HotlineInput

I genuinely believed for a while Bluesky actually has only users like her, with the same attitude towards AI. I'm so glad I was proven wrong.

Sometimes people enjoy the sight of how pathetic others are. At least in case of being put on list with malicious intent.

At one point in the past I had an experiment in recreating the looks of Windows 3.1 in Godot - it was STAGGERINGLY hard!

But a nice experiment, had some pleasant progress

No, I don't it will either and I'm tired to pretend it will

There already are open source projects with contribution guidelines banning pure vibe coding, allowing only properly assessed, only "LLM-assisted" code in PRs.

That's the good direction in my opinion, too. No sane project maintainer should ever accept half-assed unchecked code.

Genuinely never heard that actually happening.

I could be in a buble, though.

I have no idea how did I find my way here but this is my first insta-buy for a long time!

Congratulations for release and all, these aesthetics are something I've seen frequently in my dreams (and to a lesser extent in the first King's Field games) but could never find anywhere else in gaming.

I started a decompilation project out of boredom/hobby and I started comprehending Ghidra's attempt at translation to C, via Claude. I haven't touched C for over a decade beforehand.

It does miracles, I don't feel hopeless and totally lost anymore.

The way corporate managers see the slice they promise on Pizza Days

I mean, the alternative to it is a ban(?) on AI-generated content.

Plenty of people, especially here on Bluesky would be a fan of that.

I'm not so sure going back would be a net positive even for humanity altogether.

Plus, the genie is out of the bottle, it's impossible to put it back.

just saw a tiktok about a town resisting a datacenter on farmland and then it cut to the farmland and it was a cattle farm to talk about how great “nature” is.

the American people have been propagandized beyond belief. imagine thinking cattle farming is eco-friendly.

Thanks for clearing it up! Interesting time we're living, really wonder how this turns out. I don't even have the definitions at this point.

I wonder how a world will look like with this defenselessness. I mean, surely it's demoralizing but does not look prohibitively debilitating for projects.

Indeed I wasn't and had to read up on it. I think I'm getting it, you see people not having an idea whether they are complying with license and that a new standard may be necessary?

To me it seemed most people simply shrugged off the uncopyrightable nature of code. Nobody seemed to enforce it.

Yes, that's more understandable, though I'm undecided. I've got used to massive backlogs in professional SWE environment, maybe if I had concern for the inability to shorten it then it would impact me. I guess...?

Someone mentioned prioritizing or formal applications. I can support that.

I accept it's hard, but fail to see the demoralizing part. Isn't the evaluation process and pipeline the same, whether the code was written by a human, an AI, or mixed-in?

It's very possible I'm missing something from the big picture, not sure how much you yourselves test on PRs, etc.

or at least implied to be forbidding the use of such technology in PRs, but you cannot really be 100% sure it wasn't used so it seems more emotionally and ideologically loaded than protective or founded in pragmatism.

Sorry for the long reply; this has been the only comment I found worth addressing

That said, I've read the Jellyfin LLM usage policy, and I can support it; seems to be a good way to start consolidation of using AI tools.

I cannot support the 100% anti-LLM attitude, however; mostly because of the concealment argument I stated above.

One FOSS project I won't name explicitly /3

With caution, I would argue the problem you've stated could be a legal loophole, but pragmatically a non-issue.

Plus, from what I've seen so far, hunting for a "humans have been involved enough" metric is going to make genuine developers conceal the fact they used LLMs in their pipeline.

/2

Is it maybe a slip in attention? Judgement? Evaluation? Not sure of the right word here.

Or potentially a lack of interest, on demand, to hold any open-source project legally accountable for non-human code?

Can't see anyone sueing a project over it anyways. What to gain?

/1

AI is destroying open source, and it's not even good yet YouTube video by Jeff Geerling

I totally agree.

Meanwhile shit like this is spreading like wildfire on Youtube: www.youtube.com/watch?v=bZJ7...

We do not have a shared consensus world anymore.

For the record, I'm rather seeing the original caption of the meme ("see? nobody cares") in general, even from Bluesky users.

Xwitter has become a horrid cesspool, no denying that, but I'm trying not to judge the book my its cover, when it comes to users.

(my X account is inactive if that matters)

yeah...

Never ever in my life would I ever use an LLM for anything security-related.

I mean, general brainstorming about OPSEC or explicit hardening of software I develop *may* be fine, but that's it.

First time I've ever heard someone differentiating between "tech people" and "tech culture" and I had no idea how much I needed that in my professional life.

Suddenly things make more sense, from gamedev through webdev to hacking and ITSEC.

Plus, yes, I do will support every use case AI helps in.

I appreciate the restraint and wisdom in this position while the internet is flooeded with AI extremists of both sides

This genuinely feels esoteric kitchen knowledge.

That's a genius idea, I've been thinking about something like that for a while now.

How I Use Claude Code by boris tane

"The workflow I’m going to describe has one core principle: never let Claude write code until you’ve reviewed and approved a written plan. This separation of planning and execution is the single most important thing I do.

Well frankly, I have doubts. But we agree to disagree, I guess.

I'm merely striving for common courtesy, and to be excellent to each other.

Those were bullshit to begin with, LLMs have found their valid use cases within DAYS while crypto bullshit has been struggling for over a decade, you cannot compare them.

Rest assuerd, critical thinking children WILL use LLMs. Spoiler alert: LLMs don't have to be tied to megacorps.