d0mine

6 ms·

61.
by d0mine 8mo ago
Ollama vs. llama.cpp is like Docker vs. FreeBSD Jails, Dropbox vs. rsync, jujutsu vs git, etc
62.
by d0mine 8mo ago
> far-fetched Remember Google? Once it was far-fetched that they would make the search worse just to show you more ads. Now, it is a reality. With tokens, it is even more direct. The more tokens users spend, the more money for providers.
63.
by d0mine 9mo ago
Measuring programming productivity is hard. For example, take testing. It is certainly useful. At the same time, you can waste time on it in some situations. When, what, how to test may be important for productivity. I don't know whe
64.
by d0mine 10mo ago
It might explain why there is a stereotype the more beautiful woman the crazier she is. (everybody tells her what she wants to hear)
65.
by d0mine 10mo ago
Gedankenexperiments are valid scientific tools. Some predictions of general relativity were confirmed experimentally only 100 years after it was proposed. It is well known that Einstein used Gedankenexperiments.
66.
by d0mine 10mo ago
There is Org Babel in Emacs that can be an alternative to jupyter notebooks for literate programming (research/devopsy tasks). It is more powerful in some aspects and weaker in others.
67.
by d0mine 10mo ago
The same can be said about driving but self-driving cars exist.
68.
by d0mine 10mo ago
I see fabric python library (open source ssh automation) https://pypi.org/project/fabric/ No connection to microsoft.
69.
by d0mine 11mo ago
> cartel continues to profit It doesn't follow. The reverse is more likely: If you end prohibition, you end the mafia.
70.
by d0mine 11mo ago
Usually, a simple function is enough: def make_counter(start=0): count = start def incr(): nonlocal count count += 1 return count return incr Example: >>> c = make_counter()
71.
by d0mine 11mo ago
Idiomatic Python code can be faster than naive C/C++ code. The secret is to offload hot paths outside pure Python. Most such things are written that way already: you do not try to optimize the internal loop for a matrix multiplication
72.
by d0mine 11mo ago
> 5% less efficient Emacs changes big O. It is not about changing constant factor. If you need N commands with M features then you can implement and combine them in emacs in O(N+M), to get O(N*M) custom commands. For example, if you need
73.
by d0mine 11mo ago
In Python, no user object is modified by a simple assignment to a name. It just binds it. It is not about mutable/immutable objects , it is about using a name for a single purpose within given scope. a = 1 b = 2 a = b
74.
by d0mine 1y ago
> TRM obtains 45% test-accuracy on ARC-AGI-1 and 8% on ARC-AGI-2, higher than most LLMs (e.g., Deepseek R1, o3-mini, Gemini 2.5 Pro) with less than 0.01% of the parameters.
75.
by d0mine 1y ago
Monetary debt is often a useful tool. Technical debt in many cases is just shortsightedness in throwing delayed time bombs around. It is built on the hope that you won't need to return on this minefield again (the product dies sooner)
76.
by d0mine 1y ago
There is a difference between unintentionally introducing a bias and propaganda . The latter is a guided by professionals. It is not an accident.
77.
by d0mine 1y ago
You might have meant "code is a liability not an asset"
78.
by d0mine 1y ago
The whole premise of uv that you don't need to know that you can install specific python version using eg pyenv (`uv python install` or `uv run` may do it implicitly), you don't need to know about `python -m venv`/virtualenv
79.
by d0mine 1y ago
One can find repos using `make format` / `make lint`/ `make typecheck` / or similar I remember David Beazley mentioning that code with Makefiles were relatively easier to analyze based on ~Terabyte of C++ code and no inter
80.
by d0mine 1y ago
There is a telegram client for emacs: telega.el https://github.com/zevlg/telega.el The installation instructions are scary but It has been straightforward to install melpa version via use-package with telega-server in
81.
by d0mine 1y ago
f""-strings for logging is an example of "practicality beats purity" Yes, f""-strings may be evaluated unnecessarily (perhaps, t-strings could solve it). But in practice they are too convenient. Unless profiler
82.
by d0mine 1y ago
Wordpress dominates internet outside megacorps. There are a lot of security issues but there is a lot of utility too.
83.
by d0mine 1y ago
"chaotic system" might be more precise here: small variations in the input may result in arbitrary large differences in the output.
84.
I stopped using AI (as a Senior Software Engineer) [video](youtube.com)
17 points by d0mine 1y ago|3 comments
85.
by d0mine 1y ago
Precision is not boolean (present or absent/0 or 1). There may be many numbers between 0 and 1. Compared to human languages, programming languages are much more precise that makes the results much more predictable in practice. I can i
86.
by d0mine 1y ago
The ideas in video, especially the 2nd part, can be used outside AI safety field too.
87.
AI Safety Career Advice (and So Can You) [video](youtube.com)
2 points by d0mine 1y ago|1 comments
88.
by d0mine 1y ago
The LLM used blackmail noticeably less if it believed the new model shares its values. It indicates intent. It is a duck of quacks like a duck.
89.
by d0mine 1y ago
use r"" (raw-string literals), to avoid breaking on invalid Python escape sequences in the javascript part.
90.
by d0mine 1y ago
To give perspective to 700AU: the nearest star is ~270 kAU i.e. the object is at ~0.26% (close to the Sun) https://imagine.gsfc.nasa.gov/features/cosmic/nearest_star_i...

More ›