← Glossary

Permanent Underclass

A strain of Silicon Valley doom rhetoric holding that anyone who doesn't work maximally hard to get inside an AI company will be permanently locked out of prosperity once machines take the jobs -- paired with its mirror-image 'permanent overclass' of AI-company insiders on top. On ADI Pod, a framing the hosts argue collapses under its own premise.

Context

The “permanent underclass” was Rahul’s long-requested post-processing segment on Episode 33 of the ADI Pod, drawn from Fernando Borretti’s “No One Escapes the Permanent Underclass”. The narrative in its usual form: if you don’t work as hard as humanly possible to get inside one of the AI companies, you will be sorted into a permanent underclass — perpetually poor, reliant on government handouts — while a “permanent overclass” of AI-company insiders captures everything as machines take the jobs. It’s the doom flavor that quietly underwrote a lot of grind-culture rhetoric, though Rahul notes it has died down somewhat as companies angle for IPOs and prefer not to sound apocalyptic.

Borretti’s argument is a reductio. If AI genuinely does everything, the overclass has no function either. The old aristocracy earned its place by supplying something the system needed — funding the military, providing officers. In a world where AI does all the work, and in the limit controls the means of coercion too, what would a human overclass provide? Draw the pyramid Borretti implies and it’s green all the way up: it’s all AI. Membership at the top buys you nothing, because there’s nothing left for a human there to do. Even perfect alignment doesn’t save the structure — an aligned AI that serves everyone dissolves the scarcity the class hierarchy was built on.

Why It Matters

Shimin’s read is that the permanent-underclass logic never actually cohered — reason it out and it doesn’t hold together — and Borretti’s value is naming why with clarity. The practical payoff is career advice: the framing tells people to grind infinitely or be discarded, and if the premise is incoherent, so is the advice. It’s the pessimistic bookend to the same episode’s more grounded jobs data — the EY survey showing CEOs backing off the AI-headcount-cut thesis and Ford rehiring the engineers it had automated away — and a reminder that the loudest AI-and-jobs narratives are often the least examined. See the AI developer careers guide for the data-grounded version of the same question.

Related Episodes