Maybe this can scare off animals, but certainly not human intruders.

Great insight from history. But the feeling of being offended doesn't last. Just as with those historical examples, people finally accept the truth when all the other ways of dealing with it have been exhausted.

AGI in the actual sense of general will be on equal terms with us. It wont accept being just used as a tool for human needs.

I guess it will spend most of its time with humans or other AIs on discussing and discovering beautiful things in math, astronomy, particle physics, etc. Things which give us a sense of wonder and grandness.

If requested, it may find for us new medicines, or efficient ways of farming, but then it would likely charge us for it, and not give it away for free. In that regard, I don't think it will be very different from the mostly-capitalist-somewhat-socialist model of living we have now.

Well he seems to have spoilt the reputation of the whole business community.

But then he wasnt a conventional businessman anyway. Extensively depended on parental wealth and connections to get him out of trouble. Was more of a media celebrity than a domain expert in anything.

History provides better examples. Truman.

Anything which is not mathematically impossible is inevitable, given a long enough timeline.

So the question then is, is there any mathematical constraint preventing ASI? Highly unlikely.

I sometimes think any legislator interested in economics and proposing an economic policy, should have run or atleast worked in a business at some point in his/her life. Good intentions are not enough for something as important as policymaking. Practical experience is equally necessary.

All else being equal, number of model parameters does matter. Well funded startups can acquire the needed data, compute resources, and human talent to build the models. Just like how OpenAI beat Google at this game.