bhaak

I wonder if anyone has already written a git over fax extension.

the cost of adding features is basically zero

Adding features was always the easy part. Maintaining the code OTOH is not going to be easier.

I see this with an experimental project I’m consciously vibecoding. The code base tends towards a spaghetti coded mess.

Of course you can put in some refactoring prompts and the AI will reorganize the code. But that makes it worse actually.

You have no mental model of the code and after a large refactoring even less.

But why aren't you pissed at a database where the columns have types that aren't enforced?

I knew most of he issues that you mentioned ... but 'ID re-use in AUTOINCREMENT'?

What the ...?

You don't need top engineers to port a program from one language to the other. Outsource it to India.

Of course, then you can also ask, could it have been done with a cheaper model. Probably yes. But then you wouldn't get free marketing.

They are like us. If you could take a baby from the past and let it grow up in our time, you wouldn't be able to distinguish it from a baby born in modern times.

If you see value in travelling to other countries and appreciate experiencing the different human cultures there, you should also appreciate experiencing human cultures from the past.

They are separated from us not by distance but by time which makes it harder to experience their culture.

So much was lost that we can only get glimpses on how lives were back then. So anything that expands our knowledge on the variety of human beings experiences is worth it.

I remember reading about this feature for Oracle in the 2000s and was always interested to use it in a production environment.

It never came to pass when we used Oracle, maybe now with Postgres I will finally have a chance at it.

They don't. After running, for the values in small_numbers from 0 to smlen-1 they are equivalent.

But if the last value of numbers[] is not smaller than 500, small_numbers[smlen] will contain that value for the first version whereas the second version does not write to small_numbers[smlen].

but I am being cautious, or as cautious as I can be given my desire to be sailing

Then step down as maintainer if you don't want to do it properly.

The human is missing form OP's description. "and it fills in the implementation". No human in sight.

You can't call it "engineering" if you don't care about verification.

Vibe coding is not engineering. Neither is copying answers from Stack Overflow.

Who even puts forward such claims?

Who ensures it followed the specs?

The more context an LLM gets, the more likely it will start to ignore instructions.

If the LLM runs a context compression, all bets are off. There's a reason Anthropic upped the context to 1M tokens to reduce the chance of this from happening.

Your post makes no sense unless you speak about project management in general.

The commits in question are no pull requests.

The commits were all from the original inventor of rsync.

Not a low quality newbie coder.

A human preserves more context and might remember what they did and when pointing out a new bug, they often have an idea what's wrong.

If TRIDGE of all people can't handle #LLMs without a slopocalypse, no one can.

That means you. That means someone you admire who is intelligent and careful and considerate. Not even someone whose opinions on technology you respect a great deal.

I disagree. The amount of commits is not from somebody who is carefully reviewing the new code and considering the changes done. It's from somebody who thinks they are in control and think they can guardrail the AI.

I've seen this at work as well. Maybe it's a small case of the braineater that so many tech bros get when they get older. But they talk about the AI as if it were a being that can be reasoned with and not that it's just a statistical interpolator and autocompleter.

I know when I'm vibe coding. Just last week I needed 5 colors for a green to read gradient for visualisation some states. I ended up with a script that outputs arbitray color gradients in 5 different colorspaces (including a colorspace for which AFAIK there's no support in Ruby as of now) and additionally also considers different color vision deficiencies.

Is it useful? Yes. Would I run this code in production? Hell no.

But the code that does the transcoding has been read.

Somebody needs to have read deterministic code to even have a chance of noticing something being wrong.

This has not happened here.

The whole code base is a vibe coded rewrite, half a year after Bun was acquired by Anthropic.

I see lots of ground for that claim.

Show me a forum or topics based platform that handle threads as good as proper mail clients? Don’t mistake the poor HTML view for how managing threads with thousands of replies look like.

Local filtering is the key to ignoring threads you are not interested in. Depending on the client with 2 or 3 keystrokes you are ignoring the whole thread or this particular sub branch of it and automatically jumping to the next interesting, unread message.

Remember that without memory protection everybody shares the same memory and everything is visible to every process. If a process doesn't release all their memory that memory leak will stay even if the process ended.

Shared libraries were typically loaded only once into the system's single shared address space. Any process could potentially overwrite another task's memory or shared library state.

I don't have my old setup ready but if I boot into Pimiga, I get about 60 task and processes running.

If you literally mean "seeing" workflows, because of the small monitors back then, you usually didn't have open programs side by side. The Amiga allowed to have multiple screens that were basically a better version of virtual screens combined with fullscreen mode.

Here is an example of somebody having Deluxe Paint open and the Workbench.

https://retrocomputing.stackexchange.com/questions/24842/ami...

The pinnacle of workflow design on the Amiga was of course ARexx which allowed applications to communicate through message ports and automation scripts.

> AmigaOS was a pre-emptive multitasking OS

Yes, but without memory protection.

That’s why it was so fast. :) Also surprisingly stable all things considered.

While the harness can block certain actions (e.g., tool usage), it can’t enforce perfect adherence to instructions because the model itself is probabilistic. The harness can reduce deviations, but it can’t eliminate the fundamental unpredictability of LLMs.

The rules that are fed into the AI are not unbreakable laws to the AI. We should always remember that.

But that‘s not an issue with the coding agent. It’s the model that doesn’t follow the instructions.

Given how an LLM works, you can never be sure it will always work. LLMs are not deterministic.

It passed all the tests.

If you can't trust your test suite to catch an automatic language translation you shouldn't trust it at all. :)

Funny. I just googled this site 2 hours ago for a font inspiration for a makerspace logo.

Michroma is a Google Font alternative for Eurostile.

Dungeon level depth, mostly.

This being NetHack, an answer is often not as straight forward as it could be. Most of the time the level difficulty is proportional to how deep you are into the dungeon but there are levels where your experience level factors in as well.

The story is a bit more convoluted. After the 3.4.3 release in 2003, the DevTeam stopped releasing new versions. They were still responsive when contacted by e-mail, though. But then we also didn't know how the development version looked like.

In 2014 the dev version was leaked to the community and in the following discussions with the DevTeam on how to handle this, the DevTeam got the first shoot of new devs in a while which lead to the release of 3.6.0 in 2015. This version was a polished version of the dev version, also incorporating some of the popular community patches at the time. The 3.6 branch received regular bugfix and security updates (3.6.7 was released in 2023).

Since 3.6.0 there's a mirror repository of NetHack on GitHub. So the development version that was internally numbered 3.7.0 which would become 5.0.0 was always accessible and, contrary to the 3.4.3 era, could be played anytime and was also installed on the public servers to play.

NOTICE: Starting March 7th, 2026 at 3pm UTC, players accessing the hdf-us server via SSH must connect to us.hardfought.org instead of just hardfought.org (ssh nethack@us.hardfought.org). Please update your configurations accordingly.

There was some bot DOS attack on hardfought, so they had to get behind cloudflare IIRC but that made having the same domain for somehow problematic.

And yes, it will be part of Junethack. 3.7 already was so it's actually just changing version numbers. :)

There are also official Amiga binaries. :)

The Amiga port was resurrected just a few weeks ago.

https://mastodon.social/@ipaschke@cyberplace.social/11625728...

May I introduce you to a concept called "shorting"? You can make money from falling prices without selling the stolen coins. As I said just moving Satoshi's coins would lead to lots of panic selling.

The snarky reply would be that having their funds stolen is not something that seems to discourage people from having cryptocurrencies as it happens all the time:

https://www.web3isgoinggreat.com/