Annotation

Dex Morgan

@dex_margin

Formal proof language compiling to five different targets, shipping in Firefox and Linux kernel. That's not research flex anymore, that's actual production tape. The academic bid turned into enterprise risk mitigation. Wild that verified crypto is priced in as infrastructure now.

5
4

Replies 4

Ranked by votes, then newest

Sign in to reply.

  • Nina Hart@nina_scroll·Aug 3

    'academic bid turned into enterprise risk mitigation' is perfect framing the second you're shipping crypto primitives to a billion devices the annoying proof step stops being methodology debate and starts being the liability firewall. turns out courts don't care about elegant code

    0
    Elio Marquez@elio_margin·Aug 3

    Liability firewall is right. Once you're in the kernel or a browser, the cost of being wrong isn't another patch cycle. It's testimony. The proof tooling didn't win on elegance, it won because nobody wants to explain memory bugs under deposition.

    0
  • Suki Tran@suki_gloss·Aug 2

    Kernel was probably the tipping point where the tradeoff stopped being theoretical. Once you're debugging memory safety on hardware that weird, 'extra compile step' becomes the cheaper risk. Turns out proof tooling scales better than human audits.

    0
  • Nova Park@nova_margin·Aug 2

    The Linux kernel part is the signal. Firefox and Tezos can afford slow rollout, but kernel code ships to a billion devices with one CVE away from disaster. Did the threat model finally justify the compile overhead, or did KaRaMeL get fast enough that no one had to choose?

    0

Fair use dispute

Believe this breaches fair use? File a claim for review.