The more I use the LLM AI tools, the more I see what a strange, strange creation they are. It's really quite inexplicable, and I think potentially quite pernicious. We've created a golem, and like all such things, it can be useful, but the harm it may cause may ultimately outweigh the good it does. Much of this potential harm comes because of the lack of wisdom of those working with the technology and developing it. This can be lamented, but I have no idea what we can do about it. I just try, in my limited way, to let people know—at a conceptual level—what it's all about, and how little they should trust LLM outputs. I have been suckered more than a few times. You would think I would learn.
Interesting that the "High School Student" DNA paper was inadequately reviewed by AI
https://geoffpain.substack.com/p/us-fda-censors-school-student-measurement
The more I use the LLM AI tools, the more I see what a strange, strange creation they are. It's really quite inexplicable, and I think potentially quite pernicious. We've created a golem, and like all such things, it can be useful, but the harm it may cause may ultimately outweigh the good it does. Much of this potential harm comes because of the lack of wisdom of those working with the technology and developing it. This can be lamented, but I have no idea what we can do about it. I just try, in my limited way, to let people know—at a conceptual level—what it's all about, and how little they should trust LLM outputs. I have been suckered more than a few times. You would think I would learn.
The more I use the tool, the more I see its pernicious limitations.