A CPU can add 32bit numbers in a nanosecond. In my quantum factoring estimate, I also used 32bit adders. They're a bit slower. Each superposed add costs TWO MILLION nanoseconds. Quantum factoring isn't fast because of fast operations; it's fast because it uses *few* operations.

A reminder that if your paper has open data containing stim circuits, I keep a list of those and you can be on the list. I know my 2025 list so far is incomplete: github.com/quantumlib/Sti…

A circuit compilation exercise. Show that MX+CX+T is a universal gateset by constructing an H gate. MX: X basis measurement CX: controlled NOT T: 45 degree rotation around Z axis (Classical feedback and ancillas is allowed.)

People complain that the "find the cycle in a linked list" interview question is contrived. I thought that too… til I was writing the loop analysis in stim's circuit-to-dem conversion, realized it was analogous, and the standard answer was way better than what I'd been writing.

Decrement operators in C: x-- is a post-decrement --x is a pre-decrement ~-x is a temporary decrement /s

My QIP2025 talk on magic state cultivation: youtube.com/watch?v=bbR0__… A later longer version of the talk, given at the Simons quantum colloquium: youtube.com/watch?v=gxztRT…

An identity that'd be great... if it did 2 things instead of 3. If one of the two CCZs on the right wasn't there, it'd yield an n-qubit incrementer with 4̶n̶→3n T gates. If the CCCZ wasn't there, it'd yield n single-shared-control Toffolis to be done with 4̶n̶→3n T gates.