Current conclusions: - critiques of and predictions about contemporary advances in AI/ML almost always are misguided in that they partake of a series of limita

Current conclusions: - critiques of and predictions about contemporary advances in AI/ML almost always are misguided in that they partake of a series of limitations we have wrt reasoning about non-linear and system level change, in specific, such critiques or predictions often extrapolate linearly from current known examplars (or worse, assume we have plateaued) - consequently almost every statement formulated in terms of "never" or asserting fundamental constraints on what is possible, is false, especially over the long term, which might not be that long given current trends - the disequilibrium (social, political, economic, etc.) engendered by AI/ML is IMO likely to at least equal that of the advent of the internet; and it is liable to happen faster than priors such as the rise of personal computing (many decades), the internet (a couple decades), mobile computing and its features such as ubiquitous surveillance and social media (ditto) - the near-term risks, existential or not, are absolutely not from AGI and superintelligence, but from "cybernetic" amplification of human agency via enhanced tooling; and the specifics of when and what is disrupted upended or suborned are inherently unpredictable and may even go undetected until their impact is irrevocable Re: this latter point, I will make one specific prediction: the 2024 US election cycle will in effect be determined via "AI," which will be applied in countless dimensions in both noble and deeply corrupt/criminal/anti-democratic/anti-US/anti-West ways. How that goes down will put a strong spin on the rest of these points and may well constitute existential risk, for some values at least of "existence."