The optional stopping theorem states that, under some mild assumptions, you can st

You may find the comments by Melanie Matchett Wood on the discovery of a counter example to Erdős’ Unit Distance Conjecture as interesting as I did 👇🏼(see previous post for a link)

In spite of a rather crappy start* the day ended well: a student sent a message to thank my TA and me for the class, "one of the most rewarding" they've taken. *I'm under the weather, currently more potato than man

And that's a wrap for the semester! All the talks of the season are now available on our website: sites.google.com/view/tcsplus... and YouTube channel: www.youtube.com/@TCSplusSemi... Thank you to all attendees, and to our wonderful speakers—and see you in the Fall! #TCSSky #TheoreticalCS

With this, let me share an important announcement - I am actively seeking highly motivated & mathematically inclined PhD students to work on problems at the intersection of TCS, probability & mathematical physics! If you are interested, and have a strong math background, get in touch!

The OpenAI breakthrough on the unit distance problem seems genuinely impressive to my semi-layman eye. (It is!) One possible takeaway, though, is that if you throw an incredible amount of money and resources at focused research, there WILL be impressive progress. I wish we tried that, too.

A complete post is here: blog.geomblog.org/2026/05/the-... The unit distance problem: Let P be a set of n points in the plane (amen). What is the maximum number of unit distances that can be achieved? (the amen is a running joke in my old field of computational geometry) 2/n

The Zoom link for Shyamal Patel ( @shyamalp.bsky.social )'s talk on "Learning Functions of Halfspaces" is now available on our website. See you tomorrow, 1pm ET! www.tcsplus.org/welcome/next...

Last week of lectures*, concluding with the MWU algorithm and how to learn from imperfect, possibly disingenuous experts. Seems... topical. *(Except for the recap week)

Very excited to explore new roads to fault-tolerant photonic quantum computation with PsiQuantum! rms.arc.gov.au/RMS/Report/D... @arc-gov-au.bsky.social @sydney.edu.au @psiquantum.bsky.social @bartlettquantum.bsky.social @acdoherty.bsky.social @jcbridgeman1.bsky.social @quantum-roberts.bsky.social

"The nested subscripts will continue until notation improves."

🎉 Congratulations to Zhanna Sarsenbayeva, Anusha Withana, and Chang Xu on their #LinkageProjects! These will enable their teams and industry partners to empower children with cerebral palsy, cut energy waste, and deliver better aged care to all Australians! #ResearchWithImpact #USyd (More below)

🚨 #LinkageProjects #LP25r2 announcement: ❗️Outcomes announced publicly for Linkage Projects 2025 round 2❗️ See ARC's RMS for list ➡️ https://rms.arc.gov.au/RMS/Report/Download/Report/a3f6be6e-33f7-4fb5-98a6-7526aaa184cf/290 /bot

FOCS 2026 Test of Time call for nominations is out: tc.computer.org/tcmf/2026/05... Please submit your nominations! More info on the award here: tc.computer.org/tcmf/focs-te... and here's DBLP links for FOCS '16, '06, '96: dblp.org/db/conf/focs... dblp.org/db/conf/focs... dblp.org/db/conf/focs...

"But traveling is inconvenient and expensive!" No, really? Funny you only realize that when it's far from YOU. @neuripsconf.bsky.social

Hey, @neuripsconf.bsky.social, let me ask you. Say you're a student with an accepted paper, and must present it (yay!). Now you can choose to do it at the main conf, travelling there (Australia), or present it 'locally' (say Atlanta). No visa issues, just cost. Which one will your institution fund?

"But traveling is inconvenient and expensive!" No, really? Funny you only realize that when it's far from YOU. @neuripsconf.bsky.social

"AI will replace U, they said" "Nothing yoai can do aboait it. It's the faitaire."

The French ANR project ALARICE (computational complexity bounds “à la Rice[’s theorem]” for finite dynamical systems) has a postdoc position (alarice.lis-lab.fr/postdoc-nice...) and two PhD positions (alarice.lis-lab.fr/phd-fundings...) available. Application deadline 1 June 2026. #AcademicJobs

"AI will replace U, they said" "Nothing yoai can do aboait it. It's the faitaire."

This is one of the reasons I support preprints especially in outbreak science. It's not that preliminary estimates and data don't get shared otherwise. It's that they get shared on conference calls, and private emails, and "Hey Eric, thought you'd be interested in..."

For those who argue about the benefits and role of the #arXiv, let me provide two examples, and ask you this: do you want to live in a world where, to know a key result or idea, you needed to be part of the lucky few at the right conference dinner? Example: cstheory.stackexchange.com/q/28039/13319

📢 Our last TCS+ talk of the season will be Wednesday, May 20 (10am PT, 1pm ET, 19:00 CEST): Shyamal Patel, from Columbia University, will tell us about "Learning Functions of Halfspaces"! RSVP to receive the link (available one day prior to the talk): forms.gle/jeGPub2pqS5R...

There's a very specific type of papers I hate, which I think of as "territory-marking" behavior.* Find a topical topic, one others are likely thinking about too. Go for the most basic, lowest-hanging fruits, *and put your paper online first.* Now you own the topic! (You may also have killed it.)

In establishing a Centre, I am very aware of how few well-designed physical spaces for #mathematical collaboration exist. That may only be due to my personal experience, so I am reaching out to you to ask for examples. Please post ideas, photos, sketches, ideas… The more the merrier #mathsky

New blog post by @stein.ke on #DifferentialPrivacy: what happens when you modify the definition of neighbors to make it asymmetric? Why would you do that, and does that buy you anything? differentialprivacy.org/one-sided/

New blog post by @stein.ke on #DifferentialPrivacy: what happens when you modify the definition of neighbors to make it asymmetric? Why would you do that, and does that buy you anything? differentialprivacy.org/one-sided/

From #TCS4All: You are cordially invited to the TCS for All Spotlight Workshop! The workshop will be held as part of #STOC2026, and is open to all. More info: sigact.org/tcsforall/ See below for TCS for All Travel Scholarships and call for nominations for Rising Stars talks at the workshop.

Computational geometry problem: give an O(n log n)-time algorithm which, given the n-point description of a Bruce Banner, computes its convex hulk

Your annual reminder that the world makes no sense, words mean nothing, and Australia is part of Eurovision.

The existence of the Master Theorem implies that of the Margarita Theorem

We are implementing a similar policy at @arxiv.bsky.social. If there is incontrovertible evidence of LLM slop in a paper, this means the authors did not take the time to read the LLM output and we can't trust anything else in the paper. Penalty is 1 year ban from arXiv followed by...

Wonderful time at the 🏆 2026 School of Computer Science annual awards ceremony last Tuesday, where academics and staff gathered to celebrate the exceptional achievements and outstanding contributions of our students! A few of the many highlights below—and so much talent! 🎓 #USyd

Nominate "Rising Star" speakers at the TCS for All Spotlight Workshop at #STOC2026! Nominees must be senior PhD students or postdocs in TCS. (Preference will be given to speakers on the job market for postdoc/faculty positions.) Apply by ⏰ May 22 (11:59 pm PDT): docs.google.com/forms/d/e/1F...

TCS for All Travel Scholarships will cover the early registration cost at #STOC2026: the scholarships are open to both US and international students. (Preference will be given to students at the beginning of their studies) Apply by ⏰ May 22 (11:59 pm PDT): docs.google.com/forms/d/e/1F...

From #TCS4All: You are cordially invited to the TCS for All Spotlight Workshop! The workshop will be held as part of #STOC2026, and is open to all. More info: sigact.org/tcsforall/ See below for TCS for All Travel Scholarships and call for nominations for Rising Stars talks at the workshop.

🏆 Our PhD student Linwei Tao has been selected as an Apple Scholar in AIML, a fellowship recognising emerging leaders in #AI and #MachineLearning! Linwei is one of only 20 scholars in 2026, and the first recipient from an Australian University! Read more: machinelearning.apple.com/updates/appl...

Gave a lecture about learning and testing the bias of a coin this morning—or, as I like to see it, "estimating Pr[🦘]"

The TheoryFest workshops at #STOC2026 have been announced! Including one on Property Testing, by Anindya De and @shivamnadimpalli.bsky.social ! acm-stoc.org/stoc2026/wor...

Fwiw, 4 months late, and still absolutely no response from the Minister. Crickets. This makes a travesty of the Parliamentary Petition process. www.aph.gov.au/e-petitions/...

Shafi Goldwasser (left), Silvio Micali (right), and Charles Rackoff devised a way to prove that a statement is true without revealing anything about why. www.quantamagazine.org/how-unknowab...

Vertices are red Vertices are blue Graph's bipartite "k equals 2"

Fwiw, 4 months late, and still absolutely no response from the Minister. Crickets. This makes a travesty of the Parliamentary Petition process. www.aph.gov.au/e-petitions/...

Vertices are red Vertices are blue Graph's bipartite "k equals 2"

I've heard a team in Europe came up with an alternative sketching algo for frequency estimation, which works by randomly decreasing suitable counters instead of incrementing them. You maintain a few in parallel, and at the end, you return THE FINAL COUNTDOWN

Join us June 11-12 for a workshop on Differential Privacy and Unlearning in Machine Learning at University of Copenhagen! The workshop will feature tutorials, three great invited speakers (@grahamrc.bsky.social, @ahonkela.bsky.social and @koloskova.bsky.social), as well as contributed talks.

Final, replacement, and sample exams all done! Now, for the hardest part: remember which ones to upload on the class forum.

List of accepted papers at #COLT2026, the Annual Conference on Learning Theory: learningtheory.org/colt2026/acc... (h/t @gautamkamath.com)

Great turnout and wonderful talks at the Raising The Bar 2026 event last Thursday, where our very own Dr. Zhanna Sarsenbayeva discussed her work "Using AI to build better lives"! #RTB2026 #RaisingTheBar #USyd Read more: www.rtbevent.com/zhanna-sarse...

Good notation can make communication so much more efficient and concise! For instance, Me, an unsophisticated person: "The average tea bar seems to be very small" My wife, a physicist: