KajMagnus

Saying the model failed to write a competitive C compiler makes more sense.

I don't think they tried to do that though.

> today's models are not yet able to produce production software without close supervision, even when uncharacteristically good specs and hand-written tests exist.

That's a good point anyway

Or when you find some old source code at your workplace, and you're like: "this looks pretty nice, I wonder who wrote this?"

Check out Qubes OS and vault VMs and (I haven't started using yet) split SSH / GPG.

If a package developer maliciously breaks everyone's builds,

isn't that pretty great?

Because now you have learnt that you can't trust them

How do you know the hash in advance?

My impression is that they're sometimes unemployed people or students hoping to create a popular open source project, and use it to find a job.

They aren't going to care about any of the advice in the article about not posting slop -- finding a job is (of course?) more important to them.

Can't really say they are doing anything wrong, maybe I too would have? ... Just that large scale, doesn't work

The number is 9

9% not accurate answers, for Gemini 3. (I think that's a lot!)

Have you seen Talkyard Blog Comments? It's open-source and based out of Europe. https://blog-comments.talkyard.io. (I'm developing it.)

> legacy WP blog ... users should be able to ...

(What's your blog about?)

> suddenly have 215 new files to check in!

How big is your repo, if I may ask?

Personally I store vendored dependencies in a submodule, where I can squash history, if it grows too large.

Good points. But what do you mean with 3: "lockfile poisoning attacks, by making them more complicated" — making the lockfiles more complicated?

Also, 4) Simpler to `git diff` the changes, when you have the source locally already :- )

Scala could be one example? When I upgraded to a newer version of the standard library (the Scala 2.13 or Scala 3 collections library), there was a tool, Scalafix [1], that could update my source code to work with the new library. Don't think it was perfect (don't remember), but helpful.

[1] https://scalacenter.github.io/scalafix/

Then you might find Talkyard interesting: https://blog-comments.talkyard.io — reminds of Disqus, in that it's threaded, best first (optionally). (I'm developing it.)

> showing some CORS error

In my case, I found it annoying when cookies gradually stopped working, and eventually I had to make the software use custom HTTP headers instead of cookies.

> Seems that mostly spammers comment

The more interesting the contents of the blog is, the more real humans will like it and post comments? (if they can find it)

But a "Our company posts something each day, even if nothing has happened" blog, or AI fluff, attracts only spammers?

Could one say that humans are trained very differently from AIs?

If we (humans) make confident guesses, but are wrong — then, others will look at us disappointedly, thinking "oh s/he doesn't know what s/he is talking about, I'm going to trust them a bit less hereafter". And we'll tend to feel shame and want to withdraw.

That's a pretty strong punishment, for being confidently wrong? Not that odd, then, that humans say "I'm not sure" more often than AIs?

Yes, many _humans_ here hallucinate, sort of.

They apparently didn't read the article, or didn't understand i, or disregard from it. (Why, why, why?)

And they fail to realize that they don't know what they are talking about, nevertheless keep talking. Similar to an over confident AI.

On a discussion about hallucinating AIs, the humans start hallucinating.

@imiric

> Though I would suggest a general revision of the copy on this page, and removing all the "later" features.

Now done — all "Later: ..." features are gone, and I did a general revision of the copy on the pricing pages, based on what you wrote, and based on Gemini's and ChatGPT's feedback (when I copy-pasted the HTML into the AI chats — they had many pages feedback).

> These days LLMs can also be helpful with this

They were amazingly helpful. Thanks for suggesting! I'll ask them to review the whole website, like you suggested, some time later too.

> It comes off as dishonest to mention features that may be implemented at some point in the future

Thanks, good to know. (Will fix)

> a lot of asterisks and clarifications that might confuse users

Yes! People email and ask about the pricing more often than I thought.

> It's also mentioned that a credit card is not needed for the trial, but then you will email users about how to pay. Maybe this should be clarified on this page?

Hmm. Yes, there could be something about "if you want to continue after the free trial".

> It seems that you're not a native English speaker

That's right

> but I would suggest hiring someone who is, and has experience in writing technical documentation to revise your entire site, so that it can have a more professional appearance. These days LLMs can also be helpful with this

Ok :- ) First, all the LLMs, then a human I guess.

Thanks for all the ideas. Sorry for the late (8 days) reply.

Just started reading The E-Myth Revisited, hopefully that, plus the sleepiness problem being mostly gone, can put things in better order.

> I like the distinction between liking and disagreeing

Me too :- ) I'm thinking that if something gets 10 upvotes, and sth else get 10 upvotes and 10 disagree votes, then more often it's the latter that's more interesting to read?

There's also an Unwanted vote, for things that are too off-topic or rude etc, similar to how downvotes work here at HN.

Slashdot's vote system I like too :- ) (Maybe the vote system could be pluggable in the distant future.) — There's also half implemented Do-It votes and Do-Not votes, for ideas and upcoming Joint Decision topic types.

> commets ... and the links under "Old plans" return 404

Oh, thanks! Hadn't noticed. They work from here: www.talkyard.io/pricing (the www subdomain), but from the blog-comments and education subdomains, all those "Old plans" links are broken. (So I'll need to point the links to the www subdomain.)

> I would suggest a general revision of the copy on this page,

The pricing page? If you want & have time to write a bit more, that'd be interesting.

(Or if you're too short of time, I guess I can ask someone who works with UX & pricing)

> and removing all the "later" features.

Hmm, someone else has mentioned this too. Maybe there can be a "Show-later" checkbox, default un-ticked.

Thanks for the thoughts & feedback!

Doctors aren't that useful, unfortunately

It's a great book (I think), reading it now. It's the only book this far that I've brought with me to the gym (because I wanted to continue reading it, in between the exercises)

Having friends also contributes to living longer (Buffet and Charlie)

I'd like better debugging of async call stacks. Seems that'll start working better, with Java 21 and virtual threads, so looking forward to that.

And support for Scala in [web frameworks for Java].

Still on Scala 2.13, so don't know what 3.* I'd want, more than what already exist

Anyone has tried GleSYS? https://glesys.com

They're based out of Sweden, and offer bandwidth in Mbit/s, rather than GB/month (as per their pricing page, https://glesys.com/vps/pricing)

That's interesting, so, prices might not drop to pre-tariff levels, until not just one, but many, manufacturers appear in the US, and start competing on price (well, obviously, I guess some would say)

Sweden is one of only 10 countries on the planet that has developed its own fighter jet (JAS 39 Gripen, plus retired predecessors).

(At least ChatGPT lists Sweden as one of 10 countries with indigenous fighter jet programs.)

GGGP:

> They haven’t been in a war since 1814

Geography plays a role too I'd think. In a way, located in an icy corner of the world (rather than f.ex. in central Europe)

I was wondering if you're using a phpBB extension you've built yourself, and if it's on GitHub or somewhere (the extension), or ... It's not a built-in feature?

Websearched for "phpBB federation" and "phpbb subscribe rss broadcasts", found this:

"Feed post bot: This extension enables you to read any RSS, ATOM or RDF feed. It looks for new items every half hour and post them to a specified forum." https://www.phpbb.com/community/viewtopic.php?f=456&t=241159...

Intering way to use RSS

> If you make each individual bulletin board receive broadcasts from a central server

Your're doing this with phpBB? Doesn't happen to be open-source somewhere?

Would be interesting to have a look, I think I a bit like this opt-in partial federation / hivemind. Would be even more interesting if it was possible to sync comments between such forums.

**

Developing forum software myself, Talkyard. Based in Europe (Sweden).

Started thinking even more about using some European cloud, as an option. There's a Swedish hosting provider that looks interesting (I think)

Thanks! Ok, seems direct style is what I would want, after having asked an LLM: it says it's then simpler to debug async call stacks, with Java 21+ virtual threads.

Nice to know that there are good alternatives, if time it is some day to migrate away from Play.

> The language is just about perfect

Yes totally lovely :- ) And the best std lib? (On shared 1st place I'd guess, Rust looks nice too)

I just miss better debugging of async code, so I could see the execution context in other threads earlier in the "async call stack".

Tapir looks nice, didn't know about. Can I ask, do you use it together with Netty? How fast is it for you? (if you happen to have benchmarked it)

Have you tried Vertx with Scala? (Or Spring + Scala, or sth else?)

> The introduction of Guice

Personally I've wired everything statically at compile time, zero dependency injections. (Felt as if what I did went a tiny bit against the framework, but works fine.)

Now I feel curious about what do you build that makes that much use of parallel computations, races and killing the loser, etc?