Intel's 49-Qubit Chip Shoots for Quantum Supremacy | Hacker News Reader

Intel's 49-Qubit Chip Shoots for Quantum Supremacy

spectrum.ieee.org

54 pointsamaks18 comments

Did they give any numbers on the quality of the qubits? For example, IBM's 20 qubit chip has 2-qubit operations with error rates on the order of 5% [1] (some pairs of qubits are better, some are worse). Quantum supremacy experiments require thousands of operations (tens of layers of parallel operations). Qubits with even a 1% error rate per operation just won't cut it.

1: https://youtu.be/T-8uuq7Izl8?t=26m58s "Experimental quantum computing at IBM" [26:58]

(Disclosure: I work on Google's quantum team.)

> Did they give any numbers on the quality of the qubits?

No they didn't. I've also had problems finding the benchmarks for their 17 qubit processor. I would appreciate if someone could link them to me.

I guess taking every branch by default avoids the specter of branch prediction vulnerabilities?

(This is me failing my saving throw vs urge to make terrible nerd jokes.)

What makes one quantum computer superior to another? Not number of qubits, I guess? IBM has a 50-qubit machine.

IBM talks about a metric called "quantum volume" that combines number of qubits with error rate to summarise useful computing performance. See https://dal.objectstorage.open.softlayer.com/v1/AUTH_039c3bf...

(disclosure: IBMer, but very different area)

Same as classical computers, what makes Intel i7 superior to AMD Ryzen 7?

One that can perform a computation task fastest with least energy would clearly be superior to another.

I assume these are massive like D-Wave's systems? https://youtu.be/60OkanvToFI?t=392

What is it means for humanity?

Better drugs/more complex medicines should be possible. Chemical interactions are hard to do on classical computers (operating on 5D+ arrays), true quantum computers should be able to help with those computations.

Do you have easy material for me to understand how quantum computing enables that kind of things? And does it use clasic numerical method computing or using different kind of mathematical approach? Actually I'm not sure whether my question makes sense or not

I enjoyed reading this paper:

https://arxiv.org/pdf/0708.0261.pdf

Is it vulnerable to meltdown?

maybe if you run out of liquid helium ;)

No.

Maybe if it was running real world applications with system and user space. Right now these all run prototype code

Breaking Bitcoin With a Quantum Computer

http://fortune.com/2018/01/06/breaking-bitcoin-cybersaturday...

Quantum Resistant Ledger

https://theqrl.org/

It is unknown if BQP >= NP or NP-complete

NP-completeness is a property of problems, while BQP is a complexity class (set of problems). If any NP-complete problem is in BQP, then NP is a subset of BQP.

But it's not even known whether BQP > P, so we could have P = BQP < NP.

Reply on news.ycombinator.com