4 pointsby KenographerPrim3 hours ago1 comment
  • ohadkr2 hours ago
    The benchmark claims are impressive, but publishing a reproducible setup or more of the source would make them much easier to evaluate.
    • KenographerPriman hour ago
      Thank you so much for your return.

      About your advice, what do you mean?

      The whole pipeline (error correction, mapper, router, optimizer) is available with precise instructions.

      A black-hole python runner is available alongside the black-hole circuit published paper linked in the README.

      Benchpress is easily available from the IBM Benchpress Git repo.

      And IBM QPUs can be used for free (they offer 10 min of free QPU usage with a new quantum account)

      So the point is, to me, everything is available and reproducible. You could run the full Hayden-Preskill black hole scramble circuit in 10 min from now.

      I'm not dismissing what you're saying, I'm badly trying to explain that I'm doing my best to be transparent and give all the possible tools and data. So I'm not sure what else you want me to do...