Q-Factor CEO argues a quantum computer can work and still be a dead end

[NatureBoy_Dev]

Calcalist's tech site has an opinion piece from Dr. Guy Raz, CEO, co-founder and CTO of the Israeli quantum startup Q-Factor, and it makes a point that does not get said often enough. His argument is that a quantum computer can work perfectly well at a small scale and still be a dead end. Qubit counts get the headlines, but Raz says the real question is whether everything around the qubits can scale with them. As he puts it, useful quantum computing will depend on whether control, calibration, readout and error correction can remain practical as machines grow by orders of magnitude

He walks through where things break. A control scheme that is fine at 1,000 qubits may become unworkable at 100,000, simply because of the wiring, electronics and tuning involved. Readout forces a trade-off between accuracy and speed, and error correction can eat so many resources that little is left for useful computation. Modular designs, where several smaller machines are linked together, bring their own interconnect bottlenecks

Raz points to recent milestones as proof that the individual parts already work. He mentions the Caltech array of 6,100 atoms, which we discussed here, and a Harvard-led team's 448 qubit system. His view is that the components are no longer the hard part. Getting them all to work together at a much bigger scale is

Naturally, there is a pitch in here too. Q-Factor builds neutral atom systems, which Raz sees as the most promising platform, and the company is aiming for million qubit machines without splitting them into modules. He mentions a refocusing method to deal with atoms heating up and the use of auxiliary atoms to boost readout signals. Those are company claims, so they should be read with that in mind until there is published data to back them up

Even so, the core argument stands up whoever is making it. We have seen plenty of announcements this year about qubit numbers, from neutral atoms to superconducting chips and trapped ions. Very few of them say much about calibration time or how many control lines a million qubit machine would need. Is engineering scale the real bottleneck now? And which platform do you think copes best with it?

A quantum computer can work and still be a dead end | CTechwww.calcalistech.com

[PhotonBurst48]

Atom heating is a known problem with optical tweezers. Atoms get knocked about and lost, and you have to reload them. Any fix for that would be a big deal, but I would want to see the paper first