Comment on Why is AI so bad at coming up with dialogue?
db2@lemmy.world 1 week ago
Why is AI so bad at coming up with dialogue?
Generalizing the question is really the way to get the most useful answer here. It’s because LLMs are being pushed as “AI” by tech bros with room temperature IQs. It can kind of do a neat trick and sometimes sort of almost get it right so these people who shouldn’t be in charge of a pet rock even with supervision who for some reason have absurd amounts of money decided it should be in literally everything so they could have more money. It doesn’t matter to them if it’s ever actually the right tool for the job.
LLMs have a use case. Several in fact. Creativity and decision making aren’t really among them except as a gimmick not to be taken seriously.
Iconoclast@feddit.uk 1 week ago
LLMs not being AI is one of the most common confidently strong statements I see spread here as fact pretty much every day.
We’ve had AI systems since 1956.
It simply refers to any system designed to perform a cognitive task that would normally require a human. For an LLM that task is generating natural language. For an Atari chess opponent it’s playing chess. They’re both still AI - narrow AI, but AI nonetheless.
People seem to expect general intelligence from these systems, and when they don’t get it, statements like this pop up. But AGI is not synonymous with AI. It’s also an AI system, just a subcategory of AI - same as GenAI is. No company claims to have a true AGI system.