I read through it, and nothing in here suggests artificial general intelligence is close, much less artificial superintelligence.

There's a VERY basic metric you can measure this type of stuff by. The success of all those companies that went "all in" on AI and fired/laid off large portions of their staff.

Guess what, multiple studies, including a famous one from MIT, show 95% of AI implementations fail because executives vastly overestimate what its capable of and its incapable of matching, much less replacing, humans in, I believe 90-95% of instances.

I'd give links, but it would be redundant since you can just google it and find an near endless stream of articles, studies, etc. Or just google the financial filings of companies that went all in on AI and see for yourself what a failure it was. The most infamous would be Salesforce, who laid off 40% of their workforce by "replacing them with AI". Guess what? They're fast tracking bankruptcy and can't recover the human capital they let go because they've already moved on to better jobs while the AI is incapable of doing the work they used to in a sufficient manner, much less BETTER than they did.

Pretty much the only companies with any form of success with AI, are the ones who use it as a tool, an aid if you will, for human employees to make them more efficient, instead of trying to replace them with AI.

And before you argue "this is all based on old AI models, blah blah blah", no its not. Companies that're neck deep into this stuff will constantly switch to whatever model is "the most advanced" at any given point in time to stay ahead of the game, and it STILL isn't working.

General AI is only really good for math (which you still need to double check and use logic with), and coding (which it's ABSURDLY good at but still needs human oversight). It CANNOT copy a human brain or thought process fully, nor can it replicate the human ability to have independent thoughts.

Generally speaking, when AI IS successfully implemented, its either as a tool, or a NON general model that's designed SOLELY to be good at ONE singular thing. For example, a lot of farms now use AI heavy automation, but its NOTHING like what silicone valley is trying to sell everyone. It's basically just an advanced pattern recognition system that's hyper optimized so that it can detect things like ripe fruit and vegetables and designate then for picking so that it's connected robotic arms or drones will harvest them. It's SUCCESSFUL and GOOD because it's NOT AGI or ASI. It's a hyper specific AI designed for ONE singular purpose.

AGI is a pipe dream at this point. All of those "studies" that show AI being anywhere near it are the result of giving an AI instructions to do something and seeing if it would actually do it. As far as I'm aware, there has YET to be a single instance of an actually rogue AI that went rogue without a prompt making it DO so to see if it would as a test.

Likewise, this article mentioned robots at several points. That's a whole OTHER bag of worms, but its equally bullcrap. There has not been a single viable human shaped robot made TO DATE that would be functionally capable of replacing humans. They all either have energy requirements that're too high and would require being charged constantly (which has the effect of driving up energy price in operating expenses AND making it unfeasible for replacement due to having to be constantly plugged in, meaning wires would be laying around everything causing trip hazards for robots AND humans) OR they're incapable of advanced human motor function. Often both.

For example, all of those "tesla robots" aren't actually robots. They're drones being controlled by a human.

But I digress, you'd either need a car sized robot to be able to keep it working all day on one charge (which would defeat the purpose), or need to have them constantly plugged in and sucking energy (which ALSO defeats the purpose). And this isn't even debatable since its basic physics. The more complex and intricate the motor system (as in the muscle replacement system), the more energy its gonna take to run. So robots capable of fully replicating the full range of human motion AND fluid dynamics are ABSURD energy hogs. Which is the main reason not a single one has been fielded yet outside of just showing off prototypes as proof of concepts.

TL;DR: AGI isn't anywhere near a thing yet, much ASI. AI is a VERY GOOD tool, but only if used properly, otherwise its just program that doesn't work half the time without proper human input.