vanschelven

It's a pretty big part of their public image not to be but the fear is real

Surely not _all_ the courses... Utrecht was and is big on Haskell as you know... Given that you TA'd a course on it :)

It's not strictly true, but the distinction between the truth and the assertion is small enough that the ICC itself draws the conclusion that Microsoft didn't yet:

https://www.techradar.com/pro/the-international-criminal-cou...

And here's John Carmack on the subject of 1000s of lines of code in a single function: http://number-none.com/blow/blog/programming/2014/09/26/carm...

in the same vein as http://developerexcuses.com/ (and presumably many others)

This writing made me curious enough to click the provided citation though, and I'd have to say "monkeys" is really being kind if we take into account the combination of code quality issues, lack of surrounding process, and _what_ these code quality issues are affecting (the criticalness of the path).

It's a shame that the actual attack mechanism doesn't seem to be detailed on the github repo, and the link to the article is dead.

"when an LLM costs less, works faster and makes less mistakes"... indeed, but it doesn't follow at all that it's the fate of all programmers _now_... at least in my experience none of these things are true ATM.

"it can deal"... with 300 events/s easily.

I personally find "something like sentry" (but better) so easy to set up that I use it even in local development. So definitely at launch...

I wrote about it here: https://www.bugsink.com/blog/local-development/

You might want to reconsider ”no data ever leaves your machine" from the post :)

Given your target market, have you considered looking at Bugsink?[0] sentry compatible. still not local, but at least you won't have to additionally ask your customers to trust sentry/posthog.

Disclosure: that's me

[0] https://www.bugsink.com/

I do take this seriously (see link in my profile), but have to admit that doing so isn't an automatic entry tickets to places like the DoD

Both those things can be true at the same time: all software used to ship like that, and shipping like that in the current era _is a big deal_.

Spoken as someone who's core differentiator is "self hosted" [0]

[0] https://www.bugsink.com/

HN negativity strikes again.

Can't we just read this as "there are 2 wins here: security and performance"?

Which is not surprising, but still a GoodThing(TM) right?

To take matters in my own hand I wrote "tabdouse", a tiny kill-at-high-cpu script: https://www.bugsink.com/blog/tabdouse/

Quite unreliable and more a joke than anything, of course...

Do you mean that specifically a hulucibated text "copyright by not-meta" made it into evidence? Or are you talking about copyright generally?

I wonder if hallucinated copyright claims (esp. like the ZDF one at the bottom of the OP) will be introduced as evidence in one of the court cases against "big AI"

This reads like yet another attempt to pathologize perfectly reasonable criticism as some form of oppression. Calling “AI could have written this” a classist slur is a stretch so extreme it borders on parody. People say that when writing lacks originality or depth — not to reinforce some imagined academic caste system. The idea that pointing out bland prose is equivalent to sumptuary laws or racial gatekeeping is intellectual overreach at its finest. Ironically, this entire paper feels like something an AI could have written: full of jargon, light on substance. And no, there’s no original research, just theory stacked on theory.

I did this basically as a joke but I'm sure security researchers will be able to look at this post and somehow extract an attack vector from it :-)

Which makes sense because jq, with no options, acts as a formatter by default. (it's about 50% of my jq usage).

the OP wrote "This is a simple concept, but I've never seen it written down before." Putting on my tin-foil hat:

Maybe that's because predictable log volume isn’t in the vendor’s interest. Time-based logging makes usage easier to reason about. Bursty, count-based logs? Much harder to estimate—much easier to monetize.

also here, I see now:

https://wikipediocracy.com/forum/viewtopic.php?t=14242

With the link leading to a 404 (taken down?) this seems a bit hard to confirm... or am I missing something?

Skimming the graphs one would say "it hardly matters" which matches the naive assumption that any advantage would be "arbitraged away" by people spending their time researching such matters.

Missing from the article — which is funny, considering it's written from the perspective of a university student — is how deliberate Sun’s academic strategy was. In 1998 they launched the "Authorized Academic Java Campus" program, licensing Java tech to universities and setting up official training centers. Even before that, they were offering Java tools free to schools for teaching and research.

Combined with a massive branding push — Sun doubled its ad spend from 1995 to 1997 — Java ended up everywhere in CS education. By the late ’90s, first-year courses using Java weren’t a coincidence; they were the result of a planned, top-down push.

Pointing out the dangers of various tools is something that's is consider "a point" to the degree that there is a special government agency for it in every modern state in the world. This is that, but applied to a specific tool and as a meme-site.

Dan Luu quoting Joel Spolsky:

“Find the dependencies — and eliminate them.” When you're working on a really, really good team with great programmers, everybody else's code, frankly, is bug-infested garbage, and nobody else knows how to ship on time.

[..] We didn't think that everyone else was producing garbage but, we also didn't assume that we couldn't produce something comparable to what we could buy for a tenth of the cost.

from https://danluu.com/nothing-works/

If you feel Sentry, PagerDuty and full observability is overkill you might be interested in Bugsink[0].

I specifically wrote about the "overkill" bit (see the blog) and slack notifications were (finally) just introduced.

Disclosure: that's me.

[0] https://www.bugsink.com/

I'm thinking of "something like this" myself, would be happy to discuss on GitHub/Discord whatever:

https://github.com/bugsink/bugsink/issues/127

> In house or 3rd party it doesn't matter

It does matter from the perspective of privacy, because in one case your vistors deal with one party (you) and in the other case another party (Sentry) is also involved.