They were sadly on-point with ‘Biden promises peaceful transfer of power to guy he called a Nazi.’ They were just stupid about which part of that was a problem.

If it’s not happening inside your computer, it’s basically on Archive.org.

Act accordingly.

If it just works, there’s no problem.

If it doesn’t work, that’s friction.

Admiral Patrick being a crank again. Same guy one-strike permabanned oodles of people for voting wrong on comments about some prick getting shot.

His instance is proudly named after West Virginia. Y’know, the cantankerous gun-toting hillbillies who fought the rest of Virginia to end slavery and fought the army to unionize miners. But we can’t go discussing any extremist solutions to societal problems.

This would have been caught sooner if the OS itself was not infested with ads. You pay how much to get inundated with distracting corporate propaganda? In the Start Menu?

Ban advertising entirely. Brain-poisoning moral hazard of a business model.

Conveniently has The Idiot’s signature right on it.

Cutting-edge JS is for boring stuff anyway. We’ve gone from formalizing the best bits of jQuery to ‘click thing to make box appear.’ So you’re only avoiding tutorial-ass boilerplate, and quite frankly, boo hoo.

Meanwhile CSS is still in the “Jesus Christ why wasn’t it like this to begin with” phase of feature adoption.

“Okay, that was always an option.”

Gradual adoption is fine. It’s so much better than seeing StackOverflow answers from like 2013, and knowing not a goddamn one of them works anymore. It’s massively better than having to write separate code for IE and Netscape.

“How do I?” is a question. “How to” is an explanation.

It’s the hot lady room, adjacent to the bucket of truth.

Of course The Idiot does not understand anyone collecting fees on that is a collapse of the status quo we supposedly demand, because nothing is real to him unless it’s a grift.

Downvoters have absolutely fucking nothing to say, but sure are mad about something they can’t defend.

Not sure whether you’re being sarcastic here

Jesus.

“Name a publicly traded company that generated significant profit as a result of buying llm services from a provider.”

This at least is just a problem of relevance. Utility is not measured in money - especially for open-source non-profits like Zig - and people selling shovels and promises are obviously not the best place to look for success stories in mineral rights.

But I’ve seen several Youtube channels hand-wave complicated software for devices they’ve just built from scratch.

And several copyleft projects are caught in philosophical arguments against permissive rewrites with completely different code, structure, and/or programming language.

And a few ultra-niche projects simply typed “rewrite this in Rust” for an order-of-magnitude speed boost.

The kneejerk insistence that LLMs are never useful is wishful thinking. It’s less defensible than equally glib promises of institutional value, because even some of the people loudly distancing themselves do so because it worked. ‘I didn’t recognize my own codebase!’ Yeah, because you didn’t write it. The chatbot did. It’s not Inception-ing the experience of making it yourself; it’s inventing exactly the Github dead-end you’d like to find.

The linked Asimov story is only dimly related. You’d expect a culture relying on other minds, kept simple, somewhere between Brave New World and HG Wells’s Eloi. But it’s about the method of learning, not capacity or demonstrated ability. Like Bowie: “the fingerprints will prove that you couldn’t pass the test.” Your typical mid-century sci-fi civilization where The Committee has dictated how everything works, and obvious criticisms could be resolved by doing things very slightly differently.

The twist is even less relevant, because it treats originality as some vanishingly rare thing that must be tricked out of people. Hard left turn into YA protagonist syndrome. Even though the main character just got done describing the simplicity of learning incremental improvements.

Meanwhile the reality of LLM programming is, here’s an idiot robot that will do its best for literally anything you ask… and you are expected to spot where it has fucked up. The only way it’s better than letting project leaders run it themselves is if you are putting in the effort to babysit and second-guess the chatbot.

Thinkin’ about LLMs? That’s a paddlin’.

Same reason as before this policy: they care about the project itself.

If you don’t view these tools as the devil, of course you might roll your eyes and say, that’s nice, anyway, here’s an improvement.

A nice low bar, set fairly and in good faith.

The invention of photography just barely predates the first photographic hoaxes.

I keep coming back to Death of a Fantastic Machine.

Spotting a fake used to mean spotting edits, such as a repeating background, a shadow in the wrong place, or artefacts around a pasted-in object. That worked because manipulated images usually started with a real photo, leaving clues behind.

AI-generated images are different. They’re created from scratch, with no original image underneath, so those kinds of editing mistakes often don’t exist.

Even this is out-of-date, thanks to edit-models. This is why ‘replace her clothes with a micro-bikini’ is a problem. The model also doesn’t clone-tool from nearby in the same image, like us Photoshop Phriday veterans. They’re fully capable of generating an image from scratch, but replace just enough to blend into a given photo.

An image used to be reasonable proof that something happened.

Ehhhh.