ArXiV has a new LLM policy (Screenshots with alt text so you don’t have to click through to the other place and see all the stupid responses)

Happy to finally have this work from a year ago with Curt von Keyserlingk and @austen.uk published on PRR! In this work, we linked and generalised error correction / inference / learnability problems to quantum and non-optimal regimes, and studied their phases. journals.aps.org/prresearch/a...

they need to make a pokemon out of this

Howtown doesn't get enough love. I enjoyed the latest episode, especially the bit about rats on cocaine enjoying jazz: youtu.be/0ZYhyewNQMo&...

Has anyone else noticed that Norm MacDonald's moth joke sounds just like Crime and Punishment? Only started reading the latter recently and can't stop laughing when reading it, though it's supposed to be serious I assume...

coming from windows? life pro tip! create a puzzling '3D Objects' folder on your OS of choice to make it feel more familiar

I have to say, publishing results as zero knowledge proof does have one benefit: it tricks security researchers into learning about quantum circuits. This is some amazing work! https://blog.trailofbits.com/2026/04/17/we-beat-googles-zero-knowledge-proof-of-quantum-cryptanalysis/

Tailor Bird sews leaves together with plant fibre to form its nest!

Actually, I think there are infinite, but are all variations of that 7 hex case with a fully connected middle tile. Removing that, not convinced a non-trivial solution configuration exists (eg. swapping a straight line 180 degrees). Would love to be proven wrong! Here's a 9 hex case I found.

I wonder if it's easy or hard to find configurations with multiple allowed solutions

For some reason I thought the left was a giant apple

From a pen & paper theorist point of view, I find it much harder to fall into the trap described here, since I find that it *is* the point to understand every detail of the proof (as you say). In contrast, for coding, often I just want a function without understanding every detail and I start vibing

UK greenhouse gas emissions are now down 54% since 1990. They fell 2% last year. It's something the country doesn't boast about enough. Yes, some of it is switch from industry to services. But the biggest reductions are from changes to electricity supply. 👇 www.gov.uk/government/s...

don't forget sepia mode

what got me was modelling the observer's death with a replacement channel to the global thermal state ρ(β) scirate.com/arxiv/2603.2...

If you have agents who are perfect Bayesians, surely they have a prior and posterior modelling the other agents. But then they have to model that they have a prior and posterior etc... how does one stop the infinite recursion?

Have you heard the story of Barkley’s good friend Lin, a Chinese immigrant scientist in Iowa? www.wbur.org/onlyagame/20...

If you have agents who are perfect Bayesians, surely they have a prior and posterior modelling the other agents. But then they have to model that they have a prior and posterior etc... how does one stop the infinite recursion?

it’s on jooble dude just log in to bebee

omg @craiggidney.bsky.social also also uses the convention (R, G, B) = (X, Y, Z), feeling validated rn youtu.be/SnpLSvyyEx8?...

whatever man Sure let me login with wetdry world who gives a shit

now that's an exponential

Hot take: distributions with infinite variance don't really exist in real life. Everything has a cutoff due to system size. Of course, if the system size grows, then the cutoff grows. So in my opinion distributions with infinite variance are really describing a limit.

Game of Life

In somebody else's words:

Is measuring syndromes then using a decoder to infer the most likely error the best thing to do when trying to do quantum error correction? I prove cases where it is, where it isn't, and how to find the optimal threshold. Happy to get this out: let me know what you think! scirate.com/arxiv/2603.0...

Hey look it's good news

In 1874, Georg Cantor published one of the most important papers in math’s 4,000-year history. The ideas in it were stolen. 🧵

Yeah but really just two sizes right

oh, two is the only even prime? golly! three is the only threeven prime. five is the only fiven prime. pretty cool!

Thought I proved something last week. Got shown a fatal flaw by my collaborator on Monday, got very disappointed and came up with another proof on Tuesday and sent it off to just to realise myself it was also wrong. Now I think I have a legit proof but what a rollercoaster it's been 😅 (and no sleep)

I made a map of 3.4 million Bluesky users - see if you can find yourself! bluesky-map.theo.io I've seen some similar projects, but IMO this seems to better capture some of the fine-grained detail

This in academia

not sure if it stems from the love of reading but i (have to) do this

I just have questions, no answers

Apart from transitions from/to a Gaussian fixed point, are there any other examples where field theory is actually useful?

Is there an existing *random* construction of 'good' quantum LDPC codes similar to the classical case? The only ones I've seen around are highly structured i.e Cayley graphs

Please, no more endless scrolls

qLDPC: kill me....

Surprisingly LLMs have only recently surpassed the usefulness threshold for me: Generating an iCal file from a pdf or a website. Few months ago they would fail miserably, but now they can do it, with some back and forth. Now I'm fully convinced that they're going to enslave us all

interesting talk about quantum hidden markov models and inference by @sunwoopkim.bsky.social

If I gave you these evolution rules, can you tell me what will stay constant during the evolution? Very belated post on recent work on hydrodynamics of circuit systems! Work with Friedrich Hübner, Juan P. Garrahan, and Benjamin Doyon. See more here: arxiv.org/abs/2503.08788 (1/7)

Stage 1: there is Markov's inequality Stage 2: there is Markov and Chebyshev Stage 3: there is Markov and Chebyshev and Chernoff and Hoeffding Stage 4: there is Markov and Chebyshev and Chernoff and Hoeffding and Bernstein and... Stage 5: there is Markov's inequality

Le Monde reporting that a French scientist traveling to Houston to attend a conference was denied entry to US after a search of his phone & computer revealed messages critical of Trump's science cuts, "which [says CPB] conveyed hatred of Trump & could be qualified as terrorism". Computer confiscated

google search for synonyms skyrocketing

I believe the technical term is 'rawdogging'.

Here are some slides on our latest work on inferring the position of a random walker from its noisy images, now out in PRE: go.aps.org/3XbRTUp (1/8)

If I gave you noisy 'images' of a moving object, is there a threshold signal-to-noise ratio where you cannot infer its position? It turns out that the answer is 'no' in 1D, but 'yes' on a tree. This transition is like MIPTs in quantum systems. Work with @austen.uk now out in PRE: go.aps.org/3XbRTUp

Whenever I see phase diagrams of spin glasses, I see that the phase boundary of SG from the Nishimori point to zero temperature is a straight line. Is there an easy way to see why this is so? #physics #spinglass #qec