AI Extinction Risks Loom Behind Big Tech’s Pursuit of Profit
By Imani Sutton · Reporting from Atlanta ·
While researchers warn of human extinction, the industry continues to prioritize shareholder value and luxury hardware.
The most chilling admission in the latest episode of the TBPN podcast isn't about a glitchy chatbot, but a lack of a plan. Evan Hubinger of Anthropic stated he believes there is a greater than 10% probability that AI could kill all humans within the next decade, admitting that his own company does not yet have a plan to solve the alignment of super-intelligence. This is the shadow lurking behind the glossy product launches and the venture capital hype.
Profit as a Hedge Against Extinction
While Hubinger and former researcher Jacob Coxin warn that labs are "gambling with our lives" in an irresponsible race toward self-improving AI, the industry's financial machinery views this volatility as a feature, not a bug. On the podcast, Ryan Peterson argued that the industry's willingness to ignore these existential risks is actually a "bullcase" for stock investment, suggesting a total commitment to profit regardless of the consequences.
The strongest argument for staying in the room comes from Elon Musk, who suggests that those who take these risks seriously should "lean in" to solve the problem rather than "rage quitting." The logic is seductive: the only way to steer the ship away from the iceberg is to be on the bridge. But this assumes the bridge is open to those who prioritize human survival over shareholder value. As the speakers noted, the difficulty of effecting change within massive, profit-driven organizations often makes "rage quitting" the only honest intellectual move left.
The High Cost of the Final Form
While the architects of AI gamble with the species, the marketing arms of Big Tech are selling us the "flex." Apple has announced the "iPhone Duo," a foldable phone expected to cost between $2,000 and $3,000. One speaker on the podcast argued that phones are cheap relative to the time we spend on them, but the reality is more cynical. The hosts described the Duo as the "final form" of the iPhone—a device that represents a significant hardware achievement but "does absolutely nothing new" to move humanity forward.
This pattern is a historical constant. From the early days of the personal computer to the current era of "Muse," Meta’s new AI agent that can book restaurants and buy goods, the trajectory is always the same: wrap a marginal utility in a luxury price tag to maintain growth. Meta’s Muse, which provides users with a sequestered virtual machine and encrypted storage, is less a tool for liberation and more a mechanism for deeper integration into the Instagram ecosystem. It is the same playbook Mark Zuckerberg used at Harvard, where, as the speakers recalled, he referred to the students who trusted him with their data as "dumb fs."
The Ghost of Human Flourishing
The tragedy of the current AI race is the distance between what is being built and what was once imagined. The podcast highlighted a 1985 video of Steve Jobs predicting a tool that could capture a thinker's worldview—like Aristotle's—to help students learn. The speakers suggested Jobs would have demanded that AI grant people new abilities and prioritize original vision over the production of generic "slop."
We are told that we are in a new era, but the record shows we are simply repeating the mistake of prioritizing the tool over the user. The "Steve Jobs of AI" the speakers long for—an optimist focused on human empowerment—cannot exist in a system where the primary goal is to capture a "bullcase" for investors. When the people building the technology admit they have no plan for its safety, the "innovation" is no longer a service; it is a liability.
The industry is operating on the assumption that we can build a god and then figure out how to leash it, all while selling us a $3,000 foldable screen to watch the collapse in high definition. We are not witnessing a technological evolution, but a corporate heist of the future, where the risk is socialized across the entire human population and the profit is sequestered in a few virtual machines.