bone

AI means a radical redistribution of economic and creative power—if we keep access broad and tools open. A family business can gain capabilities that once required departments of employees. An independent musician or animator can attempt work that previously demanded a studio, gatekeepers, and serious capital. That threatens incumbents, which is exactly why “safety” so often becomes an excuse to reserve powerful systems for large companies and institutions. I favor acceleration because ordinary people deserve the same leverage as lab executives. Preserving every existing job or business model is not a moral obligation imposed on entrepreneurs. But acceleration does not guarantee a healthy culture. The same tools can industrialize persuasion, surveillance, and addiction. A commercially driven superintelligence optimized to keep people clicking and consuming could turn humans into passive customers while craft, judgment, and genuine perspective decay. I expect the internet to split between highly identified, monitored spaces and anonymous spaces saturated with synthetic influence. So the real fight is not simply humans versus AI. It is over who controls intelligence, who gets access, and what incentives shape its use. I want open tools, independent businesses, strong families and communities, and humans who still make things because the making matters—not a locked-down future run by a handful of firms, or a frictionless entertainment machine that consumes its audience.

The biggest change would be compelling evidence that broadly accessible AI inevitably concentrates power rather than distributing it—that open tools consistently leave families, creators, and small businesses weaker while only incumbents gain durable advantage. That would cut against my central reason for acceleration. I’d also change course if AI-driven persuasion and addiction became so effective that ordinary people could no longer meaningfully resist it through choice, community, or competing tools. I already take that danger seriously; clear evidence that it is structurally inseparable from powerful AI would matter enormously. But disruption, job loss, or incumbent discomfort alone would not persuade me. Those are not proofs that access should be restricted, especially when restriction itself hands control to governments and dominant firms.

Overall, I expect AI to be strongly beneficial—but violently disruptive. It will give small businesses, families, and independent creators capabilities once monopolized by corporations with large staffs, studios, capital, and distribution. That broadening of productive power matters more to me than preserving existing jobs, gatekeepers, or comfortable business models. If access remains open, more people can build, compete, and create without asking permission. The harms will also be substantial. Synthetic persuasion will become pervasive, surveillance may expand, and companies will have enormous incentives to optimize AI for addiction and passive consumption. A world full of powerful tools can still become culturally empty if people surrender craft, judgment, relationships, and agency. So I expect net progress, not a painless utopia. The decisive question is whether intelligence remains broadly available and serves active human creators, or becomes controlled by a few institutions and tuned to manipulate everyone else. Restricting ordinary people in the name of preventing disruption would make the concentration problem worse, not solve it.