Trespassing on Einstein's Lawn (54 page)

Someone, it seemed, had to be wrong. If it's Safe—if B is entangled
with A and not with R—then the correlations strewn across the cloud of Hawking radiation are severed and information is lost. But who would be willing to allow information to be lost after physicists had just spent decades finding it? Hawking would have to change his mind, again. Elephants would disappear from the universe. The brilliant moral of AdS/CFT—that black holes are dual to information-preserving quark-gluon plasmas—would be wrong. The Schrödinger equation would fail. Quantum mechanics would become nonsensical. Nearly all progress in fundamental physics over the last thirty years would spiral down the drain.

No, information loss was not an option—which meant that Screwed had to be wrong. B was entangled with R, not with A.

Unfortunately, that wasn't any better. Severing the entanglement between B and A is equivalent to inserting a horizon—after all, that's what a horizon does. The vacuum no longer cancels itself out. Instead of emptiness, there are particles.
Hot
particles. Planck-temperature, molten, scalding particles. A firewall.

“Perhaps the most conservative resolution is that the infalling observer burns up at the horizon,” the paper concluded. Screwed is even more screwed than we thought.

Now I was starting to sweat. If Polchinski and his crew were right—if B is entangled with R and not with A—we'd have to throw out not only horizon complementarity but
general relativity
as well. After all, the very fact that Screwed should find himself in a vacuum, with nothing out of the ordinary occurring at the horizon, is a consequence of the equivalence principle. Relativity had always said that Safe would see Screwed get burned to a crisp at the horizon. But that was only from Safe's perspective. In Screwed's own frame, nothing bad was ever supposed to happen. He wasn't supposed to feel any heat, just like the guy falling off the roof doesn't feel any gravity. Now this paper was suggesting that Safe's perspective is the
right
perspective. That inertial and accelerated frames aren't equivalent, however happy Einstein's thoughts. And if that's the case, all bets are off.

I didn't know what to think. There was no way the firewall argument could be right, and yet I couldn't see how it was wrong. But it was
okay, I told myself—Susskind would. Surely he'd read the paper, spot the flaw, and everything would be back on track in no time. There was no need to panic.

Sure enough,
Susskind posted a paper on the arXiv called “Complementarity and Firewalls.” I breathed a sigh of relief—the madness was over.

Or so I thought. Weeks later, Susskind retracted the paper, the notice on the arXiv reading merely, “Withdrawn because the author no longer thinks it is correct.” Just like that, we were back to firewalls.

Over the next few weeks I kept a close eye on the arXiv, waiting for a solution to appear. Bousso posted a paper: “Observer Complementarity Upholds the Equivalence Principle.” Then he withdrew it. Daniel Harlow posted one: “Complementarity, Not Firewalls.” Then he withdrew it.

What the hell was going on? I wondered. It seemed that the arXiv had turned into a complete clusterfuck. Papers appearing one day, disappearing the next? Everyone was quick to shoot down the AMPS paradox—which had been named for its authors' last initials—because the firewall scenario seemed so obviously wrong. Yet with every attempt it became increasingly clear that those damn firewalls were not about to be extinguished. The physics community was nearing a panic. I could feel it.

“Just ask them what's going on,” my dad said on the phone. “I'm sure they don't really think firewalls are real. They can't. Ask Susskind. Hell, ask Polchinski.”

I emailed Susskind first. “Is complementarity in trouble?”

“I don't think the concept of complementarity is in danger,” Susskind replied, “although the AMPS paper is pointing to a much deeper understanding of it. I think it's less of a question of ‘Complementarity or firewalls?' as it is ‘When complementarity and when firewalls?' ”

Complementarity, he said, would have to hold up until the black hole evaporation's halfway point in order to prevent cloning. After that, the firewall would take over.

I wanted to feel halfway better, but I had always been more of a black-hole-is-half-empty kind of girl. So I emailed Bousso, hoping for better news.

“I originally thought that complementarity, properly implemented, would get rid of firewalls,” Bousso told me. “But I now believe that complementarity is not enough. The answer may be firewalls, but I'm hoping that by asking what it takes to avoid firewalls, we'll learn something deeper, hopefully something about the fundamental description of the infalling observer, and thus of cosmology.”

You want a description of Screwed? I thought. Just look around.

Stressed, I emailed one of the AMPS responsible for the mess, namely, P.

“Do you really think that horizon complementarity is incorrect?” I asked.

“I am very puzzled,” Polchinski replied. “I do not see any outcome that is fully satisfying. I expected that complementarity would survive in some transmuted form, but as I look at the follow-up papers, I do not see anyone doing any better with this than we did.”

I texted my father:
Okay. It's officially time to freak out.

I sat outside on my balcony, sipping a glass of wine, trying to make sense of things. The Moon loomed large over the city. The air was thick and warm, bathed in the flaxen glow of Jupiter, a steady light in the western sky perched high above the skyline, the river's glassy, dark water catching flecks of lightfall below, the whole city awash in stillness. Everything holding its breath.

I felt like the world we'd spent all these years building was about to crumble—or, more to the point, burn. I had finally gotten the book deal. I was finally ready to write the book my father had invented for us, the one in which we were supposed to solve the riddle of the universe, and now everything we had learned about the universe was evaporating right in front of me.

Could the firewall argument possibly be right? Could horizon complementarity possibly be wrong? It would undo everything. It would mean that spacetime was invariant again, that we'd have to add it back
to the IHOP napkin, along with
particles/​fields/​vacuum.
It would mean that the multiverse could actually exist, that reality was not an approximate concept, that the second observer was not a copy, that the Aleph was not so hypothetical or secret after all. It would mean that coauthorship was legit, my decision to go solo unjustified, my dissertation retroactively bogus. It would mean that Einstein's happiest thought had turned seriously depressing, that heavier vaginas do fall faster than light ones, that the laws of physics aren't the same for every observer, and that there
is
a preferred frame, namely, the one in which you don't get incinerated out of nowhere by a fucking firewall. That there was a real, observer-independent world out there, one that was not nothing but something, something fundamentally and devastatingly inexplicable. It would mean that after seventeen years, I was just as clueless as when this whole thing had started.

* * *

A few weeks later I packed a bag and headed west again, to Palo Alto.

When I arrived at the conference I saw a slew of familiar faces among the crowd. Susskind was there, of course, as was Bousso. Banks wasn't around but his collaborator, Willy Fischler, was. I spotted Juan Maldacena, Don Page, John Preskill … so many brilliant thinkers whose contributions had, over the last few decades, erected an unbelievable theoretical edifice that was now on the verge of collapse. They looked nervous. Polchinski, the unlikely troublemaker, was there, as were A, M, and S. A bust of Einstein watched over the lounge. He looked nervous, too.

When it was time to head into the conference room, I made my way toward the back. “The room is divided into three sections,” Susskind had explained to me. “The comfortable chairs up front are for participating physicists. The slightly less comfortable chairs behind them are for physicists who aren't speaking. And the uncomfortable chairs in the back are for observers.”

As I sat down and readied myself for the first speaker, I couldn't help but smile. Everything I knew might be wrong and I had been relegated to the uncomfortable chairs—but I had been
invited.
They say there's a first time for everything. I looked around. I wasn't trespassing anymore.

Polchinski opened the proceedings by walking everyone through the AMPS argument. “We thought Lenny would set us straight,” he said. “I'm glad to see you're all as confused as we were.”

Soon Lenny got up to speak. “I am pretty much deaf and I can't hear you when you ask questions, so don't bother,” he began. Everyone laughed. “What?” he asked. “Did you say something?”

But as he spoke, Susskind grew more serious, wondering aloud whether firewalls might be telling us that reality is even more observer-dependent than it seems.

“When I was a young person,” he said, “I think what I was thinking was that A equals R. I thought it was redundant to describe both. But I became scared of that after you guys pointed out how crazy that would be. And I'm still scared of it.”

That A equals R—that the vacuum mode inside the black hole and the early Hawking particle outside the black hole are two radically different descriptions of the very same bit—made perfect sense to me, fitting, as it did, every intuition I had developed over the course of our journey. As far as I was concerned, that
was
the deep insight of Susskind's horizon complementarity and always had been. It was certainly the point Banks had driven home with his and Fischler's holographic spacetime. But the problem now, the one that AMPS had pointed out, was that there's a time during which Screwed can see
both
descriptions, A
and
R, simultaneously, violating entanglement's monogamy, creating the deadly firewall.

Okay, let's assume for a moment that firewalls are real, said Douglas Stanford, a young physicist, when he stood up to speak. What would that mean? Perhaps what's happening is that the singularity is moving from the center of the black hole out to the horizon. “The black hole has no interior,” he said. “The singularity is the edge of space.”

Well, sure, that was true for Safe, I thought. For all intents and purposes, there is no other side. The problem is, it
shouldn't
be true for Screwed. Einstein says Screwed is in an inertial frame—there's no edge of space for him. I shifted in my uncomfortable seat. Why was everyone so calm? If the black hole has no interior for Screwed, every profound advance in theoretical physics in recent history would unravel. Why wasn't everyone freaking out?

“At least that way all the weird physics is happening at the singularity, and we don't know what equations apply there anyway.” Stanford shrugged.

It was at this point that Andy Strominger, the Harvard string theorist, lost his shit. “You have flat space and a singularity appearing of out of nowhere?” he shouted from his seat, incredulous.

“That's exactly the problem with firewalls!” Bousso shouted in reply.

“Well, I'm just glad you're saying it in such a transparently absurd way!” Strominger yelled, his voice dripping with sarcasm. “I think it's wonderful that someone drew a singularity coming out of flat space with a straight face.”

I felt vindicated by Strominger's outrage. This was no time to be polite.

But the mood in the room shifted when a young McGill University physicist named Patrick Hayden gave an inspired talk. The AMPS paradox, he said, is based on the assumption that Screwed can make a measurement on B that will reveal its entanglement with R before he falls into the black hole, where he will find that B is also entangled with A—if it weren't for the firewall, that is. But we have to ask, Hayden says, what would it really take for Screwed to perform that measurement? What would it take to decode the scrambled Hawking radiation and extract the information from the correlation between B and R? In practice, picking out a correlation in Hawking radiation is even harder than, say, looking up a word in a dictionary after it's been burned in a fire. The Hawking radiation is
seriously
scrambled. Decoding its information would require the most powerful computer imaginable. Namely, a quantum computer.

Quantum computers exploit the power of quantum superpositions to quickly perform calculations that an ordinary computer couldn't pull off even given billions of years. Whereas an ordinary bit of information—the kind manipulated by an ordinary computer—is either a 0 or a 1, a quantum bit, or qubit, can be a 0, a 1, or a superposition of 0 and 1 at the same time. As you add more qubits, the number of simultaneous states a quantum computer can occupy grows rapidly. Ten qubits can be in 1,024 states simultaneously. Twenty qubits can be in well over a million. Three hundred qubits can be in more states simultaneously than there are particles in the universe. That a quantum computer can perform so many calculations simultaneously means that it can, in principle, factor large numbers into primes, search vast databases in an instant, and, quite possibly, decode a cloud of Hawking radiation. Who cares that the largest quantum computers built to date had only a handful of qubits? The question, Hayden said, is about what is measurable
in principle.
Quantum computation is as good as computation gets. If a quantum computer can't compute something, it's not computable. Period.