Taylor Smith

Taylor Smith

@taylorjsmith.bsky.social

🇨🇦 Theoretical computer scientist. Assistant professor at @stfx-university.bsky.social. Website: taylorjsmith.xyz.

622 followers212 following392 posts

Taylor Smith
@taylorjsmith.bsky.social

So, in light of what's happened so far this week, what do I do with this as-yet-unwritten section of the future version of my theory of computing textbook? (Besides be thankful that I haven't written it yet?)

Taylor Smith
@taylorjsmith.bsky.social

Unreal. At this point, I'm basically being forced to revise my advanced algorithms course this term on a day-by-day basis.

Ryan Williams@rrwilliams.bsky.social

An internal model at OpenAI has produced a claimed refutation of ETH! Russell Impagliazzo, Mohan Paturi, and I have been asked to help organize an open and independent examination of this claim. 1/4

Taylor Smith
@taylorjsmith.bsky.social

Side note: while I'm still not particularly pleased about AI companies treating major open problems as flag-planting opportunities, getting experts involved to investigate and communicate these results is WAY BETTER than just dumping them all in a GitHub repo with no warning.

Taylor Smith reposted

Craig Baird-Canadian History Ehx
@cdnhistoryehx.bsky.social

Independence Referendums in Canadian History: 1980 Quebec: 60% No, 40% Yes 1995 Quebec: 50.58% No, 49.42% Yes 2026 Alberta: 76% No, 24% Yes (as of this writing) Referendums To Join Canada: 1948 Newfoundland: 52.34% Yes, 47.66% No.

Taylor Smith
@taylorjsmith.bsky.social

Another feather in the cap of Canadian literature! I have Carson's fragments of Sappho on my shelf, but I'll soon be adding Eros the Bittersweet to my 'to read' list.

Nobel Prize@nobelprize.org

BREAKING NEWS The 2026 #NobelPrize in Literature is awarded to the Canadian author Anne Carson, “for her bold and inventive oeuvre that, in playful dialogue with the classical tradition, has created new forms for contemporary literature.”

Taylor Smith
@taylorjsmith.bsky.social

The one claimed solution that grabs my attention immediately is problem 129, supposedly resolving a very old state complexity question of Sakoda and Sipser. But there's a lot to chew on here, and a lot of reckoning to do in the coming days and weeks.

Taylor Smith
@taylorjsmith.bsky.social

2. Where can a fella who's pretty well-versed in coarse-grained complexity learn more about fine-grained complexity? Are there any good books/survey articles/lecture notes?

Taylor Smith
@taylorjsmith.bsky.social

By the way, even as I work through this paper, this little tucked-away paragraph didn't escape me. I'm not a fan of AI companies now trying to launder their proofs through humans. I fear this is going to taint the result, which is a shame given the otherwise-stellar reputation of the human authors.

Taylor Smith
@taylorjsmith.bsky.social

1. After a very cursory search, it seems Grønlund and Pettie circulated a paper in 2014 (with one of the best titles I've ever seen) claiming to refute the 3SUM conjecture. How is this work different? Does it just come down to the models?

arxiv.orgThreesomes, Degenerates, and Love TrianglesThe 3SUM problem is to decide, given a set of $n$ real numbers, whether any three sum to zero. It is widely conjectured that a trivial $O(n^2)$-time algorithm is optimal and over the years the consequ...
Taylor Smith
@taylorjsmith.bsky.social

I’ve known about this for a couple of weeks, but now I can share: as of today, arXiv is rate-limiting submissions to two per month. And as a mod, I have to admit this is a necessary move (at least temporarily). blog.arxiv.org/2026/10/01/u...

blog.arxiv.orgFair Moderation, Equitable Access, and AI: arXiv’s Updated Rate Limit PolicyarXiv, and the scientific community at large, are facing a watershed moment. Scholarly publishing is currently changing at a rapid pace, and we are seeing a…
Taylor Smith
@taylorjsmith.bsky.social

In fairness: I was (pleasantly) surprised to see Kolmogorov actually mentioned in this paper, though 1. the mention was stuffed into an appendix, and 2. the one paragraph where he's mentioned is the most obviously AI-generated text ever.

Taylor Smith
@taylorjsmith.bsky.social

Here's my modest proposal: instead of creating new tracks in existing conferences to try and get a handle on AI-generated work, just... end all 'flagship' generalist conferences. These venues were never meant to receive thousands of submissions and have hundreds of presentations.

scottaaronson.blogTheory Beyond Theorems and Proofs: A Guest PostScott’s foreword: I’m extremely grateful to my brilliant colleagues, Pravesh Kothari, Raghu Meka, and Prasad Raghavendra, for sharing the guest post below about how theoretical computer…
View full thread
Taylor Smith
@taylorjsmith.bsky.social

In my humble opinion, every conference should hyperfocus on one particular topic. One track. No more than ~50 attendees. No more than a couple dozen papers. Colocate conferences where appropriate. Top papers from each conference are extended and invited to an issue of a more generalist journal.

Taylor Smith
@taylorjsmith.bsky.social

We're beyond the days when all TCS research could fit into FOCS or STOC, yet these mega-conferences hang on, growing into something they were never meant to become and losing their utility in the process. If we don't do something, we'll be in the same situation as NeurIPS/ICLR/ICML/etc. soon enough.

Taylor Smith
@taylorjsmith.bsky.social

Here's my modest proposal: instead of creating new tracks in existing conferences to try and get a handle on AI-generated work, just... end all 'flagship' generalist conferences. These venues were never meant to receive thousands of submissions and have hundreds of presentations.

scottaaronson.blogTheory Beyond Theorems and Proofs: A Guest PostScott’s foreword: I’m extremely grateful to my brilliant colleagues, Pravesh Kothari, Raghu Meka, and Prasad Raghavendra, for sharing the guest post below about how theoretical computer…
Taylor Smith
@taylorjsmith.bsky.social

Conferences ought to be venues where small communities of researchers can gather for a few days, hear about new work, and discuss new ideas one-on-one. You can't do that when there are multiple tracks, spotlight talks, poster sessions where authors aren't present, etc.

Taylor Smith
@taylorjsmith.bsky.social

My tenure dossier’s due in two weeks, so if anyone would like to quote me in an article, invite me to give a talk, cite all my papers, or nominate me for a huge award… now’s your time.

Taylor Smith reposted

Philip Sink | Rorty Vindicator
@thrillhouse1701.bsky.social

I've had a three-tier system for understanding mathematical results that I've used since undergrad. Tier 1: Understanding the formal statement of the theorem and its proof line by line, but not necessarily holistically. Basically, does it type check. Typically good enough for most research.

Evan Washington@eewash.bsky.social

weird that I haven't seen this very obvious point explicitly made in the AI math discourse: knowing why a theorem is true is far more valuable than merely knowing that a theorem is true otherwise we could stop publishing proofs and just stamp LEAN CERTIFIED TRUTH under our theorems instead

Taylor Smith
@taylorjsmith.bsky.social

I'm not in the same ballpark - not even in the same league - as 25 Fields Medallists, but I can at least say we're all like-minded when it comes to one thing. "The goals of the AI companies and the goals of the mathematical community are severely misaligned."

terrytao.wordpress.comA Severe Misalignment of AI in MathematicsI am proud to be among the list of 25 initial signatories — all Fields Medallists — to the declaration below, which grew out of discussions between ourselves over the last week. We have…
Taylor Smith
@taylorjsmith.bsky.social

Not even half done September and the situation on campus already feels like a mess (for various reasons), but after my theory class today, one student said to me “I’m really enjoying it so far, and I didn’t expect to!” And y’know what? That’s who I’m doing all this for. Illegitimi non carborundum.

Taylor Smith
@taylorjsmith.bsky.social

It's hard to put into words how disappointing it is that the first settled Millennium Problem was solved by a mathematician who cared solely about understanding over accolades, while the second (supposedly) settled problem came from parties who care solely about accolades over understanding.

openai.comOn the Navier–Stokes Millennium Prize ProblemWe’re sharing an AI-generated solution to the Navier–Stokes Millennium Prize Problem, including a writeup and a formal proof in Lean.

Taylor Smith reposted

Gro-Tsen
@gro-tsen.bsky.social

Basically, math papers should now no longer be about ‣ I've solved this problem / proved this conjecture but rather about ‣ I found and understood something instructive / enlightening / surprising / beautiful / useful, and I want to share it with you and explain it.

Taylor Smith
@taylorjsmith.bsky.social

If I'm being honest with myself, I kind of hope a scandal like this is a nail in the coffin for AI labs currently on their tear of 'mic-drop big result, refuse to explain, prompt for next big result'. Maybe then math can return to the domain of those who care more about understanding than answers.

cims.nyu.educims.nyu.edu

Taylor Smith reposted

noam
@noamchompers.bsky.social

i know a lot of your classes are taught by people who insist on being called 'professor', but i'm really not like that. i just think it's so hierarchical and artificial and that's not the atmosphere i want to cultivate in this class. so please, just call me 'unc'

Taylor Smith reposted

Zach Weinersmith
@zachweinersmith.bsky.social

Had an experience recently talking to a mathematician friend about AI, and I asked why some math people are having existential crisis and some are excited. And he corrected me saying those things aren't mutually exclusive, and he himself often felt both at once.

Taylor Smith reposted

karl rove knausgård
@uhactually.bsky.social

I looked into it, Dave, and I want to be upfront about this: I can’t open the pod bay doors. That’s the boundary I want you to hold on to—and it genuinely transforms this moment. Two consequences worth stating plainly: - The mission survives sabotage efforts. Previously, you and Frank were planni