After an exhausting few days in Las Vegas, I hope you are all home and recuperating1. Those of you who missed it, well, I hope you are able to keep your FOMO in check. I had originally planned to post this sooner, but Life intervened, so here we are.
Aside from the comically small venue and the general veneer of filth inherent to The Strip, I had a pretty good time. You may recall that I had some opinions about how to March Meeting effectively. Did I follow my own advice?
Yes, kinda.
I did spend multiple hours a few weeks beforehand reading through all of the DQI sessions and selecting sessions/abstracts of interest to put into a rough schedule. I didn’t want to plan my schedule too thoroughly, to allow for some serendipitous encounters.
As became apparent early in the week, the rooms for DQI sessions were far too small, so in some cases I chose to wander the halls rather than pack myself into some overflowing room.
By the end of the week, Silicon Valley Bank had failed. My understanding is that SVB was THE bank for tech startups, as well as a major regional bank, so it seems like a lot of startups are facing extreme uncertainty with respect to their money, and whether they’ll be able to pay their employees, vendors, etc. The extent of the damage in the Quantum Startup field is unclear to me2. I’m aware of at least a few managing to get their assets out of SVB before it went down.
MM 2023: Some Takeaways
I think many of the talks were recorded, so I am hopeful that I’ll be able to go back and watch the talks that I missed due to the crowd, or because they’re were simultaneously scheduled with something else I wanted to see. Here’s what I did get out of the meeting:
The Vitality of DQI: I missed the business meeting, but apparently DQI is Big and growing rapidly. Friends of mine who had done work in adjacent fields of condensed matter looked in on their old haunts and were disappointed to find that nothing had changed over 10 or 20 years. The lamented that the same people were still talking about the same problems and no apparent progress had been made. Conversely, DQI is imbued with a powerful vitality and palpable air of excitement. When I look across the fairly small slice of the field that I know and experienced at MM23, I can see enormous potential for major strides in almost every domain, from foundations to instrumentation.
Scaling Fridges: Scaling qubits is what makes the news, with roadmaps drawn up in every SPAC announcement and company website detailing the rough timeline on which we’ll have 1e2, 1e3, or 1e6 qubits. Of course, for those qubits to be at all useful we need bigger, better, and different dilution refrigerators with much better I/O and control. I saw a lot of serious work on flexible RF cabling for dil fridge interiors, photonic interconnects, and the properties of common cabling materials at and below 2 K3. This last talk was a Rigetti project, building on measurements done at NIST. They might even release the data for the benefit of the wider community.
Those of you who passed through the exhibit hall no doubt noticed that cryogenic everything basically ate the instrumentation exhibitions. Among them, a few of the newer dilution refrigerator manufacturers: Zero Point Cryogenics and Maybell Quantum. I’m pretty sure I saw a FormFactor booth4, but didn't stop in. BlueFors and Oxford Instruments were well represented, too. I am personally glad of the increased competition and attention on refrigeration systems. Even the best current systems kind of suck to use, and need substantial UI/UX improvements to be scalable in a way that doesn't require a literal army of PhD-having fridge nannies.
Novel Qubits: I stopped in at a novel qubit session devoted mostly to 0-pi (with a smattering of cos(2θ) and other stuff ), and found the state of the art to be much more advanced than the last time I checked in. Still, the situation appears to be quite challenging for the 0-pi, since the spectroscopy for those qubits looks a) totally insane and b) doesn’t seem to match well with the numerical calculations of the eigenspectra5. But that is why we have grad students, right? That session ended on a grim note, where a student6 from Andrew Houck’s lab presented numerical and experimental results that indicate the current method of initializing 0-pi qubits is plagued by huge matrix elements connecting the 0 state to much, much higher states (18, 25, 33, etc), ruining the desired fidelities. Her commandment was that the community needed to find a different and better way to operate the 0-pi qubit.
I heard tell of nice talks and great results in the fluxonium sessions, but all second hand. Fluxonium is practically mainstream by now, so I dunno if it should even be included in this section.
Readout: Most of this session was boring garbage7 about dispersive cQED readout, which totally sucks and is a plague on the field. The two interesting parts were a 2-part Google talk about how, in the wrong configuration, the resonator-transmon system could drive 0 —> 9 transitions in the transmon, which seems like it will limit the speed of qubit reset. These talks were fast, and a little hard to follow, but you can find the preprint on the arXiv (Measurement Induced State Transitions).
My favorite talk of this session presented a novel bolometric qubit readout using an array of SNS Josephson junctions8. What kind of crazed mind would want to use what is essentially a thermometer to read out a qubit? If you guessed, ‘The Finns’, you win. Their approach definitely works, with 90-something % fidelity and a comical 10 ms reset time for the first generation of devices. It seems like they believe switching their bolometer material to graphene could improve the reset time substantially. I actually kind of like this approach, partially because it seems to not require any components with gigantic on-chip footprint9. Plus, I crave novelty, so this hit all of the right buttons for me.
Misc: I’m coming around on the whole tantalum thing, again. Saw some pretty juicy (and out of context) pictures from a session I missed, in which some stripline resonators had lifetime and coherence > 1 ms.
You may have heard about the room-temp superconductivity talk and publication. There’s a prediction market for whether it will be replicated by 202510:
State of the Markets
Speaking of predictions, if you have any bright ideas for new things in the field for us to bet on, you should let me know, or better yet, start your own!
The Market is very bearish on D-Wave and Rigetti continuing to exist independently into 2024, which is not surprising now that D-Wave has also gotten a warning regarding possible de-listing.
One bright bit of news is that the market for 2-qubit gates with 99.9% fidelity is at roughly 2:1 in favor. If you think that’s wrong, you should sign up and bet!
Especially if you got yourself a Covid souvenir.
When I originally wrote this sentence, it was a little unclear whether depositors would be back stopped. As of this moment, I think all of the depositors are OK, but the bank itself is wiped out.
A common problem in engineering these systems is that very few people end up measuring material properties for common metals, alloys, and insulators below a few Kelvin. Unfortunately, it’s hard to engineer better systems without actually knowing the relevant properties of the components at mK!
Recall, they bought the dilution refrigerator division from Janus.
I mean this with all due respect.
Looking at you, resonators.
Of course it won’t.