rorytbyrne

9 ms·

31.
by rorytbyrne 1y ago
LLMs are literally stochastic, so the point is the same no matter what the example application is.
32.
by rorytbyrne 2y ago
Apple Watch does this.
33.
by rorytbyrne 2y ago
We would need to normalise query length by the success rate to draw any informative conclusions here. The rate of immediate follow-up queries could be a decent proxy for this.
34.
by rorytbyrne 2y ago
He thinks the IQ of white people (North America and Europe) is 100. He’s implying the rest of the world are lower IQ.
35.
by rorytbyrne 2y ago
A fair point, but I think the "Old English" people (Hiberno-Normans) are not usually considered part of the protestant ascendancy.
36.
by rorytbyrne 2y ago
The “Protestant Ascendency” are not Irish, even if they technically lived and were born on the island.
37.
by rorytbyrne 2y ago
I grew up on a farm near the hills of The Burren in the west of Ireland. If you look closely, you can see walls made of stacked stones cross-crossing the hills, as if to demarcate farmland - but those hills are not arable farmland, so why?
38.
by rorytbyrne 2y ago
I highly recommend taking a dynamical systems class, if you ever get the chance. It's shocking how many seemingly unrelated things just "make sense" when you know what a limit cycle is and how bifurcations work.
39.
by rorytbyrne 2y ago
Thank you! The immediate benefit is that "what commit produced my last good result" is a local problem, so a local solution is faster and more convenient than a round-trip to an API. But more broadly, storing data locally makes it
40.
Show HN: Logis – a searchable scientific log in your Git commit history(github.com)
6 points by rorytbyrne 2y ago|3 comments
41.
by rorytbyrne 2y ago
Does anyone know of tools that go the other direction? i.e. taking a technical writeup (scientific paper, architecture docs, or similar) and emitting a candidate codebase.
42.
by rorytbyrne 2y ago
This seems useful for building new tools. It's not strictly an end-user tool.
43.
by rorytbyrne 2y ago
Oh true, forgot about that. You need someone to invite you IIRC.
44.
by rorytbyrne 2y ago
Missing lobste.rs
45.
Show HN: Cyantic – build complex objects from simple blueprints with pydantic(github.com)
2 points by rorytbyrne 2y ago|1 comments
46.
by rorytbyrne 2y ago
I do neuro/ai research, so I have to instantiate a lot of Tensors from YAML definitions (that's how I define models declaratively). I built a middleware to simplify this, using intermediary pydantic models as "blueprints&quot
47.
by rorytbyrne 2y ago
No, they are designed to elucidate the author's thought process - not the reader's learning process. There's a subtle, but important difference. Also: https://web.stanford.edu/class/ee384m/Handouts&#
48.
by rorytbyrne 2y ago
But if they want to capture (or create, commercially) the market they’ll probably price it closer to $20 or under.
49.
by rorytbyrne 2y ago
As a working scientist, this is a hugely underserved category in academic/industry science. This would be very welcome.
50.
by rorytbyrne 2y ago
The green card exists, so there must a priori be a valid path to get one. If it's not H1-B, then what is it?
51.
Show HN: Labfile, a declarative file format for orchestrating experiments(github.com)
3 points by rorytbyrne 2y ago|0 comments
52.
by rorytbyrne 2y ago
You could continually learn a distribution over AI responses and search for outliers to surface with urgency for approval.
53.
by rorytbyrne 2y ago
> the 90s IDM/jungle/trance is superior to modern EDM/house/trap This is your own fault for going to EDM/house/trap parties. Go to better parties, then you'll find better music.
54.
by rorytbyrne 2y ago
It doesn't take the whole repo as context, it tries to guess which files to look at. So if you prompt with that in mind, it works well. Haven't tried it on a very large codebase though. You can also explicitly add files, if you kn
55.
by rorytbyrne 2y ago
Just work on a branch and squash.
56.
by rorytbyrne 2y ago
Implementation quality should always be high, that is standard practice for engineering disciplines.
57.
by rorytbyrne 2y ago
MVP is a good approach when interpreted correctly: “minimum” is in reference to the feature set, not the implementation quality.
58.
by rorytbyrne 2y ago
Or a professional politician proposed a good policy and the professional political commentator made a good comment on it? Not everything is a conspiracy.
59.
Research studios, an idea bridging research labs and software engineering(twitter.com)
3 points by rorytbyrne 2y ago|0 comments
60.
by rorytbyrne 2y ago
"X is a new version of Y" is a really uninformative way to introduce a project. It's especially annoying when I click the link to Y and it just says "This repository contains the source code for Y".

More ›