London, England
Joined January 2012
- Tweets176
- Following727
- Followers632
- Likes1.2K
Sam Duffield retweeted
adrientaylor.github.io/tutor… are all very good! I would guess that Adrien has some nice talks on YouTube also.
Sam Duffield retweeted
Maybe a good time to look inwards a bit, and clarify for oneself what makes { a good proof, a useful construction, an effective bound, ... } and so on. I guess that we will be 'complaining' for some time (who knows how long), and it seems healthy to articulate our standards well.
Sam Duffield retweeted
My new book written with @sirui_lu97 and @HoldijkLars is ready to be ordered!
Major societal and economic shifts have always been anchored in foundational physics.
During the 1st Industrial Revolution, thermodynamics showed us how to harness energy to power steam engines.
Sam Duffield retweeted
Ok I do want to soften my statement a bit. I wouldn’t claim that the numerical majority of papers in these conferences are overclaiming or fraudulent. It’s more like, as a reader, if you select for the papers that seem most exciting, then those ones will be majority overclaims.
Sam Duffield retweeted
CN101 - A Digital Thermodynamic Computer for Generative AI is now live on arXiv.
Conventional chips spend energy suppressing noise to compute deterministically; CN101 is a noisy system engineered so its settling behavior is the computation, implemented in digital CMOS and evaluated on generative AI workloads from image generation across six chips to molecular free-energy estimation.
In a companion post, @ColesThermoAI, our Chief Scientist, shares reflections on the journey and why AI inference needs hardware built for its physics.
Paper: arxiv.org/abs/2608.00754
Patrick's post: normalcomputing.com/blog/ai-…
Quanta article: quantamagazine.org/thermodyn…
#ThermodynamicComputing #AIHardware
Sam Duffield retweeted
cuthberto-carlos predictions for the final 2 games of the world cup!
I for one am excited for these games. model pits all teams at around the same skill levels, so it will be an interesting to see who comes out on top!
state-space-models.github.io…
Sam Duffield retweeted
Our book "Generative AI and Stochastic Thermodynamics: A Tale of Free Energies" is out this month. Over the coming weeks @sirui_lu97, @wellingmax and I will post about some of the topics it covers.
Starting it off: heat and work in latent variable models and variational EM.
Sam Duffield retweeted
knockout stages predictions are out! 2/3 correct scores at the 90th minute i think thats pretty impressive @Sam_Duffield !
state-space-models.github.io…
Sam Duffield retweeted
so how does the model compare to other predictions? I think pretty close!
the model doesnt take into account home advantage of team USA, so we see a big discrepancy in model estimates 😱😱😱 do you guys think that the home advantage should be that big?
state-space-models.github.io…
Sam Duffield retweeted
after sweden's most recent 5-1 win against tunisia, we saw the Sweden's attack and defense scores improve! On 11 June, @Sam_Duffield 's model predicted that 2-0 was the most likely outcome. now its 1-1!
check out the model predictions here state-space-models.github.io…
Sam Duffield retweeted
Fixed point iterations for parallelizing nonlinear dynamics is all the rage:
- Newton for RNNs
- Picard for diffusion models
- Jacobi for parallel decode of LLMs
But how do these techniques relate, and when should you use them?
We show you how in our new paper 🧵
Sam Duffield retweeted
At @iclr_conf in Rio? Join us for the Normal Silicon Social on Sunday, April 26 from 6:30 to 8:30 PM at the Copacabana Palace.
Request an invite: luma.com/3ioi64wv
#ICLR2026
Sam Duffield retweeted
Did @wellingmax’s keynote at #iclr2026 leave you wanting more stochastic thermodynamics and generative AI?
Come join Max and me at poster #1023 in Pavilion 3 in the afternoon to chat about how we can learn escorted protocols for multistate free-energy estimation!
Sam Duffield retweeted
Normal Computing is sponsoring @iclr_conf 2026 in Rio de Janeiro.
Our research team will be on-site presenting papers spanning thermodynamic computing, Normal's physics-based computing architecture, and the AI methods underpinning Normal EDA, our platform for silicon engineering.
Come talk to the researchers behind the work. We'll share a preview of what we're launching live at the conference. Follow along for real-time updates in the ICLR app.
Meet us at the booth:
Thursday, April 23 – Saturday, April 25
9:30 AM – 5:30 PM daily
Sam Duffield retweeted
Parallelizing nonlinear RNNs is gaining traction!
More efficient than transformers; more expressive than linear RNNs.
My PhD thesis provides an intro guide to the math (Newton's method) behind the parallelization.
Great as a quick-start if you want to explore this new field!
Sam Duffield retweeted
Struggling with minibatch noise in Stochastic Gradient Bayesian Inference? Want your chains to naturally run on stochastic hardware?
Introducing — Stochastic Gradient Lattice Random Walk!
New work from @NormalComputing in collab with @zierhjmensch, @adnarim066, @wellingmax, and the dream team at normal computing, @Sam_Duffield, @MaxAifer and @ColesThermoAI.
Sam Duffield retweeted
We just released cuthbert, a JAX library for state-space model inference.
It provides various algorithms for Bayesian filtering and smoothing in one clean, functional API.
Repo: github.com/state-space-model…
