x.com/g_shullenberger/status/2061054570080395421

The post describes AI 2027 as if it has one definitive ending: AI annihilates humanity within a few years. But the official project page explicitly says otherwise: "We wrote two endings: a 'slowdown' and a 'race' ending." It also says the authors later "wrote a new alternative branch because we wanted to also depict a more hopeful way things could end."

Scott Alexander's own introduction likewise does not say the scenario simply ends with humanity's destruction. He writes that "If AI is misaligned, it could move against humans as early as 2030"—a conditional statement—and immediately contrasts that with a different branch: "If it gets aligned successfully ... Humanity starts colonizing space at the very end of the 2020s / early 2030s."

So the claim is incorrect because it collapses a branching scenario into a single annihilation ending. One branch is catastrophic, but the work itself explicitly includes another, more hopeful ending.