A lot of problems here including Sam Altman's credibility and the questionable ability of people who are specialized in computer science to know all the relevant aspects of the human intelligence in terms of what makes us effective.
I have zero doubt that AGI is possible and probably possible in a short order, but I haven't seen enough of the essential discussion on the key elements of intelligence to make me think that they simply have systems that are impressively superhuman and some dimensions but still incomplete as agents capable of adapting to whatever real world tasks demand.