I would suggest everyone to keep conversation respectful.
@kahliburke has been making contributions to the ecosystem, that are available for everyone.
He is very open about using AI in his work. If someone does not like it, it is their right to avoid his tools, but this level of direct attack is clearly trespassing a line.
Proliferation of vibe-coded packages will make the ecosystem unmaintainable without LLM subscription. That is not good, and just not using those packages does not solve the problem.
One thing is to say “We should rethink policies to have packages merged in the registry”, one thing is to attack directly people that are heavy users of LLMs.
Also: your rant @Mikhail_Kagalenko seems pretty much a random walk.
- Complaints about the ethics of LLM companies and of the user that was awarded some free usage of tools.
- You cast doubt about a developer abilities, bringing a concrete case that actually refers to someone else work.
- You get discuss the mantainability of the ecosystem and how this is affected by vibe coded packages.
Point 2 is clearly wrong and you should apologize.
Point 1 and 3 are worth discussing, but I am not sure whether you are trying to have a discussion or just adopting a Flood the zone strategy.
It might seem so to you, but the fact is that in every instance I have been replying to the specific comments by the LLM proponents, who went so far as to bring up for discussion the comments on a different site. For example, the “concrete case” that I brought up refers not to “someone else,” but to the originator of this topic. The discussion itself has been resurrected by the same people who now complain to the “Julia community” like a bunch of crybabies.
So I think that it’s your turn to apologize for casting aspersions on my motives instead of addressing the discussion itself.
I get it, I am also annoyed by the massive amount of AI in human-to-human communications. But at the same time you are also free to not engage with any ‘pro’ AI posts. There is already a rule against AI generated comments (which I commented on recently, you saw that). I think that 99% of the comments here on Discourse are still written by humans, so I don’t think you have to stop being active here because of AI slop.
This is definitely not fair IMO. You must know very well each new package has bugs, and mostnew large packages have a lot of bugs. If the package is of any use, then there will be a lot of discussions about something not working.
As a Kaimon user who reported a bug or another - I can tell for most time I have little to complain about
No but I would have to mute a lot of threads from the AI religion. It is not (only) about comments/answes being generated with AI, but most of the discussions being about AI…
Besides the comment above – additionally, I did not say that the non-working part was due to AI – might also “just” be a package still getting mature and hitting boundary cases where it would also fail for a non-AI-developer.
It is still AI, AI, AI, AI,… in so many threads. All is AI now. I am not AI nor do I work with AI (again, I do not have the money, nor do I feel I should invest that privately), so I feel lost and no longer helpful here, when all is about AI.
But I do take from that comment, that my future answers have to be phrased much much more carefully to not make any AI evangelists angry because a comment was not 100% precise in every single small detail.Sorry for that, will try to improve and not comment on AI again.
Take, for example, the way PrettyTables.jl is developed.
I have been developing ALMOST ALONE PrettyTables.jl for the past 7 years. In my free time, receiving absolutely nothing for this work. Last month, the package was downloaded 330k times.
You criticize a lot, where are your contributions to PrettyTables.jl?
What’s funny is that IF I had access to Claude I would not needed to release a breaking change to standardize the keywords since I am not a native speaker.
Anthropic gave me access to an amazing TOOL that costs 200 USD per month (a prohibitive value for like 99% of Brazilians to pay for AI personally) that will help a lot my contributions to the ecosystem.
You, on the other hand, complaint on discourse but, at least in my packages, I saw 0 contributions and 0 funding.
It is not just a value-neutral tool, though. And when your advocacy for it receives pushback, complaining about it seems hardly fair. Anthropic destroyed millions of print books using borrowed money to gift you a new shiny toy. Sorry not sorry for bringing that up.
I just set up a slow mode for this topic, to remind everyone that the conversation needs to remain respectful. Coarse language or imagery, ad hominem attacks and personal criticism are not welcome here.
31 Likes
Does that refer to my use of Linus Torvalds giving the middle finger? Is that too coarse? I thought it was apropos personally.
Yes. This kind of rude behavior will not be tolerated
@Ronis_BR, thanks for posting this. I am very surprised that this positive piece of news about an industry player recognizing an important FOSS contributor turned into a debate about AI in general, but I was happy to hear about it.
If you find these tools useful, just keep using them. Your PrettyTables.jl package is immensely useful for me — recently I was using it to send tables to coauthors by email as plain text. One of them remarked that my tables are always so… pretty! Of course they are, that’s what your package does.
So whatever makes your life easier to make these tools will, eventually, make my life easier by having more functionality in your package. Formatting to various backends is an inherently finicky problem with tons of special cases, well-suited for tools like Claude.