It's not - in my experience reading HN - that AGI-sceptics (like me) don't think it will happen, just that we have developed a healthy contempt for the hype-merchants and their wide-eyed followers who despite years of "machine learning this" and "deep learning that" and "gpt-3 the other" have nothing to show for it in terms of something resembling human-level intelligence, or even any sign of understanding how we think.
It's perfectly reasonable to see that and still be confident AGI will happen - just not with current models.
It's perfectly reasonable to see that and still be confident AGI will happen - just not with current models.