wilkystyle

Unrelated to the content of the article, but what a wonderful site text.npr.org is. It looks like a slightly enhanced version of my iPhone's reader mode and it displays instantly, scrolling with no lag, stuttering, or dynamic loading of page elements.

I wish more of the web were like this.

It's a great point! The 2015 has been a linux machine for ~5 years, and the 2012 is my wife's laptop, used once every couple of weeks. Trying to get her to upgrade to a MacBook Neo for this exact reason.

Agreed with all of this! To date, agent-shell.el is my favorite AI coding experience.

Similar experience. In active use in my house:

    2012 MacBook Pro
    2015 MacBook Pro
    2019 MacBook Air
    2020 M1 MacBook Pro

The longevity is incredible.

> On the other hand, the "wheeled platform powered by a box of controlled explosions" doesn't get close to capturing the impact of the car on individual freedom, the environment, the economy, and more. So if someone was saying "there's nothing to talk about here, it's just a box of controlled explosions", we can safely ignore that person too.

This is also a good point, and I agree with you 100% here. I don't think one could responsibly argue that the impact and consequences are not significant, regardless of if they believe it's near AGI or it's just fancy next token prediction.

I think the point, with a more charitable reading of the parent comment, is that people forget the fundamental underlying technology and its inherent limitations.

"It's just a four wheeled platform powered by a box of controlled explosions" isn't a terribly useful description of a car until you get into a conversation with someone who thinks their car is depressed because it spends most of its time locked up in a garage.

> are you actively pointing your LLM at the kernel to identify security bugs?

I am not actively pointing my LLM at the kernel in the specific way that would uncover these bugs, and just because it's possible to do so doesn't mean that it has happened already.

> the policy wasn't made on a whim. to quote the linux security team [...]

Great, so let the ones that arise simultaneously do so. What does that have to do with immediately publicizing a vulnerability, regardless of severity, just because an LLM was involved in its discovery?

> the term responsible disclosure is a silly thing made up by vendors

That's certainly an opinion you're welcome to have, but unless you are proposing that responsibility and disclosure are two orthogonal concepts I don't see what that opinion has to do with anything I've said.

No. I have an LLM and didn't discover it. Folks in HR have LLMs but can't produce the same software you can. Just because an LLM could discover it does not mean that it has been discovered yet. Responsible disclosure is still a valuable thing.

Yes, and I did qualify my comment with "if you can use uv"

I took the parent's comment to mean "brutal" as in an absolute burn/roast.

If you can use uv, try the --exclude flag, e.g.:

    uv pip install -r requirements.txt --exclude excludes.txt

excludes.txt is basically like an inverse requirements.txt. Your requirements.txt could have this...

    pandas
    scikit-learn
    matplotlib

...while excludes.txt has this...

    numpy

... and uv would resolve the dependencies in requirements.txt, excluding numpy from the dependency graph entirely and ignoring any dependencies that would have been introduced by numpy itself.

EDIT: You can also configure this for uv directly in a pyproject.toml file:

    [tool.uv]
    exclude-dependencies = ["numpy"]

> in practice the MOE models with few active parameters are just faster and easier to run without too much loss of intelligence

Is this true for coding tasks? Genuinely asking because I don't yet have the hardware to be able to run/compare. I thought I remembered the sentiment that the 27B dense was better than the 35B MoE