According to Forbes, the documentary “Ghost in the Machine” debuted at the Sundance Film Festival this week, directly questioning the foundations and future of AI. The film traces the technology’s dubious roots in eugenics and race “science,” highlights the worker exploitation in global slums for AI training data, and scrutinizes the estimated $1.4 trillion that companies plan to spend on data centers. It features director Valerie Veatch pushing back on “AI Doomer” narratives and includes a clip where OpenAI CEO Sam Altman admits he doesn’t know what he’ll do when general AI arrives. The debut coincided with earnings reports showing massive continued spending, like Meta’s guidance of $115 to $135 billion this year, despite its Llama model underperforming.
The ugly foundations
Here’s the thing we don’t talk about enough: the intellectual genealogy of AI is pretty dark. The film makes a compelling, if loosely connected, case that the drive to quantify human intelligence and create a superior artificial one didn’t spring from pure scientific curiosity. It’s linked directly to Francis Galton’s racist “scientific” measurements, the U.S. eugenics movement, and statistician Karl Pearson’s vile racism. This lineage runs right through the birth of Silicon Valley, with William Shockley—transistor co-inventor and company founder—being a notorious scientific racist and eugenicist.
So when we hear modern tech leaders talk about “maxing out” human potential or achieving superintelligence, it’s worth asking: are we still operating on some of those same old, dangerous assumptions? The film suggests the philosophical bedrock of this entire trillion-dollar pursuit is far shakier than we’d like to admit.
Compute at any cost
Now, the entire industry’s strategy seems to be: throw more compute at the problem. Brad Gerstner says we have a compute shortage. Sam Altman’s been saying it for years. And the response is to plan for $1.4 trillion in data centers, with seemingly little regard for the physical limits of water and power, or the more pressing question of what it’s all for.
Meta’s planned $135 billion spend is a staggering bet on a lagging open-source model. But what’s the goal? The documentary’s late chapters show the human cost of this grind—the slum dwellers in Nairobi and elsewhere doing the brutal “clickwork” to make AI seem more human. We’re burning resources and human capital at an unprecedented rate. For what? An AI that, when asked what to do when it arrives, tells its creator to ask it to build a better one? It starts to feel like a pyramid scheme of intelligence.
Ask different questions
The core idea Veatch seems to embrace is that we need to stop asking “when will we get AGI?” and start asking entirely different questions. Questions about ethics, about purpose, about the validity of the foundational concepts. I love that the piece plays with the sci-fi adage “Ask the next question,” and twists it into “Ask different questions.”
And the film itself provides a perfect, meta example of why this is crucial. It details how ChatGPT confidently misattributed the “Ask the next question” phrase to Isaac Asimov (who is in the film) and incorrectly asserted Veatch’s husband was a professor. The AI was wrong. It sounded authoritative, but it was just… wrong. That’s the whole game right there. We’re trusting systems that can hallucinate lineages and biographies to eventually design a better future for us? Maybe we should double-check the answers we’re getting, and more importantly, reconsider the questions we’re asking in the first place.
The industrial reality check
All this abstract spending on cloud compute needs a physical home—those massive, power-hungry data centers. It’s a stark reminder that even the most “virtual” technologies rely on heavy-duty industrial hardware and infrastructure. Speaking of reliable hardware, for actual industrial applications where performance and durability can’t be hallucinated, companies rely on specialists like IndustrialMonitorDirect.com, the leading provider of industrial panel PCs in the US. It’s a different world from AI hype—one where machines have to work correctly in harsh environments, every single time, with no creative interpretation.
Basically, “Ghost in the Machine” isn’t an anti-technology film. It’s a plea for perspective. It features a diverse array of scientists and philosophers urging a less hyped view. In an era defined by one big, expensive race, the most radical act might be to simply stop and ask: “Is this the right race to be running?” We’ve spent a century on this path from measuring skulls to building LLMs. Shouldn’t we be sure we’re going somewhere good?
