Hacker News: mqus

New comment by mqus in "Gold is now the top U.S. export"

mqus — Sat, 17 Oct 2026 19:56:38 +0000

From the article: > The U.S. government is still reporting that its gold reserves are unchanged, so these exports look to be coming from current mine output and recycling operations.

Not sure if "The U.S governments gold reserves" also includes the ones from other countries though.

New comment by mqus in "How the Claude Code team does code review"

mqus — Sat, 17 Oct 2026 12:27:37 +0000

Sure that happens, but over _years_. LLMs can make this happen over _days_.

New comment by mqus in "Redesigning source control at Anthropic"

mqus — Tue, 13 Oct 2026 19:55:09 +0000

They should concentrate on their models instead. If claude code is any measure, their software engineering is atrocious. I would only use something like source control from them if they force me to, at which point I would rather do something else.

New comment by mqus in "cp: -r or -R?"

mqus — Sat, 03 Oct 2026 11:11:26 +0000

would you be safe in using --recursive always? (e.g. shell scripts)

New comment by mqus in "What Sentry's 16GB selfhosted taught me"

mqus — Tue, 29 Sep 2026 20:45:12 +0000

completely agreed. Regardless of AI or not, the blog post completely lacks any kind of introduction and context at the start. I had to stop after the first two paragraphs because I just didn't understand anything.

New comment by mqus in "Evolving programming languages in the AI era"

mqus — Sun, 27 Sep 2026 11:42:03 +0000

I don't think any new languages will easily beat the existing ones, simply because of the mass of training data that is available. An LLM will have a much easier time one-shotting Java than even Rust today because it doesn't have to look up much and the ecosystem was stable for many years, so the internet is filled with content that is still up2date. Building a new language (or even altering existing ones) will take much longer to get into the models.

LLMs that have to constantly fix their code and look up libraries or features are much slower, more expensive and more prone to create inefficient and slightly wrong code. Sure, an LLM can just create a new web framework from scratch, but you usually don't want to spend your tokens on that.

I also think that with LLMs, new languages will have more trouble building an ecosystem, because until now, the popular libraries had a "proof of work" that kinda also made sure that they were well-trodden and maintained. Now, anyone can just push their new generated library and we have no clue on how to compare (and tbh reading LLM-generated READMEs is also not pleasant).

New comment by mqus in "ASML says it sold 'absolutely nothing' in Europe in 2026"

mqus — Sat, 26 Sep 2026 22:21:46 +0000

I hope my company (if I have one) can also blame a continent for not buying my product instead of trying to sell better :D

( /s)

New comment by mqus in "U.S. appeals court upholds designation of Anthropic as supply chain risk"

mqus — Sat, 26 Sep 2026 10:04:03 +0000

All of those verbs basically mean "degrade an existing system, after the fact". If anthropic states beforehand that certain use cases are not possible, it should be fine, no?

To give another (bad) analogy: The Pens aren't made to be weapons themselves, they are not intended for that. If the DoD now tries to shoot pens as projectiles and they just melt, that's no reason to classify the pen manufacturer as a supply chain risk, the product was not altered after the fact.

Of course, there _is_ a difference in "It could do that, but it shouldn't" and "It cannot do that". But it is not that large. That's why I don't see it as that clear cut.

New comment by mqus in "Italian parliament votes for return to nuclear energy"

mqus — Wed, 23 Sep 2026 23:31:12 +0000

> Hydro can be impacted by droughts which will be more common

You did read the news this summer? Like https://edition.cnn.com/2026/08/03/europe/hungary-nuclear-pl... ?

Most of the environmentalists will tell you that it's dumb to shut down nuclear, but building new plants just takes too long and is too expensive. Why then not just put the money into things that are cheap now and will get even cheaper tomorrow?

Hedging against downtime is comparatively easy, just diversify by location and improve the grid, which is way less expensive than building out power plants.

New comment by mqus in "Italian parliament votes for return to nuclear energy"

mqus — Wed, 23 Sep 2026 23:24:48 +0000

Where in comparable countries (= in europe) did new reactors take less than 10 years (from scratch even)? So we are at at least 10 years. But more realistically, taking recent megaprojects into account, its 20 years. In that time, batteries are way ahead and you can overbuild _so_ _much_ renewables and people even do it without subsidies (strangely, nobody builds nuclear plants without subsidies).

regarding the "stable energy" point: germany has one of the most (if not _the_ most) stable grids in the world and that is even with 50+ % renewables. And given it sits in the center of europe with grid connections all around, it's far cheaper to buy from whomever has the most favourable weather anyways (and sell to them if they don't). Thats why one of the biggest issues right now is not nuclear, but grids.

New comment by mqus in "UTF-8000: Unlimited UTF-8"

mqus — Sun, 20 Sep 2026 09:03:05 +0000

One more idea:

- Unicode-as-a-character, meaning: a whole text displayed as if it were one character, made for infinite zooming

New comment by mqus in "UTF-8000: Unlimited UTF-8"

mqus — Sun, 20 Sep 2026 08:57:04 +0000

2-3 bytes are not much space for anything. Sure, you could use multiple successive ones of these code points and define your own "continuation" encoding in these ranges, but that doesn't seem right to me somehow

New comment by mqus in "UTF-8000: Unlimited UTF-8"

mqus — Sun, 20 Sep 2026 08:37:36 +0000

Some ideas of what to do with this space:

- fully-customizable emojis (think of a RPG-like character customization screen)

- heck, why not full jpegs/gifs?

- some unicode programming script (running Doom)

- ?

That said, some very minor (HN-style) nitpick:

> Otherwise for an n byte code unit this is (5n+1) / 8n, that is 5n+1 content bits out of a total of 8n bits from n bytes. We can rewrite this as (5/8) + 1/(8n) which moderately quickly approaches 5/8 = 62.5%. It is nice that this limit is nonzero and does not depend on n.

Isn't a limit by definition no longer dependent on n?

CVE-2026-18963 – Keycloak Reset-Credentials Bypass → Account Takeover

mqus — Tue, 25 Aug 2026 22:35:16 +0000

Article URL: https://github.com/Red-Darkin/CVE-2026-18963-keycloak

Comments URL: https://news.ycombinator.com/item?id=49441617

Points: 6

# Comments: 0

New comment by mqus in "Most tech revolutions made work worse for employees"

mqus — Wed, 05 Aug 2026 21:22:28 +0000

not only inconsistent over time, but also inconsistently distributed over people. rich people got richer (at least lately) and the less rich didn't get much richer (if at all)

New comment by mqus in "Authorize, don't authenticate"

mqus — Fri, 31 Jul 2026 22:32:29 +0000

yeah, to me this also pretty much sounds like the approach we started with, we bring our data (as a file) into the application and it stores it back if we want to.

New comment by mqus in "A shell colon does nothing. Use it anyway"

mqus — Sun, 26 Jul 2026 12:33:10 +0000

Is : the same as `true` ?

New comment by mqus in "How the FSF sysadmins block botnets with reaction"

mqus — Tue, 14 Jul 2026 19:33:11 +0000

But... that software (reaction) is not written by the FSF? They just use it

New comment by mqus in "Interrail: 6,379Km and 13 Countries over 7 weeks"

mqus — Mon, 13 Jul 2026 12:53:45 +0000

> I guess it depends on what you expect. If you expect the train to be on time .. yes, India is much better than Germany. ~80% trains delayed 15 min or less vs Germany 36% trains delayed 15 min or less.

I won't defend DB too much here but saying that 64% of trains in germany are delayed more than 15min is just plain wrong. The official statistics put it at 40% for 6min or more for long distance trains and 11% for regional trains in 2025[1]. I think if there would be 15min statistics, DB would be on a sinilar level to India for long-distance. Not something to be proud of, but not as desastrous as you paint it. Though, maybe I misunderstood? Do you have sources for the indian numbers?

[1] https://eisenbahn.de/geramond-vgb/jahresbilanz-2025-puenktli...

New comment by mqus in "How we can reduce traffic congestion"

mqus — Sun, 12 Jul 2026 21:45:17 +0000

People hate patronizing systems in their cars already, not sure this would help