The more we work with LLMs (Large Language Models, like ChatGPT and Gemini and Claude), and the more we struggle with the stupid mistakes that they make, the more we see how very *human* their errors are. They struggle to please. They over-estimate their own ability. They fail to cite their sources. They misremember things. They are *terrible* at mental arithmetic. They fall prey to the easy answer. They don’t have good strategies for thinking about really complex problems. The more stuff they are trying to keep in mind at once, the more likely they are to forget a step, to decide that something really isn’t that important right now, that surely ignoring that rule just this once won’t hurt. They get overwhelmed. They get confused. They forget things. And by studying the parallels between how AIs “think” — or, rather, fail to think — and how humans think and fail to think, we can begin to work more effectively with their limitations. Like we do with each other. Like we do with ourselves.

This week’s speaker will be Mikki Phoenix.

After the talk, we’ll have a group discussion to explore the ideas presented, and to share related thoughts, resources, and personal experiences.

This is an online event hosted on Zoom. You can register for it here:
https://us06web.zoom.us/j/86448305839?pwd=bjmKNOytlKSdJ3PSoqH9LbaehIxcft.1