Why Space-Time Needs “Magic” to Bend

For thirty years, physicists could build a universe out of quantum particles but couldn’t get it to do the one thing universes do. It just sat there. The space-time they constructed had shape, but no give, which means stars couldn’t dent it and planets couldn’t fall. A group of researchers has now traced that missing bendiness to a property they call magic, and the answer says something uncomfortable about what gravity actually is.

The Problem Hiding Inside a Working Idea

Start with what Einstein gave us, because it’s still the cleanest picture anyone has. General relativity says space-time is a fabric, and matter tells that fabric how to curve, and the curve tells matter how to move. John Wheeler put it in two sentences in 1973 and nobody has improved on them since.

“Space acts on matter, telling it how to move. In turn, matter reacts back on space, telling it how to curve.”

The picture breaks at a black hole. When a star collapses, its mass crushes into a point so dense that the fabric doesn’t just dent, it tears clean through. At that tear, Einstein’s equations stop returning sensible numbers. You need a different description, one that works when the curvature goes to infinity, and general relativity simply doesn’t have it.

So in the late 1990s, theorists tried a different angle. What if space-time wasn’t fundamental at all? What if it emerged from a collection of quantum particles arranged on a surface, the way a hologram stores a 3D scene on a flat sticker?

Juan Maldacena, Edward Witten, and others worked out that you could describe a whole exotic universe this way. Physicists call it the holographic principle, and it remains one of the most productive ideas in theoretical physics.

Entanglement Got Us Halfway

The holographic models needed a way to stitch the particles into something with structure. Entanglement did that job. It’s the quantum link that ties particles to one another, and it turns out to behave like the connective tissue of space itself.

Take a wormhole, a theoretical tunnel joining two distant regions. Holographically, that tunnel is two entangled sets of particles. Cut the entanglement threads one by one and the tunnel narrows. Cut the last one and the regions come apart entirely. Entanglement, in other words, builds the shape of space. That satisfies Wheeler’s first sentence. Space acts on matter.

Per a June 3 article in Quantum Magazine, Charles Cao, now at Virginia Tech, spent a month as a Caltech graduate student in 2016 working through a paper by Daniel Harlow that laid out the math for this. The trick Harlow used came from quantum computing.

A quantum error-correcting code spreads one qubit’s information across many qubits, so the data survives even when some qubits get lost. Harlow noticed that holography does the same thing. A single region of space gets encoded redundantly across many sets of entangled particles.

The codes had a flaw, and it was a serious one. Known as stabilizer codes, they split entanglement into two sealed compartments, one for space and one for matter, with no bridge between them. In quantum computing that perfect isolation is a feature, because you want your data shielded from the outside world. In holography it was fatal. The bowling ball sat on the mattress and made no dent.

As Bartek Czech of Tsinghua University put it, they knew how to build a space-time, but the space-time was inert. It didn’t do anything:

“When you design codes for quantum computing, you’re doing the same kind of thing that [holography] already did for you.”

The Magic Ingredient

Cao knew entanglement wasn’t the whole story. Ning Bao of Northeastern said it plainly. Something beyond entanglement had to be there:

“It was clear that something else beyond entanglement had to be there.”

He started tweaking existing codes. In 2020, with Brad Lackey, he modified a stabilizer code and found something. The space could now change, just not in response to matter. Not gravity yet, but movement where there had been none. The frustrating part was that they couldn’t say why the tweak worked.

The next year, Jason Pollack and collaborators figured out the missing piece. If you actually tried to run the tweaked code on a quantum computer, you’d need a specific operation called a T gate, which rotates a qubit. That detail mattered more than it looked. Cao had just been at a conference where everyone was talking about gates like these, because they’re what makes a quantum computer outrun a classical one.

Here’s the part worth slowing down on. Researchers used to think entanglement alone gave quantum computers their edge. Then they discovered classical algorithms could mimic certain entangling operations on an ordinary laptop.

So entanglement wasn’t the source of the advantage after all. In 2004, Alexei Kitaev and Sergey Bravyi pointed to a different class of operations, the non-Clifford gates, which include the T gate. When a program uses these, the equivalent classical simulation takes vastly longer to run. Kitaev and Bravyi named that extra complexity magic. The more non-Clifford gates a quantum state needs, the more magical it is.

Cao connected the threads. With Brian Swingle and Christopher White at Maryland, he studied particles equivalent to an anti-de Sitter space and found them highly magical. A few years later, working with Alioscia Hamma and building on work by Xi Dong, he showed what the magic was for. It gave space its springiness. Magic is connected to space’s ability to bend, which means magic is connected to gravity. “If you have one,” Bao said, “you always have the other.”

Putting the Pieces Together

By early 2026, Cao, John Preskill, and others had everything they needed. Magic makes space bend. Codes get magic from non-Clifford gates. So they built a successor to the stabilizer codes, packed with non-Clifford gates, and the magic let the entanglement for space and the entanglement for matter affect each other. The two compartments, sealed for a decade, finally leaked into one another the way they had to.

That leak is exactly the point. Cynthia Keeler of Arizona State, who wasn’t on the work, noted that quantum gravity shouldn’t have a fixed background. “It should fluctuate,” she said. A magical code lets it.

Preskill summed up the lesson in one line:

“Without magic, things are a little too simple. And quantum space-time isn’t quite that simple.”

What the Result Actually Says, and What It Doesn’t

I want to be careful here, because the framing invites overreach. Cao himself joked at the American Physical Society summit in Denver that he was the only speaker who wasn’t really studying quantum gravity. His code is still extremely general. It doesn’t describe the space we live in, doesn’t reproduce Einstein’s specific equations, and doesn’t even include time. “This gets you a precursor of gravity,” he said. “You satisfy one of the necessary conditions. Right now, we are at step 0.5 of 5.”

That’s an honest number, and it should anchor how anyone reads this. What’s been shown isn’t a theory of quantum gravity. It’s a structural requirement that any such theory will have to meet. If you want your encoded space to bend, your code has to be magical.

The deeper claim is the one that lingers. Non-magical codes produce inert, gravity-free space precisely because they protect their information perfectly. Cao’s work shows gravity emerges from the mixing of encoded information, which means the encoding can’t be perfect. It has to be approximate, so some of what happens inside the space-time can’t be fully recovered by measuring a subset of particles.

A quantum engineer would call that a poorly written code. Czech called it “the reason Newton’s apple fell on him.” Cao finds that appealing, and his reasoning is worth keeping. Error correction is a human pursuit. He sees no reason gravity should share our preference for perfection.

So the two defining features of quantum mechanics, entanglement, and magic, line up with the two defining features of space, its shape and its flexibility. If that correspondence holds up, space isn’t a classical fabric that happens to contain quantum stuff. It’s one of the most quantum things there is.

The cautionary note is the same one Cao keeps repeating. Step 0.5 of 5 leaves a long road, and the history of this field is full of structures that looked decisive until someone tried to make them describe our universe. Brian Swingle thinks simulating high-magic gravity will demand an actual quantum computer, because no classical machine can fake it.

“If we need high magic, then we intrinsically need a quantum computer, (…) because there’s no other way, in general, to get at that kind of question.”

If he’s right, the next steps don’t just need new math. They need hardware that doesn’t quite exist yet. That’s the honest place this stands. A clean correspondence, a working proof of concept, and four and a half steps still to climb.