Hacker News Reader

rorytbyrneon Show HN: I spent 3 years building a wearable to understand my back pain

Does it depend on body shape/height/etc? Could “bad posture” for me “equal good” posture for someone with a longer back or narrower shoulders or something?

rorytbyrneon Ask HN: What are you working on? (August 2026)

I’m working on “Supabase for science”: 1-click deploy of scientific databases with validation, transformation, MCP, export-for-ML, and provenance built in.

https://amacrin.com

rorytbyrneon Call for proposals, designing new kinds of research organisations

I thought the HN community might have some ideas for this.

Unconventional org charts, beyond PI + postdocs + PhD students? New tech to enable fully-remote research collaboration? New approaches to funding? Under-appreciated ideas from neolabs/modern R&D?

rorytbyrneon Ask HN: Are you interested in building devtools/infra for science?

This is more on the ed-tech side right? I'm thinking more tooling to help a synthetic biologist design a protein, or a DARPA PM integrate the research activities of the labs they fund.

rorytbyrneon Ask HN: What Are You Working On? (April 2026)

Open Science Archive: open infrastructure for scientific databases, so that every field gets its own Protein Data Bank in 1-click.

Code: https://github.com/opensciencearchive/server.

Website: https://opensciencearchive.org/

Two demos:

https://pockets.bio

https://lingual.bio

I've got demos up and running (mirroring/extending PDB and GEO). Next I'm working on APIs with good AX, ML-friendly export, and an unified AI-driven UI that works for all scientific data types.

rorytbyrneon Books of the Century by Le Monde

Of course it's relevant to how it's seen today. French culture nurtured the author, a French publisher published it, and France didn't ban it while other countries did. This is all evidence that the book was well-liked in France when it was published, and there's no reason to think that would change over time.

If anything, it's surprising that English-speaking countries like it so much.

rorytbyrneon Books of the Century by Le Monde

Ulysses was first published in Paris during the 20 years that Joyce lived there.

>I thought only the US ranks it high

Joyce never even set foot in the United States... You could say this about The Great Gatsby, which US sources might rank in the top 5 compared to 46 in this list.

rorytbyrneon Ireland rolls out basic income scheme for artists

Why is "singing along" a relevant metric?

In Dublin's best music venues, nobody is singing along because it's brand new material from brand new artists. If you're singing along to well known songs in Temple Bar then I'm afraid you're missing some of the best music the city has to offer, in venues like Whelan's, Workmans, Sin É, The Grand Social etc.

rorytbyrneon How London became the rest of the world’s startup capital

Not sure what point you’re trying to make here in relation to startups.

rorytbyrneon Ask HN: What are you working on? (January 2026)

A domain-agnostic, open source scientific database.

“Protein Data Bank-in-a-box”

https://opensciencearchive.org

rorytbyrneon Ask HN: What are you working on? (January 2026)

A domain-agnostic, open source scientific database.

“Protein Data Bank-in-a-Box”

https://opensciencearchive.org

rorytbyrneon Why users cannot create Issues directly

Not which parts of this are sarcastic or not, but it was probably running for a few weeks. High variance on that estimate though. I was running 5+ Claude Code instances and a similar number of vim instances.

rorytbyrneon Why users cannot create Issues directly

I spotted Ghostty using 20GB+ memory a few days ago on MacOS (according to Activity Monitor). I went through all my tmux sessions, killed everything, it was still 20GB+ so I re-started Ghostty. If I see it happen again, I'll take some notes.

rorytbyrneon Resistance training load does not determine hypertrophy

Relatively speaking, no. Strength training (as opposed to hypertrophy) calls for fewer reps, around 5 per set.

Many people advise spending about a year doing more sets of fewer (~5) reps to build strength, and then switch to fewer sets of more reps (8-12) when you want to build muscle mass.

Point being, the idea of doing lighter weights until failure is already kind of there in body building wisdom.

rorytbyrneon Resistance training load does not determine hypertrophy

> heavy weights with few reps (common body builder wisdom)

It is strength training (not body builder) wisdom to use heavy weights with few reps. Hypertrophy (i.e. body builder) programmes usually call for 8-12 reps, which implies relatively low weights.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

> Who else do you suggest and why are they better?

The arbiters are just our colleagues, at the end of the day. The journal is just the organisational mechanism, one of many possible mechanisms.

For example, I follow a weekly reading list (https://superlab.ca) published by a group of motor control labs at Western University. Those people are my arbiters of quality.

I want to continue having arbiters, and I want it to be the same people (broadly speaking). I just don't want them to be organised around journals because journals are toxic and lead to concentrated power over scientific narratives.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

You pay them, they give your work a stamp of prestige that is mostly unrelated to the quality of your work.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

> At that point why even have a journal

Great question.

> the journal system isn't perfect but its the only real check we have left.

I wish I could agree but Nature et al continually publish bad, attention-grabbing science, while holding back the good science because it threatens the research programmes that gave the editorial board successful careers.

"Isn't perfect" is a massive understatement.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

Haha yes I jumped off in a very different direction. The points you raised are very much valid in the short-term. But longer term, I think journals charging authors for some kind of enhanced research presentation service is actually quite valuable, so the short-term negative effects might lead to a good outcome for the industry down the line - we hope.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

Yes exactly. Right now they are arbiters of quality but they shouldn't be, and the move towards Open Access is changing their role.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

Indeed, they are irrelevant. Right now they maintain an administrative monopoly over the peer review process, that makes them de-facto arbiters even if it's peers doing the work.

Journals should either become tech companies offering (and charging for) new and exciting ways to present scientific research, or simply stop existing.

rorytbyrneon Beginning January 2026, all ACM publications will be made open access

I disagree. We haven't chosen quantity over quality, we have decided that journals should not be the arbiters of quality. I think these new incentives are exactly what we want:

1. Journals want to publish lots of articles, so they are incentivised to provide a better publishing experience to authors (i.e. better tech, post-PDF science, etc) - Good.

2. Journals will stop prioritising quality, which means they will relinquish their "prestige" factor and potentially end the reign of glam-journals - Good.

3. Journals will stop prioritising quality, which means we can move to post-publication peer-review unimpeded - Good.

rorytbyrneon Ask HN: What Are You Working On? (December 2025)

https://opensciencearchive.org

I’ve started building a domain-agnostic scientific data archive, inspired by the Protein Data Bank. It handles the deposition, validation, curation, storage, and searching of scientific data, with plugin interfaces for domain-specific components.

The goal is to accelerate AI-for-science efforts by making it easy for scientific organisations to spin-up professional-grade data infrastructure in a matter of days. “PDB-in-a-box”.

rorytbyrneon Ask HN: What Are You Working On? (Nov 2025)

I made a HN-like discussion forum for (meta)science.

https://talk.amacrin.com/

Most (meta)science discussion is either fragmented on Twitter/Bsky, or a bit too formal. I thought a centralised place for deeper, casual discussions might be helpful, so I'm testing that theory.

Launched a few days ago, so it might have some rough edges. I'm considering making it user-invite-only soon, but for now it's fully open for signup. I'll also move it to its own domain once I think of a better name.

rorytbyrneon Libghostty is coming

In that case are scroll bars even helpful for determining position?

My terminal history is normally huge, and the log output would be some unknown percentage of the total scrollable history.

rorytbyrneon Magical systems thinking

Cybernetics is essentially control theory, and many of the methods in neural network research came from there. For example backpropagation can be traced back to Pontryagin in the 1950s.

Training a nonlinear system to behave in a way you want is the raison d’etre of optimal control theory.

But I wouldn’t say it’s the birthing place of neural networks, personally.

https://en.m.wikipedia.org/wiki/Backpropagation#History

rorytbyrneon AGENTS.md – Open format for guiding coding agents

Waymo uses a bespoke 3D data representation of the SF roads, does it not? The self-driving car equivalent of an AGENTS.md file.

rorytbyrneon Claude 4 System Card

When we use LLMs as agents, this errant behaviour matters - regardless of whether it comes from sci-fi “emergent sentience” or just autocomplete of the training data. It puts a soft constraint on how we can use agentic autocomplete.

rorytbyrneon Gemini 2.5 Flash

LLMs are literally stochastic, so the point is the same no matter what the example application is.