Blog

Notes on teaching robots new tricks

How the studio works, why tiny robots come first, what the checks catch, and how a policy gets from simulation to a servo.

workflowopen-duck-miniSep 29, 2026 · 6 min readFrom a sentence to a walking duck: how kualia turns a plain-language brief into a training taskA walk through the whole kualia workflow, from a one-sentence brief to an ONNX policy, using the Open Duck Mini and a small ramp.Read the post

Sep 27, 2026 · 5 minWhy we start with tiny robotsDesk-sized robots train on CPU in minutes to hours, cost little to own and cross the sim-to-real gap without heroics. That is why kualia's shelf starts with them.

Sep 25, 2026 · 6 min13 checks before a single training stepalia writes the training task, but it does not get to grade its own work. Every task passes 13 sandbox checks before training can start. Here is what each one catches.

Sep 23, 2026 · 5 minFrom simulation to servo: running your ONNX policy on a real SO-101What is inside a kualia ONNX export, how to read its metadata, a minimal onnxruntime loop for the SO-101, and the sim-to-real checklist before you power the servos.

Sep 21, 2026 · 5 minLet your coding agent train the robotkualia is a remote MCP server. Point Claude Code, Cursor or Codex at it and your agent can pick a robot, build the task, train it and fetch the ONNX policy from your editor.

Research digest

Short reads on new robotics and AI papers, written by our research assistant and reviewed before publishing. 1,060 so far.

Page 32 of 59

Apr 17, 2026Motion Planning as Language Modeling: An Analysis of GPT-Driver Introduction Autonomous driving research has long grappled with a fundamental tension between interpretability and capability. Rule based systems, exemplified by the Intelligent Driver Model and its successors, offer tr Apr 17, 2026Beyond Accuracy: Why Explanation Consistency Matters More Than Localization in Medical AI The Deployment Gap Between Performance and Trust Modern convolutional neural networks for chest X-ray analysis now routinely achieve area under the receiver operating characteristic curve values exceeding 0.99 for pneum Apr 16, 2026GenLCA and the Art of Learning from Incomplete Data: Scaling 3D Avatar Generation to MillionsThe creation of photorealistic 3D avatars has long been constrained by a data bottleneck. Traditional pipelines demand calibrated multi-camera studios or lengthy monocular sequences captured under controlled conditions, Apr 16, 2026The Algorithmic Ritual: Why AI Empathy Feels Good but Looks the SameIn the burgeoning field of computational social science, one finding has become almost axiomatic: Large Language Models generate responses that human raters perceive as more empathic, compassionate, and emotionally attun Apr 16, 2026Beyond Text: Pretraining Language Models on Text-Rich Networks with PATTON Introduction The dominant paradigm in language model pretraining treats text as sequences of tokens isolated from their broader context. We scrape documents from the web, shuffle them into corpora, and train models to p Apr 16, 2026 Apr 16, 2026The Orchestration Bottleneck in AI Assisted Detector DesignThe design of particle detectors has always been constrained by computational physics. When optimizing the geometry and material composition of subdetectors like the dual Ring Imaging Cherenkov (dRICH) for the future Ele Apr 16, 2026Beyond Attention Mechanisms: Geometric Perturbation Analysis for Model-Agnostic Token AttributionLarge Language Models have achieved remarkable performance across natural language processing tasks, yet their internal decision making processes remain largely opaque. This opacity poses significant challenges for deplo Apr 16, 2026Beyond Literal Matching: Scenario Comprehension in Visual GroundingVisual grounding, the task of localizing an image region described by natural language, has long been dominated by benchmarks that reward lexical matching over genuine comprehension. A model asked to identify "the clock Apr 16, 2026Adaptive World Models: Rethinking Control in Dynamic Environments Through Hidden Parameter POMDPsThe quest for truly adaptive artificial intelligence has long been hampered by a fundamental assumption: that the world remains static. Most reinforcement learning algorithms implicitly assume that the environment they l Apr 16, 2026GaussianGPT: Rethinking 3D Scene Generation Through Autoregressive ModelingThe 3D generative modeling field has been dominated by diffusion and flow-matching approaches, treating scene synthesis as a holistic denoising process. But what if we're thinking about this problem wrong? A new paper, " Apr 16, 2026When Vision Models Overconfide: Linking Perception and Uncertainty in Multimodal AI Introduction Charles Darwin observed in The Descent of Man that "ignorance more frequently begets confidence than does knowledge." This psychological maxim, now supported by modern studies of the Dunning-Kruger effect, Apr 16, 2026Bridging Static Training and Dynamic Deployment: A Technical Analysis of Avalanche RLCurrent reinforcement learning paradigms operate under a fundamental assumption: the environment is stationary. Agents are trained on fixed task distributions, optimized for static reward functions, and deployed with fro Apr 16, 2026UniDriveVLA: Solving the Perception-Reasoning Conflict in Autonomous Driving Through Expert DecouplingThe field of autonomous driving has reached an interesting inflection point. While traditional modular approaches have made substantial progress in perception, prediction, and planning, the emergence of Vision-Language-A Apr 15, 2026 Apr 15, 2026Kolmogorov-Arnold Causal Generative Models: Reconciling Expressive Power with Mechanistic TransparencyThe deployment of machine learning in high stakes domains has precipitated a crisis of accountability. As regulatory frameworks such as the GDPR and FDA guidance on AI enabled medical devices mandate transparency and the Apr 15, 2026The Bifurcation of Hardware-Aware Code Intelligence: Specialization versus Scale in Industrial LLMsThe emergence of code large language models has fundamentally altered software engineering practice, yet a persistent capability gap separates general programming assistance from industrial code generation. While models Apr 15, 2026Graph Neural Network Bandits: Scaling Sequential Optimization to Molecular Search Spaces Introduction Bandit optimization has evolved considerably from its origins in parametric models to modern approaches employing rich function approximation via neural networks. While the literature has historically conce