Mathematician, CMU. https://xitter.twitterapi.xyz/t.co/pOCv2Epa3t
Pittsburgh, PA, USA
Joined March 2020
- Tweets10K
- Following79
- Followers17.8K
- Likes16.3K
Judges/others have sometimes used historical examples of this to excuse gerrymandering as self-limiting, but I think that argument is weak.
True, sometimes people gerrymander badly or with poor foresight.
But the power to draw the lines with impunity grants a lot of free power.
Hindsight is 20/20: NC Republicans could have easily drawn a 9R-5D or even a 10R-4D map that would have withstood a blue tsunami (by drawing one more safe Dem seat in the Triad). Instead, by trying for a 11R-3D map, they're on track to end up in a 11D-3R hole.
3
795
A thought experiment for those that think
LLMs=in 10yrs we're colonizing stars and living forever
What if in 1990 we'd suddenly lost nearly all physical technology+infrastructure, but kept our (advanced!) knowledge of it?
How long would it take to recover?
1/3
This is so wild, the 20x claude max plan was never 20x pro weekly usage, the 20x factor was only for 5h usage; weekly usage was only 10x, weekly usage/$ was same as on the (5x) $100 plan. The advertised language was:
"Save 50%"
and
"Choose 5x or 20x more usage than Pro"
I lead product for Claude subscriptions. Today we're renaming our Max plans, and I want to explain why.
When we launched our top plan in April 2025, it had a single limit: 5 hours. It gave you 20x Pro's limit, so we called it Max 20x. When we added weekly limits to Max 20x, we set them to 10x Pro's (the 5-hour limit stayed at 20x).
I get that some new users assumed 20x applied to both limits. We could have done better here, and I'm sorry.
The Ultra plan keeps the weekly limit (10x Pro), but we're removing the 5-hour limit so you can keep building without getting blocked mid-session. This starts today and rolls out in stages. To help prevent fraud and abuse, a temporary limit applies until we've verified your account. You'll get an email when your limit is removed.
We've got more cooking and can't wait to share it. Thanks for building with Claude 🙏
1
8
2,521
Turns out the class action was filed over the summer claiming the naming was misleading:
plainsite.org/courts/califor…
Notably, the suit doesn't claim to know the actual multiplier (because it seems not to have been disclosed *anywhere*!), but to have estimated it to be << 20x.
1
362
WOAH!
Attention to anyone working on formal program verification, etc etc.:
Starting soon, anybody using "Daybreak Blue" access in their Chatgpt personal accounts will have their usage uploaded for model training...
***even if the model training setting is turned off!***
It's hard to imagine how this is supposed to work...
xitter.twitterapi.xyz/mdp_sec/status/2116262…
The chain below should serve as a reminder that most intellectual endeavors are not clean AI targets the way pure math has been; this presumably has outsized effects on AI capabilities because it affects not just immediate performance but the feasibility of training, etc.
Everyone knows my priors, and they are not particularly kind. Nonetheless, this should not be taken too seriously. AI is not good at making this kind of assessment, and we don’t know what the prompt was.
Some of the “errors” are actually criticisms raised in papers published after the ones analyzed, which is the normal process of science.
This is not the right way to criticize each other.
8
1,447
The bear case for huge AI impact on life sciences would be something like: chance and observation matter a lot so that the main issue is:
progress ∝ (# pertri dishes currently deployed)
You might mess with the constant a bit but overall funding/basic research may matter most.
1
2
915
I go back and forth on the significance of the fall of ETH relative to some of the other recent bombshells, but I do think it is in some sense the one that seems most like it may hint at a huge paradigm shift for its field.
I'm surprised discussion of AI safety/regulation doesn't include more calls for a regulatory requirement that all reasoning tokens be made available for inspection to the end user.
Arguments against are easy to imagine, but strong arguments in favor can also be made...
1/2
1
4
1,389
The impression I get from the AI maximalists is that if we're ever contacted by aliens, we should expect that we'll actually be talking to their version of Claude, and that the actual aliens will be watching alien fashion videos and won't understand how or why we were contacted.
How long until the last math problem of significant interest is solved, in a way that can be written up in an understandable <=20 pages that is self-contained against 2026 mathematics?
9%<6 months
22%<6 years
19%<60 years
50%longer
90 votes • Final resultsIf we get to the point where there's a math paper [or biology paper, or chemistry paper, or a how-we-harness-terrestrial-black-holes-to-power-datacenters-paper] that's truly beyond human understanding, it won't be because the AI got so smart, but because we decided not to keep up
18
1,484
I get the point but at least in isolation I find this too simplistic. In fact, medicine *currently* doesn't satisfy the suggested standard; many common medicines (Tylenol!) lack well-understood method of action.
The problem with human/AI decoupling are long-term.
1/14
Believing that mathematics as a human discipline has a fundamental role to play in society now and in the future is not enough for human mathematics to continue to flourish.
We need a new system for how we identify "awesomeness" in math.
Here is a concrete proposal:
1/
Pure math is many things to many people. No surprise everyone feels differently about AI.
One belief: benefits of math come from direct economic impact of breakthroughs on the real world via sci+tech.
Actually, newsflash, this one is almost always wrong!
1/
Narrator: the end of the earth came and went before any matrix was stored on any computing device ever for which this matrix multiplicationa algorithm was faster than Strassen. (Etc etc)
“okay, so what? it’s just math.”
no, you don't get it.
here’s what OpenAI’s math breakthroughs could eventually help make possible in the real world:
Vlasov–Maxwell → stronger foundations for fusion research, whose ultimate prize is virtually limitless clean energy to power civilization
Calderón’s Problem → portable body scanners using electrical signals, making medical imaging cheaper and easier to access
Maximum-Cardinality Matching → faster searches for compatible kidney swaps across large donor pools, helping hospitals coordinate lifesaving transplants
Inverse Elasticity Problem → scans that map tissue stiffness, helping doctors locate suspicious growths inside the body
Mumford–Shah Conjecture → better tools for spotting tissue changes in brain scans, helping doctors identify signs of disease
Bose–Einstein Condensation → stronger foundations for quantum sensors that could help vehicles navigate without GPS
Simple Stochastic Games → better safety checks for self-driving cars and robots before dangerous mistakes reach the real world
Matrix Multiplication → cheaper AI and bigger scientific simulations using the same computers
Edit Distance → faster DNA comparisons, helping researchers study genetic changes linked to disease
The k-Server Problem → robots that waste less movement and energy, making warehouses more efficient and goods cheaper to move
a reminder that math is the foundation of science.
so accelerating mathematical discovery could compress centuries of scientific progress into years.
insane timeline to be alive for!
1
17
1,378
Just wanted to take a moment to remember the breaking of the 4 minute mile, 201 years ago last week.
I think it's easy for outsiders (maybe also insiders?) to fail to appreciate just how big math is (to us); how many "famous" questions there are.
As a community we manage to have huge number of problems whose solutions would be considered "holy grail" type mega-level math.
This is a great article.
Honestly one really interesting aspect of the whole AI-in-math thing is that it has led to the opportunity to read/hear/discuss the wide range of people's perspectives on what mathematics actually is / is for.
its.caltech.edu/~matilde/AIm…
Very nice piece written by my colleague Matilde Marcolli, arguing why the current panic in the math community is unwarranted.
To appear in Newsletter of the London Mathematical Society.
14
1,486