this post was submitted on 12 Apr 2026
126 points (86.2% liked)
Futurology
4181 readers
353 users here now
founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
"The AI output is always so shitty... the workers must be the problem, they're clearly sabotaging our obviously perfect way to make perfect profits!"
The problem is that LLM's sometimes gets the right answer and then you are like Wow this is the best! And the next minute you are thinking It must be me not giving enough context? Let me try a different model. which then also fails.
Intermittent reinforcement conditioning is literally the most powerful there is.
People who are not aware of their biases are mastered by them.
Stay curious folks!