stabbles

5 ms·

31.
by stabbles 2mo ago
Your comment says he's stepping up, the HN title says he's stepping down, the article says he's stepping aside. Nobody says if it's a step forward or backward.
32.
by stabbles 2mo ago
In the Spack package manager [1] we're trying to work around this with directives: deprecated("@:1.3 +minizip", reason="cve", severity="high") So, you can define conditional deprecation, and the d
33.
by stabbles 2mo ago
Whether you agree with the policy or not, the way it's written is good. It just states the rules neutrally. See https://gcc.gnu.org/ai-policy.html . In various projects I see AI policies that state not only the rules, b
34.
by stabbles 3mo ago
Imagine this: OpenAI runs their benchmarks of unreleased models on systems physically disconnected from the internet.
35.
by stabbles 3mo ago
To be fair, many shells have command caching, so they effectively roll their own exec syscall wrapper: $ python3 -c 'print("hi")' hi $ hash hits command 1 /usr/bin/python3
36.
by stabbles 3mo ago
Indeed, it's a libc construct.
37.
by stabbles 3mo ago
Nice, PT_INTERP is the only non-relocatable thing of ELF files and typically requires wrapper scripts/executables. Regarding shebangs, I've never understood why the kernel cannot resolve e.g. `#!sh` relative to PATH instead of CWD
38.
by stabbles 3mo ago
Not really, a theorem with a hard proof can have simple but important corollaries. It's also not unthinkable that theorems exist with proofs that cannot reduce to something simple/short.
39.
by stabbles 3mo ago
Banana Paper was on the shortlist
40.
by stabbles 3mo ago
If you take a frame you see it's neither random nor dots: https://i.imgur.com/CgtyGjl.png From a single frame you can definitely identify boundaries because the dots are sliding and get truncated.
41.
by stabbles 3mo ago
So the image model's benchmark is to generate an image with the corresponding SVG sources.
42.
by stabbles 4mo ago
Nix hashes the build inputs, for which deterministic builds are not required, only desirable.
43.
by stabbles 4mo ago
Speeding up C/C++ compiler bootstrapping, starting at a single binary of <1KB. Currently it gets to GCC 4.7 in 2-3 minutes on x86_64 and aarch64: https://github.com/haampie/shpack
44.
by stabbles 4mo ago
This begs for a modern version of https://csszengarden.com/ , where the CSS is generated by different LLMs and prompts.
45.
by stabbles 4mo ago
Isn't that covered by O_CLOEXEC?
46.
by stabbles 4mo ago
A medicine for those who anthropomorphize LLMs is to run the LLMs deterministically (without randomness and memory files). It feels very unnatural to get the same conversation verbatim at a different point in time.
47.
by stabbles 4mo ago
David Hume pointed this out in 1740, and his advise still applies: A reason should be given, for what seems altogether inconceivable, how this new relation can be a deduction from others, which are entirely different from it. But as a
48.
by stabbles 4mo ago
I think this is fair criticism. It's hard to read this blog cause its premise is based on an "appeal to nature" fallacy.
49.
by stabbles 4mo ago
The Dutch version of this is to wave at the cheek.
50.
by stabbles 5mo ago
Is one day enough to find vulnerabilities? Who keeps an eye on new releases? Otherwise the problem continues to exist, just delayed by one day.
51.
by stabbles 5mo ago
Isn't it mostly the medium that's problematic? With an issue tracker it's easier to close as duplicate
52.
by stabbles 5mo ago
Isn't that rather difficult given the `.tar.gz` layers?
53.
by stabbles 5mo ago
You can look at a histogram of number of words per sentence, and you'll find immediately that it's written by an AI. When? Today. Minutes. Four years. $54 million. Collect feedback. Delivered faster.
54.
by stabbles 5mo ago
This comment is misinformed. Non-deterministic builds would also result in one tarball redistributed to all distro users. The ROP exploits don't work because of ASLR.
55.
by stabbles 5mo ago
Yeah, that suggestion made me roll my eyes. It's the wrong granularity, there's no build system support, it's inconvenient (executable wrappers? require the user to understand all transitive deps?). It also fails to mention g
56.
by stabbles 5mo ago
I noticed the same https://news.ycombinator.com/item?id=47940213 . My working hypothesis is that, given that a filter was always required (prs and issues are likely rows in the same database with a bool property to distingui
57.
by stabbles 5mo ago
What are you referring to when you say it's "fundamentally computationally inefficient"? It's pretty efficient because it's content-addressed, plus optimizations to reduce storage and data transfer with packfiles.
58.
by stabbles 5mo ago
If chess is solved and white wins, black is always in Zugzwang. We might not know.
59.
by stabbles 5mo ago
It's not necessarily an illusion. If chess is solved and it turns out white wins with perfect play, black's first move is zugzwang.
60.
by stabbles 5mo ago
This made me double check if it wasn't someone's vibe code scam website.

More ›