[jimbokun]

I agree.

This Robin Williams monologue nails exactly why LLMs make us so uneasy.

They speak fluently and confidently about experiences it’s impossible for them to have. They can’t taste a strawberry or do any of the things Robin Williams names.

There are a number of people building these machines who literally believe the machines will replace us and because they will be more powerful than us so nothing meaningful will be lost.

They need to watch this clip.

Even though they probably still won’t understand it.

[shermantanktop]

ChatGPT has taken to saying things like “What I would do now is…” or “if I were you I’d…”.

I know these are figures of speech, but it reminds me that this thing doesn’t do anything, it doesn’t learn anything, it can’t try anything and find out. And yet it uses speech patterns drawn from real humans who can and do all those things.

[zahlman]

I've noticed it commonly uses phrasing like "that's usually the next step" when I'm using it to design something that I can't find an existing implementation of.

[lgrapenthin]

Not a day passes with my LLM of choice making completely baseless claims about "many people", who supposedly share all my problems and solve them like the LLM proposes

[akiselev]

When I ask it to tweak recipes and stuff, it frequently says stuff like "my favorite way to..." or "I really like [x]".

I have a viscerally negative reaction to a machine claiming it has a favorite anything.

[shermantanktop]

Recipes! Dude, the only thing you eat is other people’s data.

[kgwxd]

They saw it, understood it perfectly, laughed at it, and continue fucking over humanity.

[SoftTalker]

It's a movie. The whole thing is fiction. Robin Williams memorized the whole monologue, or was reading que cards.

[tadfisher]

Ask ChatGPT sometime about the artistic medium of cinema, and how words combined with actors speaking them can be meant to provoke something within the viewer.

[mingus88]

And an LLM is just generating a token stream from a set of model weights

[jiggawatts]

Your thoughts are just some ions sloshing around a lump of meat.

That meat follows an ill-defined pattern encoded in fewer bits than the source code of PyTorch and its pretraining phase used a tiny fraction of the available data.

You’re a poor imitation of an LLM.

I mean… you’re fluent in, what, at most five or six languages? Can program in maybe another dozen if we’re being generous about your capabilities?

Pfft… who would trust anything to meat brains!? They’re famously prone to hallucinations!

[kgwxd]

Yeah, just like how every song sounds exactly the same live as on the album. There's no humanity in there, it's all just lifeless plans.