Taking humanoid soccer out of the lab: no instrumentation, no controlled environment. Real fields, around people. Trained in simulation. Deployed in reality. No post-training. Just pre-training. Dribbling. Passing. Turning. Scoring goals. Work led by Avi, Khai, @nolan_fey along with @martinpeticco @JohnMarangola, and Venki.

In-context continual learning requires models to accumulate experience and reuse it later in the same sequence. But an RNN compresses an ever-growing history into a fixed-size state, where each token gets a single write into memory. We study dynamic compression: letting the model revisit the past and reorganize its state as it discovers what needs to be reused.

Continual learning running on a phone. Not inference. Not a fine-tune you kick off with a dataset. A small language model (SLM) that keeps learning from its own experience while it's being used, with every gradient step on the handset. How it works (lin826.github.io/SLM-Online-…):

Excited to share SPD: simulation pre-training for dexterity. We pre-trained a policy in simulation and fine-tuned with less than 2 hours of real data (with @sarthakkamat)

You already know prompting can change what an LLM does. Turns out it can change what a robot does too — we made a robot learn a task it kept failing just by rewriting the prompt. No retraining. No new data. Better prompt in, better robot out. 🧵 1/8

We never really knew how to train nonlinear RNNs well… BPTT struggled with vanishing grads (no long-range memory) and sequential rollout (hard to parallelizable). What if instead an oracle told us the optimal memory state m_t at each step? Then the RNN could do one-step supervised learning on (m_t, x_{t+1}) → m_{t+1} labels. We call this Supervised Memory Training (SMT): a replacement for BPTT that trains RNNs without unrolling them. SMT is time-parallelizable and solves vanishing gradients. Website: akarshkumar.com/smt/ arXiv: arxiv.org/abs/2606.06479

What if the best visual reasoning steps are ones humans can’t specify? 🤔 Existing VLM reasoning is often constrained by language, pixels, and human-designed intermediates. We introduce Latent Implicit Visual Reasoning, where we show that VLMs can discover the best visual reasoning steps by themselves — no bboxes, no intermediate images, no extra supervision. Presenting this week at CVPR! (1/n)🧵

Your RL post-training may be sabotaging your LLM’s test-time scaling! Conventional RL pretends that you can collapse all reward signals *upfront* into a single *scalar reward*. We introduce Vector Policy Optimization (VPO), which natively maximizes *vector-valued* rewards, boosting test time search performance, even on the original scalar.

Eka means unity -- “one,” in Sanskrit and “first” in Finnish. We’re building intelligence for the physical world in its native language: forces. Until now, robotics faced a tradeoff — generality or speed. The real world requires both. Robotics also faced a data problem. Our Vision–Force–Action (VFA) model — the first of its kind — breaks the generality-speed tradeoff and the data barrier. It's a new foundation uniting performance, generality, and safety for putting capable robots in everyone's hands. Today, I am excited to share our journey of pushing robots beyond human limits. Today, dexterity becomes scalable. Today, I welcome you to the Era of Eka. Co-founded with @haarnoja, and so thrilled and grateful to be working with a dream team at @EkaRobotics. Learn more: ekarobotics.com