This has to be my favourite post so far from the IonQ bros.

Having looked at the IonQ material this week has resulted in my feed being flooded by the “gang” on public ionQ investors. My god!!

https://t.co/3P09JInNTq

For those who have forgotten, IonQ has been touting this 13:1 error correction overhead for years. It came from a 2009 paper from Chris, Ken and others demonstrating a [[9,1,3]] Bacon-Shor code, and it appears like IonQ never bothered to understand it. Just ran with the number

Here is a link to the D-Wave short sellers report. https://t.co/v3IRmxRRec

Senator: what are we going to get in 2027? Response: the worlds first fault-tolerant quantum computer Senator: and how many qubits is it going to have? Response: I’ll take that on notice (parliamentary speak for I don’t know, I’ll get back to you)

https://t.co/OWzzY4NBH3

https://t.co/07ZIuA1bpL

More from Australia Senate Estimates from back in December on the timeline of the PsiQuantum investment.

In keeping with the Australian tradition of thievery, we are stealing from the @unitaryfdn and launching our own Micro and MesoGrant programs in Quantum Algorithms, Software and Theory. Applications close June 30th.

So the claim the IDQ did it years ago is purely speculation. Nobody has any idea what they’re getting from the chipsets. That’s the point.

Certified quantum random number generators aren’t the same as QRNGs. This is a RNG and certification protocol using random circuit sampling first derived by aaronson. The QRNG from ID quantique or quintessence use quantum protocols for RNG, but you can’t certify randomness.

Yea, cause it’s the difficult in creating topological qubits that’s caused all the problems.

Almost any other organised effort caught out in the repeated issues of hype, borderline misconduct and allegedly actual misconduct should be looking for references to be uber drivers. And the next iteration isn’t even imaginative. Buts it’s @MSFTResearch, so all is forgiven again

It’s the arrogance i find most galling. The repeated appeals to the same playbook and acting that it’s everyone else’s problem that we just don’t take them at their word (this time) and defer to their superior expertise (this time).

Assuming literally everything else in the system is perfect except for the finite dark count rate and heralded FWHM sources.

Am I missing something? or do these numbers suggest that even in the context of a pseudo-deterministic single photon source there is still significant engineering still needed in either SNSPD detectors or in the efficiency of the downconversion sources?

And this upper bound is at the first step (single photon production) of a complex multiplexing stack needed to generate the resource states expected from the RSG's in the architecture.

Hence, even in this scenario, the maximum probability of success for this multiplexed source will NEVER exceed 50%. This gets worse if more sources are multiplexed to increase the number of photons produced or if the dark count rate is worse than 10^(-4).

Then multiplexing on the order of 10^(4) sources to guarantee, on average, at least one photon is produced per pump cycle, you are also guarenteed, on average, a dark count occurring somewhere at the same time as the actual herald signal from the source where a photon is produced