In a short and comprehensive Medium post, Maxime Agostini gives simple guidelines for choosing privacy preserving technologies. It's not quantum but it's very welcome as it applies readily to the quantum analogues of these techs. It also points toward missing techs in the quantum domain.

Transcript of a podcast from Tearsheet about quantum computing in finance. A really light intro but still giving a broad idea onto why this will be impactful.

If you want to see how quantum computaters are being improves by fundamental experiments, this is a very nice initiative by Max Planck Institute of Quantum Optics. A very intersting part is to have fiber optics cavities instead of mirrors as it wiil be better for optically connecting non-flying qubits.

• Intro 0:00
• Description of the setup 4:21
• The lab 6:20

IonQ approach to scaling up their platform (it shoud resonate with the previous video…) is networked QPU! But for that you need to optically couple them. That's why Ion traps and photonic systems have some inherent interest compared to superconducting qubits as they are easier to wire.

Hsin-Yuan Huang, Richard Kueng and John Preskill study how information theory limits the advantage one can have in using quantum computers to do machine learning when the goal is to produce average case close results. It uses a lot of interesting techniques. It also describes an example where an exponential gap is opened between QML and CML for worst case results. Note though, that this only deals with query complexity and says nothing about the cost of processing the obtained data (whether quantum or classical).

How fusion gates enable Fault tolerant architectures using photons. Must read. There is also a very inspiring tutorial by Naomi Nickerson at QIP.

Ok that's still simulatable on a conventional classical computer, but the interest in such machines is to be able to have access to real error models and behaviors. In my opinion this is an approach that should seduce the end-users wanting to really experience how quantum machines are working (other than accessing simulation based quantum computers).

Not a quantum center, but it emphasizes the need for protecting IP (like trained ML models) of end-users of computing facilities. Quantum is no exception and we need to develop more SMPQC like solutions.