I would argue that AGI will happen because we have to consensus on what AGI is. This leaves it open for these large AI companies to throw their weight around and define what AGI is before they claim to have achieved it.
You and I might agree that it's not AGI, but that's not going to stop Sam Altman from using such a bogus claim to pump share prices right before an IPO.
This has always been the issue. This is an argument I made more than 20 years ago. AGI, whatever it is as a technical problem, is mainly a TESTING problem. If you don’t solve that, then AGI is remains a matter of faith. A cult.