Hey we made the news...

Its not often you read about FreeBSD Quarterly Reports on Tech News sites.
That has to be a good thing or a slow news day.
 
Who are they calling a vulture? 😆
I'm with NetBSD about AI-generated code. Being able to see If something is generated is already questionable.
 
  • Like
Reactions: mro
why should any opensource project allow robot code? part of the appeal is to take part in a community effort; not sure using a robot should qualify as contribution(s).

1) I would venture a guess that most people could write a prompt to get some code so a "robot wisperer" should not get the same level of credit as someone who wrote it by hand. This would be same argument as me saying I am just at good at programming as [you] now that I have access to a slopbot so I deserve a raise.
2) The person uses the dummybot to only produce a test. Why would someone who obviously knows what they are doing have a problem saying they used a slopbot to produce a test(s)?

I'd say: submit the slop but you are responsible for writing tests for the slopbot tests but you also do not get credit for the submission.
 
The only thing I hope is that FreeBSD will remain true to itself by being pragmatic and measured; the question is not to ban everything, nor to allow everything.
 
In other words, interesting and quite radical stuff is happening in the FreeBSD space.
Overall, the situation is more or less clear. For me, news like this falls into the category of "well, of course it does."
I’d just like to point out that what matters to me isn't so much the push to integrate AI into the depths of FreeBSD, but rather how much this is going to complicate my life. If it just creates problems for the sake of creating problems, I always have the option to switch to another platform. We shouldn't let AI create unnecessary headaches for the end user. Sure, the people at the top might be making good money off this, but I want to be clear: I have absolutely no intention of becoming a beta tester for AI-driven models. In other words, if a version or a fork of a BSD-based system incorporating AI pops up tomorrow, I’m not taking part in that circus of madness.
And why the hell would OpenBSD want to pull in a port of Wine for Windows applications? Why on earth would anyone want this crap in the OpenBSD ecosystem?
 
Overall, the situation is more or less clear. For me, news like this falls into the category of "well, of course it does."
I’d just like to point out that what matters to me isn't so much the push to integrate AI into the depths of FreeBSD, but rather how much this is going to complicate my life. If it just creates problems for the sake of creating problems, I always have the option to switch to another platform. We shouldn't let AI create unnecessary headaches for the end user. Sure, the people at the top might be making good money off this, but I want to be clear: I have absolutely no intention of becoming a beta tester for AI-driven models. In other words, if a version or a fork of a BSD-based system incorporating AI pops up tomorrow, I’m not taking part in that circus of madness.
And why the hell would OpenBSD want to pull in a port of Wine for Windows applications? Why on earth would anyone want this crap in the OpenBSD ecosystem?
http://daemonforums.org/
Above link is the forum where are OpenBSD users active.

P.S.
Check Haiku web site.
 
I used omarchy for some time and ran from dhh aparent "extreme right wing".... anyway, I just dont like getting money for him is easy while other projects lack funding....

and about AI, wanted to post on my profile, but limit is 480 chars....


My idea of AI usage is that you really don't learn a thing if you don't do it yourself.

With natural language "as code" you become at more a technical project manager that "codes a little" 10%? less? more? anyway... I will not call myself a programmer if I dont solve the technical problems as either technical or non technical PM will call themselves programmers or project managers.

Programming is solving puzzles with the language at hand, actually solving them even if with errors or future unprooven, describing things is at most a technical writing from a project manager view, not actually solving, understanding, learning how the code is made.

I have made some things with LLMs but see that I don't learn really anything new... if your purpose is to understand and learn do it by hand or at least yourself, if your purpose is only the output, direct like any (technicall) project manager will do.
 
FreeBSD should not accept AI slop. People could already code without a slop machine, and all this will do is overwhelm the project with large amounts of crap. The code that makes it in will probably be of poor quality and contain licensing issues. I can't wait for this bubble to pop.
 
People who don't like AI should come up with a solution where normal users - and even seasoned developers - just explain in plain English what they want and get deterministic output in return, so people no longer need to learn fucking Rust to get stuff done. Long gone are the days where we had to spend weeks or even months to get something that can be done in a few hours or days.

The good thing with AI is that we no longer need to create new languages and stupid frameworks to get shit done. There's no going back.
 
My idea of AI usage is that you really don't learn a thing if you don't do it yourself.
Code is just a means to an end, and we always learn what we need to learn. Nothing more and nothing less.

Computing has failed mankind for so many decades... New languages were created because the previous ones didn't manage to capture everything, and then you also had stupid patents, licenses, copyrights, etc. Computing was hard for most people without a good reason. Now it's accessible to everyone. Some gatekeepers may not like it, but it will be like Veganism: A lifestyle choice. Most people will continue to eat meat anyway.
 
I feel its a naming problem.

There is nothing revolutionary about using computers to find Micro-Lensing signals from Planets.
Our star searches have gone beyond human detection and we need computers to help us find anomalies..

The US Military has had auto target recognition systems since the time I was enlisted in the Eighties.

Pharmaceutical companies have long used computer simulations of compounds to solve medical problems.

Just like material sciences have been using simulations to find new materials.

All of these functions are now REBRANDED AI functions.
It is a total falsehood and there is nothing Intellegent or Learning about it. You train it for desired results.
Optically compare a gun to a pitchfork.

So while AI may have legitimate uses this crap where I can make fake pictures of cats is really getting stupid.
 
FreeBSD should not accept AI slop. People could already code without a slop machine, and all this will do is overwhelm the project with large amounts of crap. The code that makes it in will probably be of poor quality and contain licensing issues. I can't wait for this bubble to pop.
In general open source projects shouldn't, and definitely not before the laws are updated to consider the lack of license that you get when generating code that's been trained on whoever knows whose code.

I think AIs are potentially fine as a way of finding things to fix, but when it comes to actually fixing them, we don't need another lawsuit threatening to shutdown entire projects due to code slipping in from another project without the appropriate licensing.
 
Back
Top