News for hackers. Stuff that matters.
Re:GPT-5 by on_the_train on Thursday August 07, 2025 @01:51PM (#44827890)
It's a misconception that almost everyone does though. I recently even saw it being being taught in a zeppelin museum!
Re:GPT-5 by sejje on Thursday August 07, 2025 @02:58PM (#44828863)
LLMs are "ask the audience"

Common misconceptions should be expected when you train a model to act like the average of all humans.

Re:GPT-5 by xeromal on Thursday August 07, 2025 @01:53PM (#44827921)
Why replace humans if make human mistakes
Re:GPT-5 by metalliqaz on Thursday August 07, 2025 @02:03PM (#44828095)
less overhead on benefits and pay raises