Can’t say I’m very surprised by the results here; too often it takes Job-like patience and persistence to navigate the editorial process today, and journals should take note that there are obvious alternatives.

ghiblified graduation

Wait, what?

🎉I am very happy to see our Barren Plateau Review published in Nature Review Physics. This article condenses 6 years of our LANL work, but also so many amazing papers by the community!! Check it here 👇 nature.com/articles/s4225…

Wow, cool paper! using off-the-shelf DRAM for low precision matrix-vector multiplication arxiv.org/abs/2503.23817

Happy quantum day

during NYC deep tech week, i gave a talk about how Normal Computing 🧠🌡️ is scaling our thermodynamic computing paradigm to tackle large scale AI and scientific computing workloads today, we're releasing a blog post based on that talk! ⬇️

Come out and join us, so we can talk about physical world AI and thermodynamic chips. Looking forward to hosting with Nathan Benaich.

posteriors 𝞡 published at ICLR! I’ll be in Singapore next week, let’s chat all things scalable Bayesian learning! 🇸🇬👋

posteriors 𝞡, our open source Python library for Bayesian computation, will be presented at #ICLR2025! posteriors provides essential tools for uncertainty quantification and reliable AI training. Join Sam Duffield at his poster session 'Scalable Bayesian Learning with

We're excited to share that our research "Thermodynamic computing system for AI applications" has been published in Nature Communications, detailing our early work on a prototype thermodynamic computer. Our stochastic processing unit used coupled RLC circuits with injected noise to demo

Can randomness in physical systems be harnessed for computation? Made with #python #numpy #matplotlib

Here’s an analogy for how research works (partly inspired by Patrick Coles ). It’s like building a pyramid; you start with the first layer, which is the largest, and takes a lot of work. Then you build another layer on top of that, which is a little smaller and a little

Interested in the science of language models but tired of neural scaling laws? Here's a new perspective: our new paper presents neural thermodynamic laws -- thermodynamic concepts and laws naturally emerge in language model training! AI is naturAl, not Artificial, after all.

I’m proud to have contributed to Normal Computing 🧠🌡️’s mission of creating a unified theory of thermodynamics and AI over the last two years. Computer science research should aim to be basic science research. I’m excited to share a self-contained introduction to thermodynamic

The brain drain is happening now. Reddit is flooded with kids who planned to do their STEM PhDs looking for advice to apply in Europe. All because a huge chunk of people in this country can’t admit that actually, having smart people around is good and we should listen to them

Interesting paper by Stephen Whitelam at LBNL that builds on our work on thermodynamic computing. In 2023, our team at Normal proposed to implement diffusion models on stochastic hardware, in our paper "Thermodynamic AI and the fluctuation frontier". Here, Stephen investigates

VerilogEval is an odd coding dataset - it gives the agent access to the golden testbench, which you will use to evaluate the final design from the model. I'm honestly surprised how papers like VerilogCoder or ChipAgents only get ~95% accuracy in this setting... That's what

New website!

There will be much fanfare at our next Proof Liturgy. As always, all math curious are welcome!