thangngoc89

Joined 11 years ago

Karma
1,369
Submissions
400
Member since
Nov 18
Email: hi [at] khoanguyen.me
View on Hacker News

Recent submissions

Comment · 10 months ago
I would recommend you to try uv for this.
Comment · 1 year ago
That's part of the experience. I guess?
Comment · 1 year ago
I remember in a video, Linus measured that the ground absorbs all the heat before it could even reaching the pool.
Comment · 1 year ago
I switched to restic (https://restic.net/) and the backrest webui (https://github.com/garethgeorge/backrest) for Windows support. Files are deduplicated across machines with good compression support.
Comment · 1 year ago
FYI: 1809/LTSC 2019 refer to the long term support build of Windows 10. Took me a while to figure this out.
Comment · 1 year ago
Also broken for me:

Safari 18.0 (20619.1.26.31.6), macOS Sequoia 15.0

Comment · 1 year ago
The script seems to be invoking yarn audit --json and does the templating.
Comment · 1 year ago
The README is kinda light on details. This is a utility written in Go that convert yarn’s audit file from json to Markdown for reporting as part of the CI pipeline.

I’m wondering if yarn’s audit is better than npm’s audit?

Comment · 1 year ago
Can a single JetKVM control multiple computers? I have several AI training machines that occasionally got stuck and need a power cycle. Most of the time this happened when I’m out of town so it has been annoying.
Comment · 1 year ago
Oh I forgot about that. I mostly work on private repo (free account) so that feature isn't available to me.
Comment · 1 year ago
Sure. Download the branch and use git.

   git checkout main
   git merge --squash pr_branch
   git commit -m "Commit message"
Comment · 1 year ago
"matching" is exactly how we do DNA sequencing right now. The current technology is called next generation sequencing (NGS), we multiply the DNA and perform matching digitally to construct the full DNA.
Comment · 1 year ago
The bottleneck on distributed GPUs training/inference is the inter-GPU connections speed. For a single node, it's doable because it utilized PCIe 4.0 connections. For a cluster, you need at least 50Gbps connection between nodes, which is expensive for cheap GPUs.
Comment · 1 year ago
Just curious, you didn’t use the e-invoice solution (載具)?
Comment · 1 year ago

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton