everythingism
@kevinroose's article is a good example of AI hype. At no point does he explain how we get from today's AI models, which for all their impressive abilities are still only outputting text on a screen based on predicting the next token, to "a general-purpose A.I. system that can do almost all cognitive tasks a human can do." All cognitive tasks should seem to suggest the ability to perceive and interact with the real world, make high-level decisions and update based on a process of continuous learning. These are all major challenges that have yet to be solved, so how are they going to happen within a few months? Roose doesn't even explore this question. Instead he cites AI CEOs making bombastic, over-the-top predictions and talks about a general feeling he has based on listening to people in the industry. "Feel the AGI" is ultimately a silly industry buzzphrase. When AGI has actually been achieved we won't need to feel it, we'll know it.