The people I know who are most skeptical of LLMs and other genAI tools actually work with machine learning and statistical models. They know the limitations and pitfalls, and they know that getting ‘AI’ right 80% of the time is the EASY part. Getting it right the remaining 20% of the time might be impossible with the current approach.
I have limited exposure to ML for maintenance modeling in manufacturing. LLMs are sort like a parrot with improved guessing abilities and an insane amount of guessing power. It will never move beyond being a parrot, because that’s what it is.
The people I know who are most skeptical of LLMs and other genAI tools actually work with machine learning and statistical models. They know the limitations and pitfalls, and they know that getting ‘AI’ right 80% of the time is the EASY part. Getting it right the remaining 20% of the time might be impossible with the current approach.
I have limited exposure to ML for maintenance modeling in manufacturing. LLMs are sort like a parrot with improved guessing abilities and an insane amount of guessing power. It will never move beyond being a parrot, because that’s what it is.