hypernormie
It's never too late to give up
- Joined
- Nov 19, 2022
- Posts
- 5,885
- Reputation
- 7,780
When they first came out they were ok and quite convenient but nowadays you can't even ask Gemini basic questions without it hallucinating.
On basically every occasion I have recently resorted to an LLM for some piece of information it immediately began hallucinating.
Like in the first 3 replies it was just saying bullshit and had I not been somewhat familiar with the topics I was asking I would not have caught it.
Very interesting how these are supposed to apparently revolutionize the workforce but can't manage to spew three sentences without completing losing track of the context or the facts.
They have just been getting worse over time and hopefully they just continue to get worse and the whole thing blows up.
I for one have almost entirely stopped using them for two reasons, one is that it's inefficient to have to constantly recheck and correct for basic errors and two because it's extremely annoying to have to constantly correct a multi billion dollar machine on basic information while it gaslights and placates.
I'm curious how many of you actually find these things useful for real tasks or real research. I for one cannot imagine using these for anything truly important as they seem to routinely fail at mundane tasks.
You tell me.
On basically every occasion I have recently resorted to an LLM for some piece of information it immediately began hallucinating.
Like in the first 3 replies it was just saying bullshit and had I not been somewhat familiar with the topics I was asking I would not have caught it.
Very interesting how these are supposed to apparently revolutionize the workforce but can't manage to spew three sentences without completing losing track of the context or the facts.
They have just been getting worse over time and hopefully they just continue to get worse and the whole thing blows up.
I for one have almost entirely stopped using them for two reasons, one is that it's inefficient to have to constantly recheck and correct for basic errors and two because it's extremely annoying to have to constantly correct a multi billion dollar machine on basic information while it gaslights and placates.
I'm curious how many of you actually find these things useful for real tasks or real research. I for one cannot imagine using these for anything truly important as they seem to routinely fail at mundane tasks.
You tell me.
