetries, in this case CPT. He says that remnants, firewalls, and even event horizons violate CPT because they introduce "new arrows of time". But the event horizons actually don't – and ironically enough, the reason why they don't introduce any time-reversal asymmetry or CPT violation is nothing else than the Hawking radiation. To be candid, I must say that the remnants and firewalls don't violate the CPT, either – although they're wrong for other reasons.
The black holes may swallow matter and their event horizons look like a one-way membrane. This would be a violation of CPT. But the claim that the event horizons are one-way membranes is just a statistical claim, like all the claims about the "mostly increasing entropy" according to statistical physics. It is not strictly impossible for matter to get out of the black hole through the event horizon, i.e. in the "opposite than normal" direction. It's just very unlikely!
In this sense, the black holes don't differ from the furnaces. Coal only seems to be moving in one direction – into the furnace – but in principle, things like coal may escape from the furnace, too. It's just very unlikely that the ashes and smoke that leave the furnace combine themselves into a large chunk of coal.
The black hole Penrose diagrams are CPT-violating – time-reversal asymmetric – but this asymmetry is not a fundamental microscopic law of Nature. Instead, the reason why we don't revert the picture above upside down is that it is simply extremely unlikely for a black hole, an object, to "emit" a Hawking radiation that looks like a star. Such a backward evolution would be reducing the entropy and this is very unlikely.
We are only discussing the diagram above – which only has the singularity in the future but the past looks like a topologically/causally empty Minkowski space (with a star) – because the time-reverted process is exponentially unlikely so it won't ever appear in practice (if the black hole is large enough). But it is not strictly prohibited. The previous sentence is the reason why Hawking's claims that event horizons inevitably violate CPT is just invalid.
The eternal black hole starts with an initial singularity ("white hole") and the configuration is CPT- or time-reversal-invariant. Large objects would be extremely unlikely to move through the "antihorizons" which is why the object above can't be combined with any other matter to yield a plausible process in the real world. The Penrose diagram must be modified by replacing the "white hole" (past portion of the spacetime) by the collapse of the star (or something else) – we get the previous realistic picture. Similarly, the previous realistic picture emits a near-thermal Hawking radiation into the future (while doing so, the total entropy of the hole plus radiation is going up, too). If time-reverted, we obtain a near-thermal incoming and convergent (high-entropy) Hawking quanta, an initial state that has no reasons to be discussed because it's hugely exponentially unlikely (the entropy would be decreasing when the quanta combine into a black hole, too).
Einstein said that the cosmological constant was his "biggest blunder" more than half a century before the experimental discovery of a nonzero cosmological constant (in the late 1990s). Most cosmologists and theoretical physicists tended to believe that the cosmological constant was exactly zero before it was experimentally found that it was strictly positive. The experimental case for the existence of black holes – with event horizons – is arguably equally strong as the case for the cosmological constant, if not stronger. So Hawking's repentance is arguably much more embarrassing than Einstein's. The Hawking radiation hasn't been experimentally found but it's extremely well theoretically understood and it's theoretically clear that black holes with event horizons and the Hawking radiation don't violate CPT and don't suffer from any other lethal problems suggested by Hawking.
For decades, I have been impressed by Hawking's being so right about so many cutting-edge questions (except for some important exceptions where he lost the bets, of course, like the information loss, the existence of the Higgs, and some others), despite all the hurdles he is overcoming while forming his own opinions. Unfortunately, his rejection of the black hole has changed this impression in a significantly negative direction. It's silly not to believe that black hole horizons exist.