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 31 of 59

Apr 18, 2026The Threshold Query Pattern: Selective Language Model Intervention for Out-of-Distribution Reinforcement LearningReinforcement learning agents, despite their impressive performance in controlled training environments, often collapse into random or hazardous behavior when encountering states that deviate from their training distribu Apr 18, 2026The Recursive Ceiling: Analyzing the Limits of Bilevel AutoresearchThe dream of automated scientific discovery has always contained a recursive temptation. If a system can conduct research, why can it not research the process of research itself? In Bilevel Autoresearch: Meta-Autoresearc Apr 18, 2026The Compositional Tokenization of Light: Why Disentangled Control Matters in Generative Image RelightingLighting defines the visual world. It sculpts form, directs attention, and establishes mood. Yet once a photograph is captured, the lighting is fixed, immutable. Traditional post-production offers limited recourse: globa Apr 18, 2026The Verification Tax: Why Better AI Models Are Harder to AuditThe most cited calibration result in deep learning history, the post temperature scaling Expected Calibration Error (ECE) of 0.012 on CIFAR 100 reported by Guo et al. in 2017, sits below the statistical noise floor. For Apr 18, 2026Rethinking Prompt Selection in Continual Learning: The Geometry of Prototype-Guided Learning Introduction Continual learning remains one of the persistent challenges in modern machine learning. Apr 18, 2026Beyond Binary Detection: Why Provenance Matters More Than Origin in LLM Text ClassificationThe detection of synthetic text has reached an inflection point where technical capability must confront regulatory nuance. While the field has spent years refining binary classifiers to distinguish human writing from ma Apr 18, 2026Measuring the Paper Itself: Local Verification Infrastructure for the Age of AI-Assisted Science WritingThe scientific publication ecosystem currently offers researchers two inadequate options for quality assurance, each failing in distinct ways. Journal gatekeeping promises to verify both integrity and contribution throug Apr 18, 2026Target Policy Optimization: Decoupling What to Learn from How to Learn ItThe field of reinforcement learning from human feedback (RLHF) has been grappling with a fundamental architectural problem that becomes particularly acute under sparse reward conditions. A new paper titled "Target Policy Apr 17, 2026The Tacit Knowledge Gap: Why LLM Judges Fail to Capture Developer Preferences in Code EvaluationAs large language models become ubiquitous in software engineering workflows, the question of how to evaluate their outputs has shifted from academic curiosity to practical necessity. The dominant approach, LLM-as-a-judg Apr 17, 2026Active Observation Planning in Video Understanding: An Analysis of LensWalk Introduction Video understanding has long been constrained by a fundamental architectural limitation. Traditional pipelines treat visual perception as a preprocessing step, where videos are uniformly sampled, compressed Apr 17, 2026Cubic Discrete Diffusion: Unifying Visual Understanding and Generation Through High-Dimensional Token PredictionThe pursuit of unified multimodal architectures has long been hindered by a fundamental architectural schism in how we represent visual information. While language models operate on discrete tokens that seamlessly serve Apr 17, 2026Inference-Time Scaling Meets Hardware Synthesis: Evaluating General-Purpose Agents in HLS OptimizationHigh Level Synthesis has long promised to liberate hardware designers from the granular complexity of Register Transfer Level coding, yet the reality remains stubbornly labor intensive. Despite mature toolchains such as Apr 17, 2026Intent as Infrastructure: Why Egocentric World Models Must Model Goals, Not Just GeometryGenerative world models have achieved remarkable fidelity in simulating autonomous driving scenarios, drone cinematography, and static natural environments. Yet the domain of egocentric vision, the first person perspecti Apr 17, 2026The Automation of Environment Optimization: Coding Agents as Infrastructure for Reinforcement Learning The Bottleneck of Simulation Reinforcement learning research has long been constrained by a fundamental asymmetry. While neural architectures have scaled to hundreds of billions of parameters and training frameworks hav Apr 17, 2026The Semantic Abyss in Text-to-Audio-Video Generation: Why Looking Good Is Not EnoughThe transition from silent Text-to-Video synthesis to Text-to-Audio-Video generation marks a fundamental shift in how we interact with generative media. As frontier systems including Sora 2, Veo 3.1, and Kling 2.6 demons Apr 17, 2026PhyEdit: Bridging the Gap Between Image Generation and Physical RealityThe field of image editing has witnessed remarkable progress with diffusion models, yet a fundamental challenge persists: how do we ensure that generated images respect the laws of physics? A recent paper, "PhyEdit: Towa Apr 17, 2026Structured Semantics and Continuous Geometry: A New Framework for Language Conditioned Dexterous ManipulationIn unstructured human environments, functional dexterous grasping remains one of the most challenging frontiers in robotics. Unlike simple pick and place operations, functional grasping requires robots to understand task Apr 17, 2026Think Anywhere: Adaptive Reasoning Mechanisms for Code GenerationThe current generation of reasoning optimized Large Language Models, from OpenAI's o1 to DeepSeek R1, has settled on a consistent architectural pattern. These models perform extensive upfront computation, generating long