frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Robotic Adaptation Suffers from Data Collection Bottleneck

2•aunmesh•1h ago
How can we ease Robotic Adaptation, or the critical last mile of reliable robotic policy deployment?

Vision-Language-Action Policies (such as Gr00T, Pi0, OpenVLA) are designed so that an adaptation stage is needed before reliable deployment (reliable meaning with very high success rate) of the policy model on a specific combination of application environment and robotic task.

This adaptation stage, often called fine-tuning or post-training, caters to the narrow set of tasks, done in the exact deployment environment, atop a specific embodiment (such as single arm robots, or bi manual robots, or humanoids among others).

The recently released World-Action-Models (WAMs), while having improved generalization to novel tasks and environments, also need adaptation stage to reliably execute a given task on a given environment.

The key requirement of robotic adaptation is high quality adaptation data, typically human teleoperation data of the robot performing the required task in the given environment. Typically 50-200 examples of human teleoperation are required, depending on the base policy being finetuned, as well as the desired performance levels and the complexity of the task.

Thus, given any change in the environment or task where we want robot to perform reliably, there will be a need to collect such teleoperation data again and again.

Thus, robotic adaptation, which is the critical last mile of reliable robotic policy deployment - suffers from the data collection bottleneck - which needs to be eased.

Thus, an important research question needing solutions is -

What are effective approaches to ease human-teleoperation data collection effort required for robotic policy adaptation?

Show HN: Free Bible study that explains every chapter you read

https://learnofchrist.com
1•not_wowinter14•2m ago•0 comments

Nushell

https://www.nushell.sh/
2•handfuloflight•2m ago•0 comments

$550M in Medical Debt Only Cost $5.5M [video]

https://www.youtube.com/watch?v=Cimg-Xg043Y
2•foxfired•3m ago•0 comments

Wordsmith

https://segue.app/wordsmith
2•_continuation•4m ago•0 comments

Kimi K3 is not cheap

https://www.alexinch.com/blog/kimi-k3
3•ainch•5m ago•0 comments

New funding rules will make US research slower, costlier and harder to share

https://theconversation.com/new-trump-administration-rules-on-federal-science-funding-could-make-...
2•teaman2000•8m ago•0 comments

Look at Silicon Valley history from 360 degrees over the years

https://ValleyOfIdeas.com
1•chenpaul•9m ago•0 comments

Building Political Superintelligence

https://freesystems.substack.com/p/building-political-superintelligence
3•m-hodges•10m ago•0 comments

I've converted my website to grayscale

https://rz01.org/qmx/
2•exitnode•13m ago•2 comments

Optical Memory Link Could Boost AI in Robotics

https://spectrum.ieee.org/ai-in-robotics
1•jonbaer•13m ago•0 comments

Persistent Money Struggles Linked to Faster Brain Ageing

https://www.ucl.ac.uk/news/2026/jul/persistent-money-struggles-linked-faster-brain-ageing
1•karakoram•20m ago•0 comments

Show HN: Punter.js – single-file JavaScript game engine for teaching programming

https://punterjs.org/
1•john-doherty•21m ago•0 comments

LLM Cliche Highlighter

https://tools.simonwillison.net/llm-cliche-highlighter
2•saikatsg•22m ago•0 comments

America's schools admit screens make students worse off

https://fortune.com/article/schools-abandoning-chromebooks-laptop-programs-as-screen-time-hurts-l...
1•jedberg•25m ago•0 comments

Kitty [terminal emulator] chases the mouse

https://lwn.net/Articles/1080821/#Comments
2•leephillips•25m ago•0 comments

In a Ukrainian bomb shelter, Laura Loomer explains shift on Russia's war

https://apnews.com/article/laura-loomer-trump-ukraine-russia-interview-101852f6deab17e1d8d215fdde...
2•haburka•26m ago•0 comments

Please ship APIs, not AI

https://iamwillwang.com/notes/please-ship-apis-not-ai/
2•wxw•26m ago•0 comments

Illume Labs – Your 24/7 personalized health companion

https://www.illumelabs.ai/
3•parilatawa•27m ago•2 comments

OpenAI: A Bubble Bigger Than Dotcom

https://www.youtube.com/watch?v=zDtvrme-L-0
5•cable2600•28m ago•1 comments

Go server-less for projects that only need to scale from 2-200 users

https://seamside.com/docs/
2•akumpf•30m ago•1 comments

Iran used roaming systems, ad tech to track US troops in Gulf: Report

https://www.turkiyetoday.com/business/iran-used-roaming-systems-ad-tech-to-track-us-troops-in-gul...
5•billybuckwheat•31m ago•0 comments

BSD seq(1) on Linux

2•1vuio0pswjnm7•33m ago•0 comments

What Happened to Tagging? (2019)

https://daily.jstor.org/what-happened-to-tagging/
2•downbad_•33m ago•0 comments

OptMem: Plug-and-play infinite memory for any agent

https://github.com/VictorTaelin/OptMem
1•LightMachine•34m ago•0 comments

The bitter lesson for skills and coding agents

https://twitter.com/jim_dowling/status/2081455720336179313
1•jamesblonde•36m ago•0 comments

Doom running on nothing but regex find-and-replace

https://github.com/4RH1T3CT0R7/doom-regex
1•amichail•37m ago•0 comments

Learning two pointers from scratch after solving 100 LeetCode problems

https://leetcopilot.dev/blog/two-pointers-intuition-for-leetcode-beginners-step-by-step
2•alexwang24•38m ago•0 comments

I Made Open Source Adobe Acrobat – Windows 11

https://github.com/jonahgigi534-wq/PDF-Pilot
1•Jonathan_Gigi•38m ago•1 comments

We need to rethink tech stack

2•vivekk7190•40m ago•1 comments

Angels in Coptic Magic I: Introduction

https://www.coptic-magic.phil.uni-wuerzburg.de/index.php/2026/01/16/angels-in-coptic-magic-i-intr...
2•jruohonen•44m ago•0 comments