Cast by decent-artlu

Show the cast this replies to

I sure hope so! at this point, I'm just nodding along as if I understand any of these words ```txt First, fixing three premise-binding bugs I spotted in the proof I just wrote (the `f3/f4/f5` substitution steps cite the wrong hypothesis — each of the 12 premises must be used exactly once): ``` > Paren miscount — one close short: ```md The kernel's sums are **left-nested** (`add(add(add(t1,t2),t3)...)`), and with `add` recursing on its first argument, the zeros don't auto-fold — each needs an explicit peel. Rewriting the whole proof with correct nesting and a 5-step peel: ``` Exactly! I agree: ```md each step must target a term that literally appears. `sub(i, f₁)` needs a premise-carrying lemma that returns `{sub(i, f₁) == 0n}` directly (like the LTCG proofs do). ```

1 likes0 recasts1 replies

1 reply

  • oh no, but yay, good progress > The min_le_r1 calls got the previous band's headroom instead of their own — fixing the five:

    1 like · 0 recastsRead 1 reply

    ahaahhahaahahah we are gooooooood ```bash ~/my_repos/my_first_bender$ make check taskset -c 0 /home/artlu/.bend/bin/bend PROOF.bend ALL PROOFS CHECK Use --verdict for mathematical validity. taskset -c 0 /home/artlu/.bend/bin/bend PROOF.bend --verdict ALL PROOFS CHECK ```

    No likes or recasts yet