Recently, there's been this weird obsession in Silicon Valley about escaping the "permanent underclass." This could be defined as a feudal structure where upward socioeconomic mobility becomes increasingly difficult. The reason we will enter this feudal structure, is because the working middle class will be gutted and replaced by the very software many of those middle class engineers built. The basic idea is that after getting replaced by AI, those who were middle class lose that social leverage they once had, whether that be education, familial guidance, etc. And those who are on the rich end of the economy, those who otherwise own the infrastructure, reap its benefits.

Most pro market, pro innovation people would say that Marxist ideology is completely infallible, and communism isn't practical. On the other hand, Marx raises real concerns that even market supporters must acknowledge to make the system most efficient. Specifically, one of those concerns is in regards to the asset owners congregating wealth into their respective companies, and stifling innovation. Many AI companies talk a lot about democratizing the technology so everyone can benefit from intelligence, not just those with access to it. But it's important to consider the fact models require constant tuning, servers, and maintenance. This means that to participate in AI, you need enough capital to swallow the losses of managing servers, tuning, and more until an unproven demand catches up. Following that logic, that means the AI market is inherently becoming anti-capitalist. Not only are discussions of government bailouts being seriously considered, but that means there's no innovation to challenge the giants. They can get comfortable in their complacency, because no small company will have the capital so survive against the giants. We can see that AI advancement consequentially is leading innovation in practical application, rather than the sci-fi superintelligence investors expected.

Since the superintelligence promise has started to lose steam, Meta, who's invested billions into their infrastructure, yet always lagging behind, exemplifies the issue previously outlined. A company can spend billions on R&D, without any clear path to success. As a result, Meta strategically pivoted to Muse, and other strategies many companies have followed suite with that aimed to increase day to day AI applications. But Meta's pivot is highlighting the key tension: billions don't ensure innovation. If capital can't even buy success, what makes a small start up think they have any chance besides going through the same investors, companies and people driving its technological advancement?

As previously stated, since dependence on success in AI depends on the recruitment and investment of a company made by the same wealth groups who believe in its success, there is no real growth potential. Companies can just buy out competition too, as we saw xAI do to Cursor, or the countless companies OpenAI and Anthropic have begun to swallow. And recognition, investment, and beyond would likely remain enclosed to the wealthy, since those in the "underclass" may be cognitively dependent on the very software the rich profited with, rendering their creative input redundant. It also touches on their other favorite idea, "agency."

Agency as I've come to understand it in Silicon Valley is the ability to autonomously take control over your actions and circumstances. In modern society, the idea of people asking permission and trying not to step out of line has become prevalent. Tech thinkers suggest that agency is one of the defining differentiating characteristics of those who will become successful today. Agency, as celebrated and normalized by the wealthy elite, will likely become an exclusivle luxury as those less wealthy sedate their situations through the cheap dopamine the wealthy are routing into their environments. And as the underclass grows more dependent on the wealthy, loosening that grip to retain the very agency that permits the possibility of socioeconomic mobility may become increasingly difficult to attain for the poor.

Someone who doesn't know my views might read this post as anti AI, or anti technological progress. But I would pushback, because I highly support AI & tech development. This piece wasn't meant to critique technology, communities or industries. Rather, the point is outlining the issue is in the push towards the K economy. And whether that push is intentionally driven by select groups, or a natural result of human tendencies, it's worth acknowledging, including the potential negative impacts it could have on our society over the long term. It's easy to say societal wealth should be equitable, but noticing the K economy is also important in regards to the fact the middle class is foundational for our functionality. The middle class being gutted will only impact everyone else in society. In other words, the technology isn't the problem. The problem may rest in who owns that technology, how people depend on it, and how that dependence further reinforces economic and behavioral trends.