Against AI Utopianism
Artwork (Unknown) from Cosmos
The current public discourse surrounding artificial intelligence has been thoroughly hijacked by a tech-utopian narrative. Figures like OpenAI’s Sam Altman and Meta’s Mark Zuckerberg command the headlines, positioning their proprietary models as the ultimate panacea for humanity's most intractable crises. We are told that by surrendering our societal foundations to these “next-token” predictors, we will unlock rapid breakthroughs in molecular biology to cure chronic diseases, deploy complex algorithmic systems to (ironically) engineer climate repair, and automate economic workflows to achieve unprecedented efficiency.
This promissory rhetoric conceals a profound and destabilizing shift in engineering methodology. Modern AI is not really an exploration of cognitive mapping or human understanding. It is the raw culmination of a statistical turn that began decades ago (during the Chomsky era, as I discussed with Bill Bain) and materialized fully with the introduction of the 2017 transformer network blueprint. By utilizing massive computation to process immense volumes of text, these systems do not possess “intelligence” so much as they are highly optimized mathematical engines designed to predict the next likely word in a sequence.
Predatory commercialism now dominates this landscape, completely overshadowing genuine scientific inquiry and masking a devastating material reality. To sustain the immense computational force required by these massive models, tech conglomerates are colonizing the earth with data centers, imposing a severe and extractive tax on small towns across the United States and the global south. As we have become well aware, these massive, concrete warehouses consume vast amounts of local electricity and divert millions of gallons of water from municipal resources to cool humming server racks. This creates a huge strain on local grids and inflates utility costs for everyday residents. The industry speaks poetically about "the cloud" — a deception designed to obscure the ecological destruction happening on the ground. There is a pressing necessity to decouple this hardware footprint from local communities, to force a transition toward decentralized and self-sustaining computational networks before the physical cost of digital intelligence bankrupts the planet's remaining resources.
Meanwhile, the popular cultural anxiety that this technology is diminishing human intellect partially misses the mark. Cognitive degradation is not an inevitability of the tool, but a direct consequence of passive consumption. When treated as an intellectual crutch for automated plagiarism or mindless convenience, the technology certainly erodes our capacity for critical thought, but when used as an active research instrument, it functions as a powerful cognitive lever. For the autodidact, it can synthesize disparate fields of knowledge into a personalized curriculum, adapting to individual cognitive styles to teach complex subjects on demand, democratizing the ability to scale a business by automating overhead that previously required massive capital. The machine does not always dull the human mind — instead, it reflects and magnifies the level of intellectual curiosity or laziness that the user brings to the screen.
