RE: https://mathstodon.xyz/@bgavran/116660321337179277 Regarding recent news

Finally added a reinforcement learning section to the list: https://github.com/bgavran/Category_Theory_Machine_Learning I'm sure I'm missing a few papers

GitHub - bgavran/Category_Theory_Machine_Learning: List of papers studying machine learning through the lens of category theory

List of papers studying machine learning through the lens of category theory - bgavran/Category_Theory_Machine_Learning

github.com

Almost forgot to mention: I'll be presenting TensorType today at 16:30 EEST at the ACT conference https://actconf2026.github.io/programme.html

Our work on compositional behavioural semantics for state abstraction was accepted at ICML! Link: arxiv.org/abs/2606.25357 State abstraction compresses an RL state space while trying to retain decision-relevant behaviour. What survives? 1/

Compositional Behavioral Semantics for State Abstraction in Reinforcement Learning

State abstraction plays a key role in scaling reinforcement learning to complex but structured systems. In studying such systems, a wide range of behavioral structures have been studied in reinforceme...

arxiv.org

RE: https://mathstodon.xyz/@julesh/116658899165675527 C3, but with a caveat:

It was pointed out to me that the factor in this blog post is off by 1000x. AlphaZero achieves superhuman performance compared to GPT4 not with 30x fewer parameters, but 30000x fewer, which is an even more stark difference.

Bruno Gavranović@bgavran.bsky.social · 7mo ago

I just published a new blog post! Types and Neural Networks ( https://www.brunogavranovic.com/posts/2026-04-20-types-and-neural-networks.html ) It's about what comes into view once dependent type systems become a part of neural network training.

I just published a new blog post! Types and Neural Networks ( https://www.brunogavranovic.com/posts/2026-04-20-types-and-neural-networks.html ) It's about what comes into view once dependent type systems become a part of neural network training.

New blog post coming up soon! I'm very excited about some of the work we've been doing at GLAIVE about making the output space of frontier models *typed*. Stay tuned :)

I never wrote about it here, but as of some time ago I figured out a basic implementation for named axes in TensorType: https://github.com/bgavran/TensorType This means that now you're: a) forced to assign some meaning to all your axes)

something strange and horrible and somehow fitting that we should have a very real threat of madman civilizational destruction at the very moment when we also have humans on the dark side of the moon taking pictures that show how small, precious, and beautiful our world and existence is

"Coalgebras for categorical deep learning: Representability and universal approximation" https://arxiv.org/abs/2603.03227

Where are the nuanced left-wing takes on modern AI and LLMs? So much of the discourse around this tech is centered on rejecting it because of who currently owns it. But like all tech, it can be used for both oppression and liberation. Who is focusing on the latter?

Lots of exciting stuff has been happening at GLAIVE: https://mathstodon.xyz/@julesh/116279930318085086 https://mathstodon.xyz/@Andrev@types.pl/116255766617293854

Discworld QOTD, from Monstrous Regiment “Stopping a battle is much harder than starting it. Starting it only requires you to shout ‘Attack!’ but when you want to stop it, everyone is busy.”

Micah@rincewind.run · 8mo ago

Discworld QOTD, from Feet of Clay

Masteful sequence of quotes in Mike and Ike which I never noticed before

never, ever, ever, ever accept "how will you pay for it?" as an argument against social programs.

Leah McElrath@leahmcelrath.bsky.social · 8mo ago

Preliminary Pentagon cost estimate of the war in Iran is $1 BILLION a day. So far.

RE: https://mathstodon.xyz/@julesh/116103778140182700 Some more work I've been a part of:

julesh (@julesh@mathstodon.xyz)

Attached: 1 image New blog post! Autodiff through function types: Categorical semantics the ultimate backpropagator https://julesh.com/posts/2026-02-20-categorical-semantics-ultimate-backpropagator.html

mathstodon.xyz

New blog post! Autodiff through function types: Categorical semantics the ultimate backpropagator https://julesh.com/posts/2026-02-20-categorical-semantics-ultimate-backpropagator.html

The transition from “AI can’t do novel science” to “of course AI does novel science” will be like every other similar AI transition. First the over-enthusiastic claims that are debunked, then smart people use AI to help them, then AI starts to do more of the work, then minor discoveries, & then…

there is a widespread belief among people with even a little technical and scientific savvy that Alchemy was a bunch of hooey, that they never published their experiments, that alchemists had to each discover on their own not to drink mercury, etc but this is urban legend!

This is not what is happening at all. The amount of misinformation on BlueSky about AI is insane, and it keeps promising that AI is all hype that is going away soon. A really dangerous position that cedes all AI policy and decisions about how it will be used to others. Also Futurism is clickbait

I'm quite proud of how far I've been able to get with TensorType: https://github.com/bgavran/TensorType What started out as a casual "I wonder if I can implement type-safe tensors" question has now evolved into a fully-fledged library

GitHub - bgavran/TensorType: Framework for type-safe pure functional and non-cubical tensor processing, written in Idris 2

Framework for type-safe pure functional and non-cubical tensor processing, written in Idris 2 - bgavran/TensorType

github.com

Come and visit me at the poster session at #QIP26 today! :)))

If you're interested in some of the design choices behind TensorType, have a look at the great blog post that @Andrev just posted: types.pl/@Andrev/1159... TLDR; Tensors in NumPy are secretly built out of the composition product of containers

Dear USAans, ICE will not let you complete an electoral process which might result in a government that might hold them to account. If you want your democracy back, you have to get rid of them *first*. Abolish ICE, you say? How, I ask?

On the drive home I was idly thinking about what changes I'd make, if I could, to our system of governance after this administration is - ideally - gone. I suppose in no particular order, here is a list of what I'd do, sorted by the mechanism for doing it.

If you've been curious what I've been up to, the recently published report from GLAIVE reveals a part of it: https://glaive-research.org/2025/12/08/q4-report.html

Different products of containers, with examples on the List container: List ⊗ List -> rectangular array List ∘ List -> ragged array List × List -> two lists List + List -> a boolean value and a list

Remember kids, theoretical physicists have irreparably ruined public trust in science by saying they *hope* we *could* find BSM evidence at the LHC, but Elon gets to predict his products will "do the sci-fi" and be wrong 100 times per year for profit but he's just giving them hope

🇵🇸Tim Henke (tɪm 'ɦɛŋ.kə)@timhenke.bsky.social · 12mo ago

The brave renegade truth teller: "the people must be made to believe obvious lies to pacify them"

Had an absolutely fantastic time in Scotland. It's been great seeing a lot of friendly faces, and making exciting research progress with Glaive