This is why I hate search engines promoting AI results when you are researching for something. It is confidently giving incorrect responses. I asked for sources on one LLM model before while using Duckduckgo, and it just told me that there are no sources and the information is based on broad knowledge. At one point, I challenged the AI that it is wrong, but it insisted it doesn’t. It turns out that it is citing a years old source written by a different bot long ago. But on the one hand, most of you are probably familiar that on occasions that the AI is incorrect and you challenge it, it will relent, although it will be a sycophant even though you yourself are actually incorrect. This is Schrödinger’s AI.
This is why I hate search engines promoting AI results when you are researching for something. It is confidently giving incorrect responses. I asked for sources on one LLM model before while using Duckduckgo, and it just told me that there are no sources and the information is based on broad knowledge. At one point, I challenged the AI that it is wrong, but it insisted it doesn’t. It turns out that it is citing a years old source written by a different bot long ago. But on the one hand, most of you are probably familiar that on occasions that the AI is incorrect and you challenge it, it will relent, although it will be a sycophant even though you yourself are actually incorrect. This is Schrödinger’s AI.