Your comments have made it QUITE clear that you have no idea.

Odd, I can say the exact same thing about your comments on the subject.

We are clearly at an impasse that won’t be solved through this discussion.

Why are you so focused on just the training?

Because I work with LLMs daily. I understand how they work. No matter how much I type at an LLM, its behavior will never fundamentally change without regenerating the model. It never learns anything from the content of the context.

The model is the LLM. The context is the document of a word processor.

A Jr developer will actually learn and grow in to a Sr developer and will retain that knowledge as they move from job to job. That is fundamentally different from how an LLM works.

I’m not anti-AI. I’m not “crying” about their issues. I’m just discussing the from a practical standpoint.

LLMs do not learn.

You do understand that the model weights and the context are not the same thing right? They operate completely differently and have different purposes.

Trying to change the model’s behavior using instructions in the context is going to fail. That’s like trying to change how a word processor works by typing in to the document. Sure, you can kind of get the formatting you want if you manhandle the data, but you haven’t changed how the application works.

What part about how LLMs actually work do you not understand?

“Customizing” is just dumping more data in to it’s context. You can’t actually change the root behavior of an LLM without rebuilding it’s model.

Unless you are retraining the model locally at your 23 acre data center in your garage after every interaction, it’s still not learning anything. You are just dumping more data in to its temporary context.

… That keeps making the same mistakes over and over again because it never actually learns from what you try to teach it.

Better than PEBKAC.

AKA, the update where they start charging you to use your own hardware.

The use case of “make a shit ton of money licensing a proprietary standard” is kind of mutually exclusive with other use cases. It would be hard to cover.

My 3090 works flawlessly

This is what I hate about flatpak. I shouldn’t have to just know what random permission something needs. It should be marked with what permissions it requires, and then prompt for whether or not to grant them when it’s launched.

Sounds like it was just a small section of incorrectly mixed (possibly contaminated) filament at either the beginning or end of a production batch. When they start mixing the plastics there can be some irregularities in the flow of materials and you get a bit more of one thing and a bit less of another. That’s normally cut away and discarded once everything is flowing correctly. Sometimes it’s missed or they didn’t cut enough.

I’d consider it a one off, or very rare incident, and not a general overall quality issue unless it starts happening regularly.

True. I should have been more specific. No strings in that there should be no expectation of receiving anything in return.

Going back to 60 from 120, I can say that it’s very noticable. Even outside of games, in things like the window manager, just resizing or moving windows feels very choppy.

Nothing was stolen. The authors choose to give it away, for free, with no strings. That’s not theft.

No one forced them to choose that license, and no one forced anyone to contribute to that project.

Back in my day, we used our teeth.

To shreds you say?

Probably “great grandma” by then.

I just looked it up and wow, how have I managed to not hear about this game until now? This is going to be my next obsession.

It’s just the extreme pessimism that comes with modern society, not necessarily a conspiracy thing.