Post

it is not "acting out" it is doing exactly what it's being programmed to do, and the failures come because it's being overseen by incompetent people with no ethics

Look man, I hate AI too, but yeah, a core principle of LLMs is that they use randomness to generate results. Saying they programmed them to do these things is like saying you programmed a craps game to roll a 7. You programmed a roll of a dice, but the actual result is still down to random chance.

More like an unbalanced roulette wheel. The result always lands within a probability distribution defined by the training set, just not always at the exact same spot.

The probability distribution isn't just defined by the training data though. It is also influenced by the previous random results. Give two instances of the same llm the same training data and the final probability distribution of the two llm instances will not be the same.