Gabriel of the Libre Solutions Network has an uncomfortable question for the AI industry: have we been here before?
His new essay, published March 11, 2026, draws a direct line between today's AI adoption frenzy and the social media gold rush of the 2010s — a period that promised connection and productivity and delivered surveillance capitalism and monopoly lock-in instead. It has circulated widely enough in free software circles to be worth attention from anyone following the agent space.
The sharpest observation is what Gabriel calls "Gell-Mann's Apathy" — a twist on the Gell-Mann Amnesia effect. Programmers, he notes, are rightly skeptical of AI-generated code. Artists question whether generative images have any real substance. But those same skeptics often drop their guard entirely when AI is applied outside their own domain. Well-founded doubts never accumulate into the broader skepticism they should produce.
Vibe-coding gets particular attention. Gabriel frames these agentic loops — where AI iteratively generates and error-corrects code until something compiles — as computing by brute force. The resource cost is real; it's just that end-users rarely see it. The computational bill gets quietly socialized.
None of this makes him a techno-pessimist. He separates the current commercial AI landscape from what a freedom-respecting alternative might look like: open hardware, open software, genuine user autonomy. His argument isn't that AI is inherently wrong, but that the AI we actually have is centralized, monopolistic, and structurally incompatible with privacy. And that by the time most people recognize the trade-off, the window for a different path may have closed.
The essay is available on the Libre Solutions Network site — and, pointedly, via Tor and I2P as well. That last detail is probably the most efficient summary of where Gabriel stands.