HNHacker News
TopNewBestAskShowJobs

jasonjmcghee

4,910 karma · joined March 25, 2015

Website: https://jason.today

GitHub: https://github.com/jasonjmcghee

submissionscomments
jasonjmcghee··on WebMCP Proposal
In response to microphone or camera access proposals you could have said "so we're going to let every website have a vector to spy on us?"

This is what permissions are for.

jasonjmcghee··on NPMX – a fast, modern browser for the NPM registry
Cynically, if you can attract a representative sample, you could aggregate and sell analytics data.

Another could be to have an "alternatives" section based on semantic similarity and / or some other features that have signal.

jasonjmcghee··on Beyond agentic coding
Yes please. There are many use cases where failure modes are similar to not using AI at all, which is useful.

Many very low risk applications of AI can add up to high payoff without high risk.

jasonjmcghee··on Steam Hardware: Launch timing and other FAQs
Look at this shadertoy to get a sense. It’s crazy.

https://www.shadertoy.com/view/4dsXzM

jasonjmcghee··on Data Processing Benchmark Featuring Rust, Go, Swift, Zig, Julia etc.
What's up with the massive jump for 20k to 60k for nearly all languages?
jasonjmcghee··on OpenClaw security assessment [pdf]
The account's stars are mostly a "system prompts" collection repo fwiw.
jasonjmcghee··on I have written gemma3 inference in pure C
I guess llama.cpp isn't quite as popular as I had assumed.
jasonjmcghee··on Clawdbot Renames to Moltbot
My twitter timeline was dominated by it for a few days and would see periodic star stats posted, but certainly didn't monitor the repo.

I've seen the author's posts over the last while, unrelated to this project, but I bet this had quite the impact on his life

jasonjmcghee··on Clawdbot Renames to Moltbot
I’m out of the loop clearly on what clawdbot/moltbot offers (haven’t used it)- I’d love a first hand explanation from users for why you think it has 70k stars. I’ve never seen a repo explode that much.
jasonjmcghee··on Kimi Released Kimi K2.5, Open-Source Visual SOTA-Agentic Model
> if you want to understand the effects of quantization on model quality, it's really easy to spin up a GPU server instance and play around

Fwiw, not necessarily. I've noticed quantized models have strange and surprising failure modes where everything seems to be working well and then does a death spiral repeating a specific word or completely failing on one task of a handful of similar tasks.

8-bit vs 4-bit can be almost imperceptible or night and day.

This isn't something you'd necessarily see playing around, but when trying to do something specific

jasonjmcghee··on Dithering – Part 2: The Ordered Dithering
Is it self-promotion or just "hey cool I care enough about this I built something too"

It's ok for people to get excited about shared passions

jasonjmcghee··on A macOS app that blurs your screen when you slouch
Except the form factors are swapped, but yes.
jasonjmcghee··on ANN v3: 200ms p99 query latency over 100B vectors
So many missing details...

Different vector indexes have very different recall and even different parameters for each dramatically impact this.

HNSW can have very good recall even at high vector counts.

There's also the embedding model, whether you're quantizing, if it's pure rag vs hybrid bm25 / static word embeddings vs graph connections, whether you're reranking etc etc

jasonjmcghee··on A macOS app that blurs your screen when you slouch
Yeah- this and the upcoming steam frame seem like the best options today.

There's something very attractive for me personally about the sunglasses form factor.

Safer in public, draws less attention, more portable, less headset fatigue, etc.

But obviously trading quality and features.

Also AVP is like $3k, steam frame will probably be $800+, xreal are like half that

jasonjmcghee··on A macOS app that blurs your screen when you slouch
I'm not sure how you can use a laptop with good posture. An external monitor at the right height seems like a necessity.

I'm also optimistic about monitors in the form of glasses- even less effort needed to set yourself up for perfect posture. But the sweet spot problem is still very much a thing from what I've seen- can't wait until it's normal for them to have eye tracking, foveated rendering and streaming, and be wireless.

jasonjmcghee··on Nobody likes lag: How to make low-latency dev sandboxes
If you don’t have control over all these pieces and just need terminal (ssh / remote vim etc), highly recommend https://mosh.org/
jasonjmcghee··on Show HN: Whosthere: A LAN discovery tool with a modern TUI, written in Go
Big missed opportunity to call it “Whose LAN is it anyway?”
jasonjmcghee··on Tree-sitter vs. Language Servers
Most of them are in the language pack (https://github.com/Goldziher/tree-sitter-language-pack)

For others, this is a sub optimal answer, but I’ve played with generating grammars with latest llms and they are surprisingly good at doing this (in a few shots).

That being said, if you’re doing something more serious than syntax highlighting or shipping it in a product, you’ll want to spend more time on it.

jasonjmcghee··on Vibecoding #2
Any time doing hand-rolled on demand spinning up of ec2 instances, be sure they are properly spun down later.

It's very easy to get hit with a massive bill due to just leaving instances around.

jasonjmcghee··on GLM-4.7-Flash
Fwiw Sonnet 4.5 is very far ahead of where sonnet was a year ago
jasonjmcghee··on The recurring dream of replacing developers
If you can build a luxury brand, sure.
jasonjmcghee··on The recurring dream of replacing developers
Many companies aren't selling anything special or are just selling an "idea".

Like liquid death sells water for a strangely high amount of money - entirely sales / marketing.

International Star Registry gives you a piece of paper and a row in a database that says you own a star.

Many luxury things are just because it's sold by that luxury brand. They are "worth" that amount of money for the status of other people knowing you paid that much for it.

jasonjmcghee··on Ask HN: How are you doing RAG locally?
How does performance (embedding speed and recall) compare to minish / model2vec static word embeddings?
jasonjmcghee··on Is Rust faster than C?
Thanks
jasonjmcghee··on Is Rust faster than C?
> Rust programmers may as well sprinkle threads all over the place regardless whether that's a 16x improvement or 1.5x improvement

What about energy use and contention?

jasonjmcghee··on Show HN: Ferrite – Markdown editor in Rust with native Mermaid diagram rendering
(not associated, just looked at the code - no js interpreter)

https://github.com/OlaProeis/Ferrite/blob/master/src/markdow...

jasonjmcghee··on What happened to WebAssembly
SIMD and multithreading support really helped with closing the performance gap.

Still surprised about the 5% though- I’ve generally seen quite a bit more of a gap.

jasonjmcghee··on A prediction market user made $436k betting on Maduro's downfall
Yeah I don't understand why it would be the case. How is it any different? It's trading on privileged information.
jasonjmcghee··on A prediction market user made $436k betting on Maduro's downfall
What are the laws around polymarket?

I've never used it, but continue to see posts about employees making bets on privileged information and that being legal.

jasonjmcghee··on 'Doomsday fish': Once-in-a-lifetime sea creature encountered in Monterey Bay
I was about to post the same comment- so just blatant click-bait misinformation by sf gate...
← PreviousPage 8 of 34Next →