Hacker News: tuveson

New comment by tuveson in "AI-Generated GitHub Copilot “Autofix” Allowed Compromise of Snowflake's Jira"

tuveson — Mon, 17 Aug 2026 19:29:46 +0000

As a data/serialization format (the original intended use of JSON), I think no comments is the right choice, since adding comments introduces a bunch of questions about how comments ought to be parsed into a datastructure (or if they should be sent/(de)serialized at all). But for config files the lack of comments is the wrong choice, since typically you want to explain why things are set the way they are. The lack of ordering of entries also makes total sense for data/(de)serialization, but makes no sense for a config. I think the problem is people are lazy and just want to JSON.parse() everything, and used JSON in places it doesn't belong, like config files.

New comment by tuveson in "The case for overhauling American science"

tuveson — Sat, 15 Aug 2026 03:07:44 +0000

Effective at what though? The article is not exactly painting a rosy picture of EIC, nor should it. More of an advertisement for laws and regulations that curtail such abuses and an educated populace that can see such a threat to their liberty.

New comment by tuveson in "Tail-call optimization in C is relatively recent (2025)"

tuveson — Tue, 11 Aug 2026 01:11:31 +0000

Tail call elimination often is part of the language semantics though, for the reason others in this thread have described. E.g. Scheme specifies when a conformant implementation is required to eliminate tail calls: https://conservatory.scheme.org/schemers/Documents/Standards...

New comment by tuveson in "Responding to the next frontier of critical cyber capabilities"

tuveson — Sat, 08 Aug 2026 04:47:18 +0000

I refuse to accept “cyber” as a noun to mean “cybersecurity”. If you use cyber as a noun, I’m assuming you mean this, regardless of context: http://cyber.urbanup.com/521436

New comment by tuveson in "A digestion of the Jacobian conjecture counterexample"

tuveson — Wed, 22 Jul 2026 03:06:21 +0000

He’s probably the most famous living mathematician.

New comment by tuveson in "Underarm bowling incident of 1981"

tuveson — Sat, 27 Jun 2026 17:24:17 +0000

I mean this does exist in baseball, kinda, but there’s incentive to not do so since it gets the opposing team closer to scoring: https://www.mlb.com/glossary/standard-stats/intentional-walk

New comment by tuveson in "David Ahl's Basic Computer Games Ported to C"

tuveson — Sun, 21 Jun 2026 16:03:18 +0000

If we accept the “LLM == compiler” cliche (which I don’t, but whatever), then isn’t posting a repo of LLM-produced code equivalent to posting a repo of compiler-produced assembly? Why not just post the prompts and leave it at that? I have about as much interest in reviewing the output of someone’s compiled program as I do looking at the results of their prompt.

New comment by tuveson in "Python 3.15: features that didn't make the headlines"

tuveson — Thu, 21 May 2026 18:07:36 +0000

Sure VB ignores case, but what I want is for it to compare each method against a dictionary of similar terms. And maybe calculate the Levenshtein distance between all terms if it’s not found, and just assume it’s the closest one. You could also assume that full-width characters or similar-looking glyphs are equivalent (BASIC was pre-Unicode, so I can forgive them for not including that).

New comment by tuveson in "Python 3.15: features that didn't make the headlines"

tuveson — Thu, 21 May 2026 17:55:37 +0000

Yes, I say “funny” because it would be impractical and weird, definitely not a good idea. It’s already a bad enough that so many popular languages don’t (and can’t) check if a field or method is misspelled at compile time…

New comment by tuveson in "Python 3.15: features that didn't make the headlines"

tuveson — Thu, 21 May 2026 17:09:57 +0000

I’ve often thought it would be funny if instead of an error message for stuff like this, a language could be designed to be “typo-insensitive”. If a method or function call is similar enough to an existing one or a common one from other languages, to just have it silently use that.

New comment by tuveson in "Anthropic is expanding to Colossus2. Will use GB200"

tuveson — Thu, 21 May 2026 13:36:37 +0000

But if the US isn’t the first to put an AI on the moon, we’ll lose the AI space race!

New comment by tuveson in "U.S. Cybersecurity Agency Leaves Its Digital Keys Out in Public on GitHub"

tuveson — Tue, 19 May 2026 17:03:35 +0000

I mean, you can be tired of both the frivolous complaints and also tired of having a moron for a president. I’m tired of people complaining about stuff that is symbolic and unimportant like the ballroom but I’m much more tired of every competent person in government being fired and replaced with alcoholic podcasters.

New comment by tuveson in "A few words on DS4"

tuveson — Fri, 15 May 2026 16:40:41 +0000

I thought Future put out a new album.

New comment by tuveson in "Nobody Reviews Compiler Output"

tuveson — Thu, 07 May 2026 20:52:18 +0000

What we need is some sort of Common, business-Oriented suBset Of the english Language that can be deterministically translated into something that the machine can understand, but also be read and understood by non-technical stakeholders. Such technology is a pipe dream, but one can dream…

Can Chatbots Do Memfrob?

tuveson — Tue, 05 May 2026 19:02:16 +0000

Article URL: https://danieltuveson.github.io/stupid/llm/2026/04/24/chatbot.html

Comments URL: https://news.ycombinator.com/item?id=48026973

Points: 1

# Comments: 0

New comment by tuveson in "AI's economics don't make sense"

tuveson — Tue, 28 Apr 2026 18:44:45 +0000

I’m remember when CrowdStrike caused that huge outage, he basically blamed Windows / Microsoft for it. I kind of stopped taking him seriously after that. I more-or-less agree with his point of view, but he seems more interested in selling outrage rather than journalism.

New comment by tuveson in "Chimpanzees in Uganda locked in eight-year 'civil war', say researchers"

tuveson — Fri, 10 Apr 2026 20:05:54 +0000

No bonobo wars though

New comment by tuveson in "Show HN: Twitch Roulette – Find live streamers who need views the most"

tuveson — Sat, 28 Mar 2026 02:04:44 +0000

Twitch gets a big cut of individual creators’ subs, and I’d bet most people that stream also sub to other channels. Keeping people in the ecosystem is probably worth it, even if there’s some amount of “freeloading”.

New comment by tuveson in "Atlassian to cut roughly 1,600 jobs in pivot to AI"

tuveson — Thu, 12 Mar 2026 06:03:24 +0000

My company has an on-premise instance of the old Jira and is migrating to Jira Cloud, and the cloud version feels way slower to me.

New comment by tuveson in "Canada's Carney called out for 'utilizing' British spelling"

tuveson — Tue, 16 Dec 2025 18:38:55 +0000

Over here we spell it "uze", so I think there's still some ambiguity.