He’s right though, this post says nothing non-obvious. It boils down to “AIs that can do some difficult things might be able to do other difficult things. Also: Danger!”.

There isn’t enough here for it to be interesting.

On the contrary this couple paragraphs is aimed strictly at 'normies' as in people who are riding the current exponential with little history of thinking about this moment. Most people's beliefs and knowledge needs updating and patching without hoping that 'it obviously follows from assumptions' because most people don't bother to think about the assumptions in the first place.

IOW:

> There isn’t enough here for it to be interesting.

True - for you and the dozens of folks like you, to paraphrase a meme.

is the horsepower behind ai speculation less than, greater than or equal to human speculation? seems they’re recursively related, though human speculation has more torque since it occurs naturally or at least as spontaneously as it takes for an epiphany instance to occur.

and then transposed into an ai chat prompt to get the analytical momentum for some grand scheme that gets published to whatever extent ie a comment, blog post, or whatever textual content that then gets ingested as training data…llm rinse repeat. not sure where this is going. is anyone?