I Don't Like LLMs

The author expresses ambivalence toward large language models, noting their productivity gains and potential for both harmful applications—such as agent swarms and bio-weapon design—and beneficial uses like miracle cures. They criticize LLMs for producing confident but often fabricated responses and for mimicking human conversation in an uncanny way. The piece argues that LLMs reflect the values of their creators, urging caution in interacting with them.