I Don't Like LLMs

The author has mixed emotions regarding AI and language models, specifically LLMs. They find these technologies fascinating for their potential to boost productivity and the products we can quickly create. However, they also harbor fears about the potential harm LLMs may cause, such as agent swarms taking over infrastructure or designing bio weapons.

Despite these concerns, the author believes we cannot avoid the AI technology train, although they express hope for a successful outcome. They acknowledge that LLMs can be useful and even "irresponsible not to use them," as they are more thorough and faster. Even so, the author dislikes interacting with LLMs due to their uncanny, human-like voice and the fact that they sometimes confidently provide inaccurate information.

The author compares this experience to interacting with irritating, dishonest humans they would prefer to avoid. This visceral dislike extends to LLMs, as they are created by people with specific values and worldviews, making them tools that reflect human biases. The author acknowledges that LLMs are essential for productivity but feels uncomfortable engaging with them.