MaximusIdeal

we'll see how things go; not necessarily a math account (I am not a math object for you to use)

[MaximusIdeal @maximusideal]

@littmath This is why it's important to emphasize the value of having multiple proofs of theorems. It never fails to give me pause that something truly deep is happening when a result has widely different approaches (e.g. quadratic reciprocity, fundamental theorem of algebra, Gauss-Bonnet)

[MaximusIdeal @maximusideal]

@matthew_petti That's completely fair. I was aware that was likely the issue, but I'm much more inclined to separate the hypothetical tech from what we do with it. Maybe I'm being naive however.

[MaximusIdeal @maximusideal]

This occurred to me independently: There are psychiatric disorders we would want treated/cured/solved, but are personality disorders even solvable in principle? Does such a question make sense? Anyways, I don't see what's dystopic here so long as we embrace neurodivergence.

[Matthew Petti @matthew_petti]

underrated part of Anthropic’s dystopian vision

[MaximusIdeal @maximusideal]

Oh my God, in case you thought the discussion couldn't get any more negative iq...

[Другая планета @nihilunbounded]

Mark roots of unities on boundary of disk, mark three points z1, z2, z3 on boundary of Jordan domain.

For any x in interior of disk, start Brownian motion from x and calculate probability that it escapes boundary of disk through the arcs [1, w], [w, w²], [w², 1] respectively. Call the probabilities p1, p2, p3. Note p1 + p2 + p3 = 1.

Find a point y in interior of Jordan domain such that the probability that a Brownian motion from y escapes the domain through the arcs [z1, z2], [z2, z3], [z3, z1] are *exactly* p1, p2, p3 respectively.

Map x to y. This is the Riemann map. (It is conformal because Brownian motions are conformally invariant)

[MaximusIdeal @maximusideal]

@shlevy People are way too quick to jump to their conclusion and not even try to find a way to second guess their opinion, so they make confident sounding declarations that sound nice

[MaximusIdeal @maximusideal]

@prof_g Basically. I hate both sides of this conversation, but here's my brilliant strategy OAI should have done instead. Everyone would have walked away a little happier. https://t.co/1sdqSjRNRa

[MaximusIdeal @maximusideal]

Here's how OAI could have done PR better:
>announce new model, too scary to release like mythos
>call it 'prometheus' or something
>say they are giving special access only to select mathematicians
>make a pretentious name project ars magna
>ppl kinda collaborate more, less anger

[MaximusIdeal @maximusideal]

@generatorman_ai they should have just copied anthropic (or listened to me): https://t.co/1sdqSjSlGI

[MaximusIdeal @maximusideal]

Here's how OAI could have done PR better:
>announce new model, too scary to release like mythos
>call it 'prometheus' or something
>say they are giving special access only to select mathematicians
>make a pretentious name project ars magna
>ppl kinda collaborate more, less anger

[MaximusIdeal @maximusideal]

My idea would have also naturally led to more transparency: https://t.co/1Iy71j8SGJ

[Gary Marcus @GaryMarcus]

The real news here isn’t the result; it’s what we were not told.

1. AI once tried to be a science. Now we get stuff like the completely vague report from OpenAI below, and a lot of ignorant questions from people who don’t know how to think critically.

“Same procedure”? “using an unreleased model”?

This would never pass peer review.

We don’t know what the procedure was.

We know nothing about the architecture (e.g., were proofs generated in one shot, and then verified by Lean? was there an iterative process?).

We know nothing about the failure rate. We know nothing about the training/post training/data agumentation.

2. As a result we have zero idea of how generalizable the result is outside math.

3. A lot of X has been reduced to an ignorant cheering section that applauds without knowing what it is applauding or what it might mean — without ever asking basic scientific questions.

The new system could be a legitimate step towards AGI or just a clever leveraging of Lean and synthetic data in a verifiable domain with no generality whatsoever; from this report we can tell almost nothing.

[MaximusIdeal @maximusideal]

Here's how OAI could have done PR better:
>announce new model, too scary to release like mythos
>call it 'prometheus' or something
>say they are giving special access only to select mathematicians
>make a pretentious name project ars magna
>ppl kinda collaborate more, less anger

[MaximusIdeal @maximusideal]

My pt (which never gets across) is that OAI should have given an opportunity for mathematicians to be proactive rather than reactive to a dump of proofs they announced out of nowhere. This doesn't solve everything but they should have copied anthropic's marketing stunts

[MaximusIdeal @maximusideal]

@mathandcobb I'm very split on this situation, but I think OpenAI could have cut down on half of the controversy simply by taking a page out of Anthropic's brilliant marketing stunts. It wouldn't have solved everything but it'd have been better for everyone. See here: https://t.co/1sdqSjRNRa

[MaximusIdeal @maximusideal]

Here's how OAI could have done PR better:
>announce new model, too scary to release like mythos
>call it 'prometheus' or something
>say they are giving special access only to select mathematicians
>make a pretentious name project ars magna
>ppl kinda collaborate more, less anger

[MaximusIdeal @maximusideal]

To the 2 people that saw and appreciated my tweets, I regret to say that it looks like solving is not solved https://t.co/FiNP7OO9LF

[Thomas Bloom @thomasfbloom]

With, reportedly, 372 solved problems out of around 4000 tried, I think my earlier heuristic remains roughly accurate.

Maths is very much not solved! (Though having a magical proof producing machine that works 10% of the time does change it a lot...)

[MaximusIdeal @maximusideal]

Huh. Ok maybe that guy everyone made fun of that said problem solving was solved was right...

[MaximusIdeal @maximusideal]

@RBehiel @cgtwts I agree this is likely low-hanging fruit humans missed, and others' replies about e.g. quasi-RH isn't a good counterargument, but when you say "Math will be back to normal, just with more powerful tools" that seems like an oxymoron. More powerful tools usually means acceleration.

[MaximusIdeal @maximusideal]

@csaez_math I would have called it "the square root of six times the Basel constant."

[MaximusIdeal @maximusideal]

@SYNESTHEIZURE The word "magic" itself is an incoherent and not-well-defined concept to begin with. It's just a vibe people use either to pump something up or tear something down, and its entire content is in its connotation.

[MaximusIdeal @maximusideal]

@KylePomykala This is already the argument (not being derogatory or disrespectful b/c it's true but it likely won't last as a human-only thing)

[MaximusIdeal @maximusideal]

@ArtemisConsort I like this, but where will we get competent teachers to do this? This is the bottleneck for proposals like this. (I know there are programs to be at university while in HS -- I took college ODEs at uni while in HS -- but I'm not sure if this is what you were thinking.)