Eventually there will not be enough human workers creating training data for LLMS to hallucinate with, and they will just be training each other with more and more incorrect bullshit.
I do find GitHub CoPilot helpful to save typing, but last night it took me longer to track down the mistakes it made than if I had just did a combination of copy/paste + regex.
Eventually there will not be enough human workers creating training data for LLMS to hallucinate with, and they will just be training each other with more and more incorrect bullshit.
I do find GitHub CoPilot helpful to save typing, but last night it took me longer to track down the mistakes it made than if I had just did a combination of copy/paste + regex.