I’m not denying the decades of work of the contributors which has pushed Haiku far beyond what BeOS was.
I swear a few years ago I found some sources saying OpenBeOS was based on leaked source code… somehow I can no longer find those articles.
A lot of articles in Internet are telling lies. I know no evidences that Haiku used any closed BeOS code. Actually Haiku design is quite different from BeOS in many parts, for example app_server design and app_server protocol. Note that some BeOS code was officially opened under open source license, for example Tracker, Deskbar, FAT driver etc..
1 Like
That won’t change in any case, either we use LLMs to develop Haiku software or not.
1 Like
Yes, but I will not be personally responsible for it and I feel much better this way.
You do whatever you want with your own conscience.
Yes, from Jean-Louis Gassée “50 years in tech” article series for example. But he was mistaken, and he later amended it to rectify that. Probably some confusion with Zeta OS, for which the situation was never fully clarified. If I understand it correctly, a BeOS distributor in germany had a contract allowing them to make changes and fixes needed to support their users, and decided to take a very liberal interpretation of that contract. But I wasn’t there, so if you want to know what was up with that, you’d have to ask people who worked there 
But OpenBeOS was cleanly implemented from scratch, except for code which was released under appropriate license from Be.
2 Likes
First of all, dismissing facts as “left wing propaganda" is actually right wing propaganda. The only opinionated part of what you quoted was “and no one cares“ – which is actually hardly propaganda either.
Science is about facts. If you acknowledge men made climate change, it’s not propaganda, it’s acknowledging a fact. It’s a fact that CO2 reduction (among others) can stop the climate change.
Now how you react to these facts is politics. Denying facts is not. That’s just stupidity.
There are people who believe everyone should be free in the sense that they can do what they want no matter what that means for other people, and there are people who believe everyone should be free without hurting the freedom of others. One part is mostly selfish, the other part tries to care.
There is good reason to leave these things out of the Haiku discussion channels, as it easily creates heated discussions that usually go nowhere, and only leave bad feelings behind.
7 Likes
As I mentioned before, there is also an option to consider current consumption rates more important than potential harm of climate change. And make a decision to not reduce CO2 emissions. It is perfectly valid political decision, but someone may not agree as with any other political decision. Claiming that reducing CO2 emission is the only mankind option to survive is factually false and manipulation.
Anyway it is definitely not Haiku subject of responsibility. Haiku is not environment protecting organization.
There is nothing wrong with being selfish as long law is obeyed. There are no universal humanity morals. Not admitting that a lot of people are selfish is delusion. Selfish people are allowed to participate to Haiku I think.
Fully agree. So let not talk about CO2 emissions and saving the planet when making Haiku decisions. We are not UN.
3 Likes
@nipos Thanks for the link, I’ll read it with interest. I’m the first to acknowledge that current LLMs have enormous limitations and that the hype is often disproportionate. My position isn’t “AI is infallible”, it’s “it’s a tool, use it if it’s useful to you, ignore it if it’s not”.
For me it was useful. Without LLMs these projects for Haiku would never have existed:
- Sestriere native MeshCore client for LoRa mesh networks
- DenteBlu Haiku’s Bluetooth stack completed with the missing parts for pairing and communication with phones
- BubiCam camera application
- haiku-uvc-webcam UVC webcam driver
- AkaiAPCmini support for the Akai APC Mini controller
Many of these projects are already six months old and should be updated given the improvements that have happened since. For others it might not be useful, and that’s completely legitimate.
2 Likes
@us3r1d Your position is the most balanced I’ve read in this thread and I respect it. You’re right that tolerance has to go both ways.
On the filter though I have a practical problem: how do you implement it reliably? Today I can tell you that Sestriere is written 100% with AI. But if tomorrow I update a library that in the meantime received a commit written with Copilot, does the flag change? Who verifies it? Who updates the data? And if gcc itself has parts written with an LLM, should my code compiled with gcc be flagged too?
I’m not opposed to the idea in principle. I’m opposed to the fact that in practice it risks becoming a stigma based on a voluntary declaration that’s impossible to verify, not a useful piece of information.
2 Likes
perfect is the enemy of good. Such a label does not have to hit everything in order to be usefull, just as our “Native” and “Desktop” labels probably don’t cover exactly everything. (and there is always overlap or “fuzyy borders”), but even with fuzyy borders you can still clearly point at stuff and say “that is made with an LLM!” without having to solve the ship of thesius problem.
2 Likes
@PulkoMandy On the environmental point though I’d ask you to be careful: you’re writing this message from a computer connected to the internet, on a forum hosted on servers that consume energy. Professionally I work in sustainability, I’ve been driving electric for ten years, and through my choices and investments I generate more green energy than I consume in a year. I’m not saying this to brag, I’m saying it because sustainability is a complex field and before using it as an argument in a technical discussion it’s worth making sure you know what you’re talking about.
I don’t want to keep fueling this discussion. The tutorial is there, it works, the projects are public and open source. Whoever wants to use them is welcome, whoever doesn’t is free to ignore them.
3 Likes
Those don’t cancel out, why did you bring it up?
Also, on the previous topic of logical fallacies, like the previous strawman of “there exists a group on this forum, i tell you, and they have bad arguments against LLMS, which are these Let me disprove them!”, this is an argument from Authority, the “I work in sustainability” part, It does not work as proof of anything, even moreso if you ignore the arguments entirely, it only raises doubt on your claim to work in that space.
This is also a logical fallacy, this is “what aboutism”. “How dare you critizise people consuming endless water somewhere else, when you are also consuming (some) water??”
1 Like
It’s not left wing and it’s not propaganda. But I think what bothers you is the (assumed without evidence) “left” part. I am guessing if it was “right” you wouldn’t complain that much, if at all.
If there is one person in this thread that talks politics, that would be you, by seeing “propaganda” wherever you want to see it.
@nephele Sheriff, I think it’s time to restore order in town. 
The discussion has gone quite off-topic: we ended up talking about climate change and energy politics. Maybe it’s time to bring things back on track. 
1 Like
Among the many reasons I want to generally avoid “AI” (most already mentioned in this thread), I have another I get back to, the few times I think about this topic.
I guess I might experiment with LLMs in “code-complete on steroids” and/or “‘talking’ rubber-ducks/code-analisys” roles… IF, and only IF, I was able to run “Claude” (or whatever else is cool) entirely locally/offline, and with my very modest hardware (fat chance of that).
I just don’t like at all the idea of having a subscription to “out-source my thinking process”.
Sure, you may not pay for “Claude” (or whatever else is cool), for now. But end game seems clear to me: make a lot of people over-reliant on tools outside people’s control, then extract as much money as possible, and “locking out” everyone else one way or another.
World already has too many “haves and havenots”. I just don’t want to contribute to create this new one.
Anyway… goes back to slowly read the ACPI 6.6 spec, with hopes of submitting a small driver patch, eventually.
3 Likes
@nephele On the authority point: I didn’t bring it up as proof of anything, I brought it up to respond to someone using sustainability as an argument without the context to do so. It’s not “trust me because I’m an expert”, it’s “be careful before judging other people’s choices in a complex field”.
On the whataboutism: my point wasn’t “you also consume energy so be quiet”. It was that the line between acceptable technology and technology with unacceptable impact is not as clear as it’s being presented. That’s not a fallacy, it’s a legitimate question about the consistency of the argument.
2 Likes
@BiPolar The point about dependency on external tools is legitimate and I partly share it. But it’s worth knowing that local models like Ollama with llama3 or mistral already run on modest hardware, no subscription, completely offline. They don’t have Claude’s capabilities, but for “rubber duck on steroids” they already work well today.
Good luck with the ACPI spec. 
1 Like
This is how propaganda works, by hiding a fact that it is a propaganda and claiming that it is a common cense. It is not. And of course right wing propaganda also exists and also not appropriate here.
2 Likes
It would have been a question, if you did present it as such, that is writing it as a question, for example asking what energy usage one considers OK for what purpose, but as presented it is still a fallacy. That you now elaborated does not really change that.
In this case it is very clear, because the enegy use and use of other ressources (like water for “cheap” clooing) is orders of magnitudes worse. Pulkomandy never said “all energy usage is bad” but your original statement implies that they should have, or that they fail to consider their own direct use of ressources.
Since your original message also does not move the topic forward except for issuing a warning to “make sure you know what you’re talking about and “I’d ask you to be carefull [With your claims regarding sutainability]” I really can’t read it as anything other than an appeal to (your own) Authority.
The point is, Pulkomandy presented arguments, you did not adress them, simply stating that, and implying with this message, that they can’t know the facts (by presumeably not beeing an expert) and therefore can’t form the “correct” opinion. They did not judge your use of these tools, they judged the possibility of using them for Haiku and Haikuports.
2 Likes
We really need a “tutorial only” category!
For Games, Apps, bug report and a tutorial about self control…
4 Likes
@nephele Let’s argue then.
On the appeal to authority: I cited my experience not as proof, but as context for a request for caution. “Be careful before using sustainability as an argument” is not “trust me”, it’s “this field is complex”. Sustainability is not a binary argument: a data center running on 100% renewables has a radically different impact from one running on coal. Using “data centers consume energy” as an argument without this distinction oversimplifies a problem that requires much more granular analysis.
On the whataboutism: my point wasn’t “you also consume energy so be quiet”. It was that any argument based on energy consumption requires a defined threshold: below which level is it acceptable, above which is it not. PulkoMandy never defined it. Without that threshold the argument hits everything indiscriminately: the internet, Haiku’s servers, this forum, the computer I’m writing from and the one he’s writing from. Asking where that threshold is isn’t a fallacy, it’s the minimum prerequisite to make the argument coherent.
1 Like