I’m going to be honest, I have no idea how open source works. I can’t imagine maintaining anything more than a tiny library that I can ignore six days of the week.

Also: open source relies on good jobs. You can only do it if you have a well paid low stress job with good hours. Those have been in short supply recently.

I think the free time covid gave, followed by the free time the layoffs gave, and AI have been patching / hiding the fact that the core model of open source is completely unsustainable in its current state.

I don’t know a ton about Gitea, but I’ve recently starting looking for a simple git server + decent web UI

Gitea and Forgejo are the main recommended ones, but they both seem overly complex. (3D File previews?? Who needs that?)

“misrepresent” is a vague term. Actual graph from the study

The main issue is usual… sources. AI is bad at sources without a proper pipeline. They note that Gemini is the worst at 72%.

Note, they’re not testing models with their own pipeline. They’re testing other people’s products. This is more indicative of the product design than the actual models

Pffsh, we can allow a few thousand violations of the ceasefire

Holy shit boycotting works?

That’s a terrible metric. By this providers that maximize hardware (and energy) use by having a queue of requests would be seen as having more energy use.

It’s cheaper though, so very likely it’s more efficient somehow.

Original source instead of blogspam: https://www.tryprofound.com/blog/ai-platform-citation-patterns

It’s because technological change has a reached staggering pace, but social change, cultural change, political change can’t. It’s not designed to handle this pace.

Scroll was just for reading websites though. Musk seems to want We chat style super app

Some sort of universal microtransaction layer is the dream. I believe there’s also a proposed web standard for it.

Scroll was also making it work before they got bought by Twitter

Yeah I think we’re going to be grappling with this issue for at least the next decade. The traditional web model falls apart under AI

If you find yourself in weird corners of the internet, schizo-posters and “spiritual” people generate staggering amounts of text

the ability to rationalize and see through difficult to explain situations has never been a human strong point.

you may be misusing the word, rationalizing is the problem here

saying it’s rare for a person to think and do things that I do.

probably one of the most common flattery I see. I’ve tried lots of models, on device and larger cloud ones. It happens during normal conversation, technical conversation, roleplay, general testing… you name it.

Though it makes me think… these models are trained on like internet text and whatever, none of which really show that most people think quite a lot privately and when they feel like they can talk

I… guess he was trying to stop more of it. Regret?

Yeah likewise. I think it shows the primary weakness of Llms right now is not skill or understanding, but context.

It can’t use functions or docs that it doesn’t know about. Neither can I. RAG systems are supposed to solve this but in practice they don’t seem to work that great

Arguably we made it worse. They would probably have regressed towards the mean had we not waged war on them for 20 years, tried to shove our supposed values by force, in the process only further convincing them that such values are hollow and meaningless and encouraging the leadership of the more violent, radical people.

This is correct.

In this case it is true though. Soon after grok3 came out, there were multiple prompt leaks with instructions to not bad mouth elon or trump

They’re “how dare you tout a chromium browser” votes. The FF circlejerk is and has been crazy for a while now