Skunkleton's comments

On: CLI agents make self-hosting on a home server easier and fun

[Skunkleton]

Yes, that is really all it is.

On: Linux kernel security work

[Skunkleton]

In the context of the kernel, it’s hard to say when that’s true. It’s very easy to fix some bug that resulted in a kernel crash without considering that it could possibly be part of some complex exploit chain. Basically any bug could be considered a security bug.

[SSLy → Skunkleton]

plainly, crash = DoS = security issue = CVE.

QED.

[michaelt → SSLy]

BRB, raising a CVE complaining the OOM killer exists.

[pamcake → michaelt]

Memory leaks are usually (accurately) treated as DoS. OoM killer is a mitigation to contain them and not DoS the entire OS.

[worthless-trash → michaelt]

I could be wrong. But operation by design isn't considered a bug.

[samus → worthless-trash]

It is if some other condition is violated that is more important. Then the design might have to be reconsidered.

[suspended_state → worthless-trash]

If it is faulty, then it's not a bug, it's a flaw.

[lfllfkddl → suspended_state]

It is possible to design a security vulnerability.

[worthless-trash → lfllfkddl]

Oh, now that is an exciting area.

[SSLy → michaelt]

you either get OOMed or next malloc fails and that's also going to wreck havoc

On: Lessons from the PG&E outage

[Skunkleton]

> The spontaneous explosions become so common and normalized that just about everyone knows someone who got caught up in one, a dead friend of a friend, at least

That’s an extraordinary claim.

[pastel8739 → Skunkleton]

This is a metaphor; do you think it’s an extraordinary claim to make for traffic accidents or even traffic deaths? To me it isn’t, at all.

On: Vm.overcommit_memory=2 is the right setting for servers

[Skunkleton]

Over commit is a design choice, and it is a design choice that is pretty core to Linux. Basic stuff like fork(), for example, gets wasteful when you don't over commit. Less obvious stuff like buffer caches also get less effective. There are certainly places where you would rather fail at allocation time, but that isn't everywhere and it doesn't belong as a default.

On: 73% of AI startups are just prompt engineering

[Skunkleton]

The question that isn't answered completely in the article is how useful are the pipelines for these startups? The article certainly implies that for at least some of these startups there very little value add in the wrapper.

On: AMD GPUs Go Brrr

[Skunkleton]

This meme is tired. Let it rest boss.

On: How apps are using Liquid Glass

[Skunkleton]

Maybe I’m grumpy, but the old designs all look better and more functional to me.

On: Why aren't smart people happier?

[Skunkleton]

If you find yourself getting in trouble, maybe you are solving the wrong problems?

On: ChatGPT Atlas

[Skunkleton]

A browser using your keychain seems like the least questionable bit, if anything.

[Analemma_ → Skunkleton]

Right, but most browsers aren't owned by money-losing startups desperate for any bit of training data they can get their hands on as scaling taps out.

I really doubt OpenAI consciously wants my passwords, but I could absolutely see a poorly-coded (or vibe-coded, lol) OpenAI process somehow getting my keychain into their training set anyway, and then somebody being able to ask Chat-GPT 6, "hey, what's Analemma_'s gmail password?" and it happily supplying it. The dismal state of LLM scraper behavior and its support (or lack thereof) of adherence to best practices lends credibility to this.

On: Vapor chamber tech keeps iPhone 17 Pro cool

[Skunkleton]

No case is great. I’ve taken to slapping a screen protector on my phone with no case. Keeps me from feeling bad about setting it face down.