Episode 403 of RevolutionZ takes a side trip away from our current ideology sequence to revisit the issues surrounding the development of Artificial Intelligence. Media is full of smart sober folks saying the end is near but also with smart sober folks saying utopia is near. Are we on a deadly trajectory to hell? Or is heaven on earth the next stop on the technology express? Or maybe it is just two-headed hype. Just another shiny new technology but barely more than that? More immediately, what has happened of late? What is likely to happen next? What should we ourselves work to cause to happen?

AI is getting “better” so fast that arguing about today’s mistakes can miss the main fact that matters most: the slope of the curve. We consider what happens when large language models evolve into agentic systems. When being able to spit out "the next word" so effectively that passing the Turing Test is a lark, to being able to plan, to autonomously act, even to recruit other AIs, and to self-improve in ways even their creators can’t predict not only impinges on what used to be how humans earn and live, but also, how we manifest our humanity. We ask a simple question with big implications: if you had a three-position switch, would you let AI continue full speed, pause it until strict enforceable rules exist, or shut it down forever? 

We dig into the motives behind AI's trajectory. The story involves innovations; but the crux is motivations and incentives. When CEOs believe the first country and company to reach artificial super intelligence, or ASI, wins everything, for them to seek safety becomes in their eyes a competitive handicap. The public becomes collateral. We consider how the emerging "who is going to win, it better be me" emphasis means for AI safety, regulation, data centers and "compute," and why “trust the companies” and "trust the government," is not a plan. 

We also get specific about real-world impacts: deepfakes that shred shared reality, AI-enabled fraud that scales without limit, surveillance that tightens lives, and job displacement that turns productivity into pink slips. And then we step outside the box of familiar AI discussion to ask even if we imagine perfect intentions--no nefarious apocalypse of greed and hate--we still have to wrestle with the impact of unintended consequences like outsourced reading, outsourced care, and the slow erosion of the skills and bonds that make us human. 

This episode is unsettling. Share it with a friend who thinks AI is just another tool, all that matters is who owns it, and ask yourself and others: do we need fast, faster, and fastest development, as now, or do we need a moratorium, a full shutdown, or something else entirely? Data centers are certainly horrible. To oppose them is more than warranted. But if the companies were to find a way to ASI without data centers, would that suddenly be something to celebrate--or to terminate? Is AI an existential danger? Is it a path to civilized living at last? Or is it just a lot of hype?

Support the show