Nice essay but the core point of AGI is that it will universally disempower humans in all domains, send the value of human labor to 0, and construct a world far beyond human comprehension that will be disorienting and horrifying.
It’s similar to the No True Scotsman fallacy: any bad outcome due to AI could be dismissed as “well no aligned AI would do that”. Alignment becomes an asymptotic north star which very well could be theoretically impossible, but would be impossible to know whether it was.
I'd argue we can't align humans to neccessarily degree to ensure even semi stable future. Or nations, or corporations.
Intelligence isn't magic. It can only overcome so much brute force and humanity, as a non-intelligent system made out of intelligent parts, carries an ungodly amount of brute force.