Thanks for sharing this! I agree with some of the things namely
effect on WMDs in the near future dodgy
Scepticism about near-term quantum computing: I worked on NISQ algorithms (basically near-term quantum computing—Noisy Intermediate Scale Quantum) for a little bit, and agree with your impression that there have been negative results, with quite a bit of pessimism in the community about near-term prospects (and maybe a “quantum winter”, a hype correction like you mentioned). It’s why I didn’t include quantum machine learning and cautioned about “quantum jobs”. However, on the other side of the spectrum, my impression is that some people are really positive on recent hardware progress and advancement with error correction (towards fault-tolerant quantum computing). On this side, I have talked to someone who believes there could be some kind of GPT-moment in the next decade when a “killer app” is demonstrated. Would be curious to hear your thoughts on this.
I tried to stress the early state of the technology and large uncertainty, but your comment makes me realize that I should have front-loaded more heavily that nearly all the things I mentioned are long-term concerns.
On the “why now rather than later”, I partly disagree:
Once (if ever) dangerous applications are around the corner, we will also have economically meaningful applications at least equally close. I think that would make it much harder to put in place guardrails.
I am not sure we will have a lot of advance warning (although I agree that most likely we will): I read a good commentary which agrees with you, but I think it neglects that some approaches to quantum computing (e.g. photonics) could face very different scaling challenges than those currently dominant (i.e. superconducting). (There is a DARPA project “Underexplored Systems for Utility-Scale Quantum Computing” which looked at this.)
I think for the bifurcation/great power competition part (which I personally think deserves most attention), there is a strong case that action now would be most effective in addressing them. I will put up a separate post on this.
In the case of cryptography the need to act is clearly now (see the section in appendix), although I do agree with you that a lot of actors are already incentivized to do so (even though I personally believe more should be done).
Finally, on the “lumping”, it is motivated by the possible future of quantum networking for distributed quantum sensing and interconnecting quantum computers with quantum sensors. Maybe I should have made more explicit in the post that right now, quantum sensors are, as you said, just a slightly better way of doing things we can already do.
Thanks! I partly agree on the cryptography part, I definitely don’t think that it would be effective to fund efforts towards transitioning to quantum-safe cryptography. Nonetheless, I do believe that working on the quantum-safe transition could be a good career option for someone who already has expertise in the field (feel free to push back on this).
I think in the future, it could be promising to work on capacity building and helping places that lack behind catch up to new cryptography standards. Especially, if real negative consequences start to materialize (which maybe won’t primarily be quantum-enabled hacks but rather compatibility challenges).
Finally, just to clarify, I think that cryptography is the most clear problem, but probably not the most important one from an EA perspective.