quillette.substack.com/p/an-open-letter-to-scott-alexander

The article presents “no describable scenario in which AI is conclusively an extinction threat to humanity” as the RAND report’s conclusion. But the report itself used that sentence as a falsifiable hypothesis in its approach section.

Evidence:

• RAND’s own report summary says the researchers “found that it would be immensely challenging for AI to create an extinction threat, although they could not rule out the possibility.” It also says that in the three scenarios studied, extinction “would not be a plausible outcome unless an actor was intentionally seeking that outcome.” That is much weaker than saying RAND concluded there was no describable scenario.
• A Scientific American article by Michael J. D. Vermeer, the RAND report’s lead author, explicitly explains the sentence’s role: “My team’s hypothesis was this: No scenario can be described in which AI is conclusively an extinction threat to humanity.” In other words, it was the team’s starting hypothesis, not the finding they ended on.
• A text mirror of the report shows the structure directly: under Approach, the report says “Our analysis proceeds from a falsifiable hypothesis: There is no describable scenario in which AI is conclusively an extinction threat to humanity.” Later, under Findings, it says “We found that all three scenarios posed global catastrophic threats...”

So the problem is not merely wording. The essay turns a methodological premise into a reported conclusion, which materially changes what the RAND study actually said.