Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
stared
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
46 ms
·
841.
▲
by
stared
2y ago
It is a simple tool, yet it makes a day-and night difference when traversing directories.
842.
▲
by
stared
2y ago
It seems that people pay for Ghost, https://github.com/TryGhost/Ghost
843.
▲
by
stared
2y ago
Minecraft is LEGO bricks for the modern age - and, well, it is popular. As the blog post mentioned, "SpaceChem and Kerbal Space Program sold over 1 million and over 2 million copies on Steam." If you mean "putting a game in a
844.
▲
Science-based games and explorable explanations
(p.migdal.pl)
68 points
by
stared
2y ago
|
13 comments
845.
▲
Claude 3.5 asked to create a superstimulus for itself and humans
(twitter.com)
1 points
by
stared
2y ago
|
0 comments
846.
▲
by
stared
2y ago
In this topic, there is a wonderful summary of concrete examples of dos and don'ts: https://accessibility.blog.gov.uk/2016/09/02/dos-and-donts-o...
847.
▲
by
stared
2y ago
Omni (necklace) +2 Charisma +1 Intelligence A necklace with a helpful spirit living inside a silicon crystal. The spirit grows with every word it hears. "First you wear it, then it wears you"
848.
▲
by
stared
2y ago
From time to time, I see a tool to present a discussion as a tree with arguments for and against it. Unless it is a school essay, arguments don't go that way. It is usually harder to encompass what a node (an atomic fact) is and what a
849.
▲
by
stared
2y ago
I sense personal animosity and almost feel honored that it is behind a throwaway account. The talk is publicly available (in Polish) at https://www.youtube.com/watch?v=ChEsmwe7YN0 . Which part of it do you find objectionable
850.
▲
by
stared
2y ago
Looks interesting! Does it work with the Fish shell? And, in case, how do I update or uninstall it?
851.
▲
by
stared
2y ago
I understood all the words there. There is a big difference between jargon and, in general - presenting things using unnecessarily complicated language. Give this abstract to any editor, and I reassure you - they will rewrite it. Even a sim
852.
▲
by
stared
2y ago
My brain involuntarily assembled these letters into PTSD. If there is a single programming language that rewired my the brain, it's PostScript. I used that a bit two decades ago when I wanted to programmatically generate a variant of F
853.
▲
by
stared
2y ago
I think that this quotation captures the essence of the problem. It is nice to have the boring things automated - be it proofreading, generating mundane pieces of code, and stuff. The sad part In programming (at least, for now), I feel that
854.
▲
by
stared
2y ago
Thank you for sharing - especially with the green flag of "no account required"! I wanted to go habit tracing in a minimalist way, and ended up using checkboxes in Obsidian daily notes (and take them anyway, previously in Evernote
855.
▲
by
stared
2y ago
There is a lively stage of SC2 YouTubers, be they commentators (like LowkoTV) or gammers. Harstem is dark and edgy, including his iconic series "Is it IMBA or do I suck?". At the same time, it is lovely to watch challenges by MaNa
856.
▲
by
stared
2y ago
I am speaking of my experience, and I am an enthusiast for new things. I used the Jupyter Notebook before it was mainstream, or, say, PyTorch, in times when it was obvious that TensorFlow was the default option. However, in general, I belie
857.
▲
by
stared
2y ago
Thank you! If it is of public interest, I am happy to discuss it here. If you want to shoot me an email, that would be great. My path is in "What I do or: science to data science" ( https://p.migdal.pl/blog/201
858.
▲
by
stared
2y ago
I am a data scientist myself, one who moved from academia, vide https://p.migdal.pl/blog/2016/03/data-science-intro-for-math... . I used Jupyter Notebook before it was popular and back when I was a PhD student
859.
▲
by
stared
2y ago
https://www.cursor.com/ - an AI-first VS Code clone VS Code (and Cursor) has so nice Jupyter support that I find it much better to use it for my workflow, rather than using any dedicated solution for Jupiter Notebooks only.
860.
▲
by
stared
2y ago
Unless I miss something, having this type of local news on the Hacker News homepage is surprising. IMHO, it falls under the category of "boat programming" ( https://meta.stackexchange.com/a/14486/174092 ).
861.
▲
by
stared
2y ago
Thank you for bringing it up - it is a fair point that any set of axioms can be turned into a physical device. At the same time, it is not anything specific to the axiom of choice.
862.
▲
by
stared
2y ago
The (post) title mentions a Rust library, I see code in TypeScript, and the installation instructions involve Python. Frankly, I am confused. EDIT: To make it clear, my point is that the description could have been clearer, especially regar
863.
▲
by
stared
2y ago
Well, in principle, the Universe can be infinite. Sure, we cannot measure infinity, but to be fair, all mathematical concepts (when looked at closely enough) are not something we measure directly. Even if a kindergarten-level maths of "
864.
▲
by
stared
2y ago
"The Axiom of Choice is obviously true, the well-ordering principle obviously false, and who can tell about Zorn's lemma?" Well, the axiom of choice gives a lot of counterintuitive examples, with the Banach-Tarski paradox bei
865.
▲
by
stared
2y ago
So, I think that we differ at a fundamental level. While you prefer to work in a Neoplatonic world of ideas, I prefer empirical facts and convictions that all models are approximations. English grammar is not fixed per se; it evolves with r
866.
▲
by
stared
2y ago
Well, again, I'm not sure what your prompts are, but here we go: https://chatgpt.com/c/7ead1985-83be-4dce-9fd7-29aafb248f01 > The statistical model is confused by the superficial similarity between (a) and (b),
867.
▲
by
stared
2y ago
Well, prompting is not a nice addition to LLM; it is a necessary thing to do. > Nonetheless, native speakers agree on which interpretations it can and cannot have. GPT-4 does not have the same capacity, as far as I've been able to d
868.
▲
by
stared
2y ago
Thank you for providing these examples. I asked GPT4o, and it has no trouble with understanding 1a: https://chatgpt.com/share/daea469f-d823-45e2-9d6b-f6bea82a26... As a side note, my instinctive reading is on the telli
869.
▲
by
stared
2y ago
Attention is one of the core parts of transformer architecture, so I would be surprised if they have any trouble understanding this asymmetry. Could you provide a testable hypothesis? I would be happy to test it on GPT4.
870.
▲
by
stared
2y ago
I mean, contrary to Chomsky, that certainly advanced statistical models are able to capture all grammar - at least, nit worse than ordinary humans. It feels that Chomsky’s understanding of ML stopped at the level of Markov chains.
More ›