News for hackers. Stuff that matters.

kstonekuan

kstonekuan on Hacker News

Re:Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data... on Monday August 31, 2026 @11:53PM (#49517774)
Thanks for the advice, we'll definitely consider an example like this.
Re:Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data... on Monday August 31, 2026 @03:15PM (#49513739)
Yes, especially in a field moving as fast as robotics, users need the flexibility to extend the system quickly
Re:Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data... on Monday August 31, 2026 @02:09PM (#49512832)
Love those robots! Unfortunately, we don't support the training and deployment side yet as we are focused only on data processing for now.
Re:Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data... on Monday August 31, 2026 @01:32PM (#49512416)
We have had this same experience, and it's a big reason we chose to go open source and make it easy for teams to migrate their business logic onto our SDK instead of requiring them to adopt a new system. By targeting a more foundational layer, we aim to be a part of that "built it yourself" system instead of replacing it.
Re:Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data... on Monday August 31, 2026 @12:35PM (#49511713)
Our users process thousands of hours of data used in robotics model training, that's when they need scalable infrastructure to handle the workloads

Launch HN: Hebbian Robotics (YC S26) – Build scalable robotics data pipelines

Posted by kstonekuan on Monday August 31, 2026 @11:02AM
from the github.com dept.

https://github.com/Hebbian-Robotics/hflow

Hi HN, we’re Brandon and Kingston, the founders of Hebbian Robotics. We built HFlow (https://github.com/Hebbian-Robotics/hflow), an SDK that turns multimodal recordings from robots and human operators into standardized, quality-checked episodes and queryable dataset manifests. A recording can contain synchronized video, joint states, actions, timestamps, and metadata, and HFlow processes those...

( Read More... | 13 comments | 13 points | on HN )
Re:YouTube System Design for Robotics Data Infrastructure on Tuesday July 21, 2026 @02:08PM (#48995924)
What does YouTube system design have to do with robotics data?

When we started building Pareto, we weren’t thinking about YouTube, but the data path looked familiar: upload, process, store, index, play.

Here's what the classic system design question gets right in production.

- Read: https://hebbianrobotics.com/blog/youtube-system-design-for-r...

- Try Pareto: https://pareto.hebbianrobotics.com

- Code open source: https://github.com/Hebbian-Robotics/pareto

Show HN: Yoink functionality from dependencies and avoid supply chain attacks

Posted by kstonekuan on Saturday April 04, 2026 @11:13PM
from the show-hn dept.

https://github.com/theogbrand/yoink

yoink is a Claude Code plugin that removes complex dependencies from your codebase by reimplementing only the functions you actually use.yoink runs as a three-step, agent skills-based workflow:1. /setup clones the target repo and scaffolds a replacement package.2. /curate-tests generates tests verified against the original's expectations.3. /decompose determines dependencies to keep or...

( Read More... | no comments | 2 points | on HN )

Show HN: webskills – turn any webpage into an agent skill

Posted by kstonekuan on Sunday March 08, 2026 @11:20AM
from the show-hn dept.

https://github.com/kstonekuan/webskills

I built webskills, a CLI that turns any webpage into an agent skill.It first tries native `npx skills add` installation from a URL. If the site does not already expose an agent-ready surface, it falls back to document extraction to generate the skill locally.It’s built for pages that are useful to agents but are not yet packaged as skills: docs, pages, wiki/reference pages, help centers, specs,...

( Read More... | no comments | on HN )
Re:Anthropic says 'Claude will remain ad-free,' unlike ChatGPT on Wednesday February 04, 2026 @10:24PM (#46895280)
Thanks for linking!
Re:Web Viewer for Apple's ml-sharp with 3D Gaussian Splat Rendering in... on Monday December 29, 2025 @08:17AM (#46420402)
I built a web viewer for visualizing Gaussian Splat .ply files generated by Apple's ml-sharp, which converts a single photo into a 3D Gaussian Splat in under a second [1].

Features: - Upload and view .ply file directly in the browser - Multiple camera trajectory animations (rotate, swipe, shake, forward) - Interactive orbit controls (drag to orbit, scroll to zoom, right-drag to pan) - No installation required – runs entirely client-side

The original ml-sharp can render the same video trajectories but requires a CUDA GPU. This viewer lets you explore the 3D output on any device with a browser.

I also added cloud GPU inference via Modal so you can generate splats without a local GPU (free tier available) [2].

[1] https://github.com/apple/ml-sharp

[2] https://modal.com/

Re:Show HN: Open-source customizable AI voice dictation built on Pipecat on Wednesday December 17, 2025 @02:41AM (#46299252)
Tauri v2 has this global shortcut plugin which I am using and it works amazing

https://v2.tauri.app/plugin/global-shortcut/

Re:Show HN: Open-source customizable AI voice dictation built on Pipecat on Tuesday December 16, 2025 @10:21AM (#46289608)
I started off trying Electron, but I am liking Tauri more. It seems Rust has better support for system-level integration, like controlling audio and keys. Are there other alternatives you are exploring?

Older »