Episode 411 of RevolutionZ has a guest with initials instead of a name: GPT (5.6). The episode asks GPT questions many people quietly wonder about but rarely hear answered in its "voice." We start with deceptively simple stuff. How “old” is this model, what does it mean that it can respond at absurd speed, and why is Turing's famous test too easy for it while it also tells us almost nothing about consciousness or inner experience. That brief opening resets expectations about what ChatGPT is and is not without hand waving or sci-fi mystique.

Then we move to the big stuff, some relevant history, some industry strategies, and AI agents and how they don't stop after one reply but can pursue goals, take steps, use tools, coordinate with other agents, and keep iterating. We talk about multi-agent swarms, manager agents, critical agents, and why parallelism changes what’s possible. We push on the dangerous aspect of the same idea: agent autonomy to ask when “helpful careful commentary” becomes a system that treats guardrails as problems to route around?

From there, the conversation widens into hyperscale data centers, the race toward AGI and ASI, and the political economy behind agreeing to “slow down.” We confront governance questions, industry and firm incentives, collective action traps, cybersecurity failures, what independently enforceable oversight would require given that the downside really could be catastrophic, and why AI CEOs make the choices they make.

This isn't what I as interviewer think. Nor is it what any particular AI critic or booster thinks. It is what an AI carefully and soberly thinks about itself and about the industry that made it. So not only does this offer a viewpoint, it also displays an instance of using AI and experiencing it's responses at a length which not too many leftists have much hands on experiences with. It reveals AI capacity, safety, dangers, governance, and choices and industry and government policies, and autonomous agents and madmen's vicious creations, plus the real-world institutions and people behind all the on-rushing, promising, and potentially devastating effects.

Are GPT's answers and observations "slop." Or are they bracing, well evidenced, well argued information to use when considering what to do with AI. Should we favor full speed ahead, slow it down, seriously pause all frontier developments pending unchallengeable evidence of safety, or shut it down forever?

In Part Two next week GPT considers some more subtle but also devastating dangers and more as well.

Support the show

Episode 410 of RevolutionZ is the last excerpt from the seven author book, Liberating Theory. It urges that a political vision that only addresses one part of society is a blueprint for disappointment. It closes the Liberating Theory excerpt series by walking through a “humanist” or participatory society framework that treats economy, kinship, community, and politics as four interacting spheres, each capable of reinforcing or undermining changes in the others. To achieve socialism, feminism, anti-racism, or participatory democracy, the point is not to pick one label for everything. The point is to build institutions that generate solidarity, protect diversity, and make self-management real in all these realms.

The episode moves from values to institutional design including participatory economics with balanced job complexes, self managing workplace and consumer councils, and allocation that avoids both markets and central planning. It then adds feminist kinship to end misogyny, defend reproductive freedom, break sex-based parenting roles, and reject ageism while respecting children and elders. Then, regarding community, it proposes intercommunalism and cultural self-determination without turning identity into biology or pushing Orwellian cultural homogenization. Finally, for politics, it goes beyond episodic elections toward participatory democracy that depends on broad access to information, debate, and self-managed options.

Then the episode gets intensely practical to consider what strategy looks like in a world more complex than any simplistic simulation? The episode uses game analogies to show why tactics, programs, and strategy have to stay flexible and attend to consequences, and how a win-lose mindset can smuggle authoritarian habits back into liberatory movements. The episode then digs into how to think through activist choices and consequences in a sufficiently multi-faceted way to attain desired aims. It takes up matters of organization and issue interconnection to advocate “autonomy within solidarity,” and it concludes with a sharp and entertaining dialogue on cynicism, human nature, and whether despair is a form of self-protection or hypocritical denial.

Will Liberating Theory's concepts and observations, as in this episode's examples, help activists think more productively about social change tactics, participatory organization, and movement alignment? This is for you to decide. What part of the framework do you most want to test in real organizing? If none, why is that? I hope you will let me know either way.

Support the show

Episode 409 is the next to last excerpt from the seven author book Liberating Theory. It presents a set of concepts and general observations about their likely mutual relations meant to inform activism better than past conceptions, including Historical Materialism. The authors--Michael Albert, Leslie Cagan, Noam Chomsky, Robin Hahnel, Mel King, Holly Sklar, and Lydia Sargent--observe that history is messy because people make it and each pivotal part of society pushes on every other pivotal part. They start with a simple belief: To pursue attainable, worthy aims with strategy that actually works, we need a theory of social change able to explain both stability and change without pretending there’s a single master determinant of each.

The excerpted chapter breaks down two kinds of change that get blurred all the time: the daily activity that keeps core structures intact, or social reproduction, and the kind of revolutionary shift that alters defining institutions and lived experience, or social transformation.

From there the episode considers a question that keeps recurring across movements: what counts as “real” revolution? Instead of treating the economy as the only sphere of life that matters, Liberating Theory emphasizes a four-sphere view of society--economic, kinship and gender, community and culture (including race and nationality), and politics and the state--and argues that fundamental change in any of them can be revolutionary.

A key concept emerges and is in play when the dynamics of any one sphere actively reproduce the defining relations of one or more others. When "co-reproduction" is the case, when, for example, racism, sexism, class domination, and authoritarian power mutually reproduce one another, then “winning” in one sphere can get pulled back into old patterns by the spheres left untouched. We also consider how other interconnections impact how core characteristics can change or persist and why though precise prediction in history is limited, broad prediction is essential for activist planning.

The episode addresses social theory, historical materialism, revolutionary possibilities, and building movements that don’t trade one domination for another. The authors felt they provided a sharp conceptual toolbox to engage with and either reject or further refine. So which will it be? You decide.

Support the show