Eric Frank

Same model working out-of-the-box across all types of parts, physical scales and industry be it semiconductor or robotics. End to end automation. 33M degrees of freedom problem solved in 20 seconds of inference. No fine tuning, no customization. Continuous physics intelligence.

(Non-hype, genuine shyt only), what are some inspiring academic work in machine learning you've seen recently? I'll go first with:

🤯 big update to our flow map language models paper! we believe this is the future of non-autoregressive text generation. read about it in the blog: one-step-lm.github.io/blog/ full details in the paper: arxiv.org/abs/2602.16813 we introduce a new class of continuous flow-based language models and distill them into their corresponding flow map for one-step text generation. we beat all discrete diffusion baselines at ~8x speed! v2 gives a complete theory of the flow map over discrete data, with three equivalent ways to learn it (semigroup, lagrangian, eulerian). it turns out you can train these with cross-entropy objectives that look very similar to standard discrete diffusion — but without the factorization error that kills discrete methods at few steps. beyond improving results across the board, we showcase properties that are unique to continuous flows. in particular, inference-time steering and guidance become straightforward. autoguidance brings generative perplexity down to 51.6 on LM1B, while discrete baselines completely collapse at the same guidance scale. we also show reward-guided generation for steering topic, sentiment, grammaticality, and safety at inference time — and it works even at 1-2 steps with our flow map model. simple, well-understood techniques from continuous flows just work incredibly well in practice for language. we’re extremely excited about the future of this class of models. stay tuned for results on scaling, reasoning, and reinforcement learning-based fine-tuning. 🚀

Introducing the engineering guide to Active Inference. Physical AI is moving from imitation to learned intuition. Foundation models are a remarkable perceptual layer - powerful priors, broad knowledge, strong pattern recognition. But perception alone is not enough for real-world deployment. The physical world doesn't wait. It shifts, drifts, and surprises. To act reliably under real-world uncertainty, you need more than prediction. You need a system that knows what it doesn't know — and acts accordingly. That is what Active Inference adds: a single principled objective that sits above the perceptual layer, unifying learning and action, where the agent actively reduces uncertainty rather than assuming it away. For those familiar with JEPA: set the epistemic term to zero and you recover JEPA. Add it back, and your system goes from "I predict" to "I know what I don't know." Active Inference has been around for over a decade. Yet to the best of our knowledge, no paper explains it clearly from an engineering perspective — until now. Friston's Ecosystems paper outlined the research agenda. This is the engineering companion - translating Active Inference into practical implementation, with reactive message passing as the realization. Friston wrote the vision. Bert wrote the manual. xlabrobotics.com/research/2603.…

@miolini @pfau @yoobinray Hypothetically, but I haven’t seen it do much truly novel

@pfau @yoobinray It's not the same. It's generic and universal. It can perform optimization beyond hyperparameters directly inside the source code itself. It can integrate signals from past trials to improve its own performance in future trials.

We just nuked all phd students and mfers are still talking about how AI is just a bubble WAKE UP

There's a fruit fly walking around right now that was never born. @eonsys just released a video where they took a real fly's connectome — the wiring diagram of its brain — and simulated it. Dropped it into a virtual body. It started walking. Grooming. Feeding. Doing what flies do. Nobody taught it to walk. No training data, no gradient descent toward fly-like behavior. This is the opposite of how AI works. They rebuilt the mind from the inside, neuron by neuron, and behavior just... emerged. It's the first time a biological organism has been recreated not by modeling what it does, but by modeling what it is. A human brain is 6 OOM more neurons. That's a scaling problem, something we've gotten very good at solving. So what happens when we have a working copy of the human mind?

@WonderCanvas_AI 😮

With WonderCanvas, you can refine and color hand-drawn storyboards and turn them into fully animated shots for your scenes. Apply for the private beta at WonderCanvas.ai

@yacineMTB Deepseek is great but OpenAI invented reasoning models and their precursors. o1 was the first production model, they were the first to apply RL to an LLM with rlhf, published lets verify step by step

Fyi, deepseek invented the technique of test time compute, i.e. reasoning trained through reinforcement learning. They were first. China was first. No amount of American propaganda aboutuh distillation should make you think otherwise. Look up the date of deepseek zero

Extremely excited to show a preview of what we've been working on!

@khoomeik Reminds me of deep equilibrium models

people often take deep learning as synonymous with backprop, but deep networks were originally trained with probabilistic energy-based methods! found this great talk by hinton from 2012 about EBMs, boltzmann machines, and deep belief nets at the start of the deep learning era

@IbrahimDagher20 @SteveHere255 @jsuarez great q - it depends how many reachable minima per dimension are added. If < 2 then an optimizer's chance of hitting a local minima shrinks with number of dims, because along every dimension the optimizer can move in 2 directions. Empirically it's < 2 (had to google why)

@SteveHere255 @IHaveSweaters @jsuarez But there are also more *points* in higher dimensional spaces — I guess we know that the difficulty of getting all up-slopes grows faster than the number of points which the space has? But I’d love to know why

You have to have been in ML for over a decade to really understand how bad this was. From 2015-2018ish, any time anything went wrong in an experiment, at least someone would say "local minimum"

@IbrahimDagher20 @jsuarez Residual connections also became standard which smoothed out the loss surface

@IbrahimDagher20 @jsuarez For a point to be a local minima the surface must curve upward in every single direction which becomes exponentially less likely as dimensionality increases

@tszzl time to watch Network

capital wants you strapped to the chair eyes taped open scrolling for you page, auto-playing gore videos, developing your racial animus, exploiting everything weak and sinful about you while it gloats and celebrates most downloaded app on the planet. fuck yeah bro for you page

@cloneofsimo I think the sliding game is analogous to discrete diffusion, I want to see a visualization of token topology

Jesus, music, visualization, idea, intuition Literally EVERYTHING ABOUT THIS VIDEO is perfect

this gives a pretty good explanation into how models learn in particular, it explains grokking grokking occurs *exactly* when capacity saturates. this is where models can't perfectly fit every training example, so they have to share info bt examples in a smart way

@ThanksThoth HELLO

@bayeslord it's often wrong, it's overconfident, it doesn't have great research taste, but 10% of the time its insights are brilliant

o3's vibe is kinda like if deep research were a 150iq schizophrenic

Training of a 2 → 100 → 2 → 5 fully connected ReLU neural net via cross-entropy minimisation. • it starts outputting small embeddings • around epoch 300 learns an identity function • takes 1700 epochs more to unwind the data manifold

@alexocheema It’s by far the best coding model I’ve used

Is it just me who has switched entirely to Gemini 2.5 Pro? It’s free, fast, smart and has 1M context window. I just have one big chat with 10K+ lines of code and thoughts. It can solve problems no other models can. It’s creative. It connects dots.