I wrote about how dismissing the possibility of AGI is not an excuse to dismiss the possibility of significant AI risks.