sorted by: new top controversial old
Haley Stevens doesn’t regret endorsement flip-flop in Michigan Senate race in c/news@lemmy.world
Haley Stevens doesn’t regret endorsement flip-flop in Michigan Senate race in c/news@lemmy.world
Hakeem Jeffries Says Impeachment and Getting Even Aren’t Priorities in c/politics@lemmy.world
Proton users cancel subscriptions following CEO Andy Yen’s “loser thinking” remarks in c/technology@lemmy.world
[–] 10 points 3 days ago (2 children)

Your first sentence is basically impossible. When reviewing code, your insight is always inferior to the person who wrote it. Often my code reviews consist of questioning the programmer on why they've done what they have, because I need their perspective as well.

You can't do that with an LLM because it doesn't really know why it did things.

And testing properly....? Nobody does that. LLM or human.

  • permalink
  • fedilink
  • source
  • parent
  • context
    • Proton users cancel subscriptions following CEO Andy Yen’s “loser thinking” remarks in c/technology@lemmy.world
    Psssst in c/politicalmemes@lemmy.ca
    Psssst in c/politicalmemes@lemmy.ca
    [–] 27 points 3 days ago (9 children)

    God the cynicism in this thread is depressing.

    You know why your vote doesn't matter? ....because you talk yourself out of getting of your lazy ass and using your vote.

  • permalink
  • fedilink
  • source
    •
    Haley Stevens’ chief of staff and others resign amid Mike Rogers endorsement in c/politics@lemmy.world
    Did you hear Stevens endorsed Rogers? in c/politicalmemes@lemmy.ca
    Starlink Reportedly Tried to Sell Ukraine a More Expensive Internet Plan After Musk Cut Service in c/technology@lemmy.world
    People who use AI are actual NPCs in c/showerthoughts@lemmy.world
    Why Is the Media Ignoring the Stunning New Trump-Epstein Allegations? in c/politics@lemmy.world
    [–] 1 point 5 days ago

    ...because allegations don't get it done.

    We need him bang-to-rights. No wiggle room. Conclusive evidence. He's the most powerful man in the world and he shrugs off allegations.

  • permalink
  • fedilink
  • source
    •
    104
    LLM mathematical proof exploits theorem proover bugs (infosec.exchange)
     

    Okay, we have a new contender for Most AI Thing to Ever Happen

    1. July 25th: someone messes around with an LLM and posts a proof of the Collatz conjecture that does, in fact, verify in the theorem prover. (The AI use is not disclosed on the github page) https:// github.com/xrchz/CollatzLean

    2. July 26th: several serious bugs are posted in the theorem provers, that in principle could allow a false statement to be "proven" true. They're serious, yes, but no need for panic, because you're not going to blunder into accidentally exploiting the bugs while writing a proof, probably. https:// github.com/leanprover/lean-kernel-arena/pull/81

    3. July 28th: someone who was right to be very skeptical of the Collatz proof, and had the expertise to study it with a fine-toothed comb, discovered it was exploiting a bug https:// github.com/leanprover/lean4/issues/14576

    4. The "proof" turns out to be exploiting multiple similar but distinct bugs to pass different solver variants!

    5. the human who posted the proof acknowledges the AI use and claims they did not knowingly point it towards the bugs it exploited. https:// leanprover.zulipchat.com/#narrow/channel/270676-lean4/topic/Counterexample.20to.20the.20Lean.20Conjecture.20.28Soundness.20Bug.29/near/613135216

    Note that the proof was posted shortly before the related bug reports were posted. It is an open question if the AI found people discussing the bugs shortly before they were formally posted and "decided" to exploit them, if the AI "knew about it" as a learned strategy from the training stage (putting every single "proof" it's ever made and ever will make into profound doubt), or if it's recently been repeatedly blundering into it by sheer stupidity and that's how people noticed the bug at about the same time.

    Theorem provers aren't magic, and have bugs just like all other programs. They are tools to help us double-check our reasoning. When you skip the reasoning and ask an AI to "prove" something for you that's over your head, you're entering an adversarial pact with the monkey-pawed Devil of Customer Satisfaction.

    view more: next ›