Doomerism Explained: The Case AI Will End Humans

I have expressed a lot of skepticism about AI "Doomers" on this channel—but am I being too dismissive? To really test my view, I decided to put on the doomer worldview and "steelman" their argument.

In this video, I give my level best attempt at arguing that AI poses a serious existential risk to humanity. I move past the skepticism to formulate and defend the core doomer argument:

1. AI will likely be strictly more powerful than humans.

2. AI will likely be misaligned with humans.

3. We cannot escape the misalignment.

4. If (1), (2), and, (3), then humans face an existential risk from AI.

5. So, humans face an existential risk from AI.

From the "Guiding Principle of Irrelevance" to the history of "Post-Human" Chess AI, I explore why some believe we are heading toward a catastrophe that we cannot negotiate our way out of.

Stick around for the end of the video where I break character to share my actual takeaway. Did building this argument change my mind, or am I more confident in my anti-doomer stance than ever?