(terrytao.wordpress.com)
It seems we’re making tremendous progress on the back of existing abstractions and concepts with the brute force ability to generate language and verify proofs.
For example, Tao says about the Riemann proof:
> appears to use the full power of modern developments on the Langlands program and arithmetic geometry
What I wonder: how do we create new abstractions and concepts? Like whatever “Langlands program” is? That then feedback into the system, and allow for even more sophisticated mathematics.
Are we assuming AI will do this? Or does this still need to be something at least agreed-upon by a mathematics community - in the same way programmers talk about a “linked list” without having an hour long conversation to define it axiomatically? And without that social aspect, without breaking down long proofs into Lego block common abstractions, math may plateau? In part because we lose the ability to ask meaningful questions?
Compare Elliot Glazer a month-ish ago:
> I occasionally talk to experts on all the MP's. Their rough summaries on what is needed to solve them usually go:
> Hodge: not a clue
> BSD: not a clue
> RH: not a clue
> P vs NP: not a clue
> Yang-Mills: "Well..."
https://news.ycombinator.com/item?id=50252673
So if we're in "not a clue" territory, maybe that's that.
I fear that it gets muddled again with e.g. the concurrent human announcement and that in retrospect RH turns out to be easier than expected, like many other problems before.
Math is hard, but how hard math is is also hard...
Langlands is interesting because it is most conjectural (and not completely clear what the precise statements are -- hence "program" and not "conjecture"), but proving special cases turn out to be tremendously powerful. One special case implied Fermat's Last Theorem, for example.
That being said, its probably going something akin to how experimental and theoretical physics have an interplay with each other.
Can we develop more and more all encompassing abstractions (theories) that explain and verify experimental results; translates to, can we come up with more and more theorems/abstractions that explain formal proofs (and perhaps provide explanations for other known proofs)? Probably yes, and likely will also be accelerated by AI.
How this fits into the discipline of mathematics is something we are currently finding out.
Yes.
The impact of the RH proof (assuming nothing torpedoes it) is any of the new machinery involved in the proof.
That said Just because something is solved doesn't mean we don't work on making the proof simplified and easier to understand.
Compared to the volumes of pure math proofs which are useful?
>so the significance is fairly lessened until the result is boiled down and understood
Significant enough to make a blog post about it the same day it occurs, but not significant enough to mention the significance of it within the blog post itself lol
I imagine Tao is just being careful and conservative with what he's saying it this point. His previous posts discussing AI in mathematics probably garnered way more attention (including negative attention) than he expected, so it'd be fair if he kept his perspective a bit more neutral and muted now. But it's pretty obtuse to assume that Tao doesn't consider this to be one of the most important events in his life. You don't need a PhD in mathematics to appreciate the significance of this.
Room temperature superconductors are still a ways off imo.
I'm starting studies to be an electrician course next year. I want to work in an industry where I'm expected to do high quality work, made to last decades. I want to be as far away as possible from the slopfest.
At a very baser level being a software engineer makes me feel like an important man in society being an electrician would not give me the same feeling. Obviously this is totally arbitrary, but I think the vast majority of software engineers think this way…
9-10 years ago I was in a smaller structure and basically doing close to SOTA work in the field of NLP: contributions to `spacy`, some of the best word2vec models models for our language (french) and good english ones, at least models that crushed other word2vec models of the time on our large corporate benchmarks (both french and english), and were on par (english) and sometimes better (french) than the official benchmarks of the time. Mostly through minor model tweaks but major data sanitation and preparation work. And then custom transformer-based architecture for other NLP tasks that were also better than all off-the-shelf NLP solutions of the time. And even on my free time I could do some cool ML/NLP stuff no one else was really doing.
Now I am in a larger structure with more resources, yet I know that both in terms or resources and honestly to a degree in terms of skill I feel left out of the current NLP/AI SOTA landscape.
It sounds like you were already doing statistical NLP (the thing that has been clobbering everything rule-based in performance) but have seen that field itself start to require much greater resources than what you were working with.
It's 4CT on speed.
Is mathematics simply a list of true statements? Or is there some intrinsically human aspect?
What a time to be alive!
For now, same old stuff. Go to work, keep up with the latest changes, get the kids to school, plan to send them to college. Hard to say if that's the right move at the moment, but what else is there?
I basically just try over in the moment if each day and hope we have many left.
When all this starts to extend to the physical world
Maybe GRH was as well, I haven't seen much about how these were put together. But, it's a curious difference in approach if not.
I find that a lot of the times that people see the post from Terrys site they attribute the authorship to him when he does a significant amount of reposting.