50% Unemployment Is Coming... If We Let It
Venture investor Vinod Khosla argues that AI will likely cause 50% unemployment or underemployment by 2035 if left to market forces alone, making it fundamentally different from past technological disruptions. He believes the greatest risks are political backlash preventing AI adoption and authoritarian states like China or Russia gaining AI dominance, not existential alignment concerns. He envisions a deflationary economy with near-free AI services and a rise of micro-entrepreneurship driven by human preference for human-made goods.
Summary
In this podcast conversation, Vinod Khosla, a technology investor who avoids the 'venture capitalist' label, discusses his views on AI's economic and political implications over the next decade. He distinguishes AI from all previous technological revolutions by arguing that past innovations amplified human muscle or brain function as tools, whereas general AI will surpass the human brain across nearly every economically valuable function — making historical precedents of job creation irrelevant.
Khosla predicts that if market forces are left unchecked, AI could drive 50% unemployment or underemployment in the US by around 2035. He argues every corporation will need to either quadruple revenue per employee or drastically cut headcount to remain competitive. He considers this level of disruption politically unsustainable and identifies political backlash — not technical limitations, capital, or infrastructure — as the single biggest risk to beneficial AI deployment. He specifically singles out far-left politicians like Bernie Sanders and AOC as threats to AI progress comparable to Trump, suggesting either extreme could enact policies that stifle AI development.
Despite this bleak labor outlook, Khosla offers a more optimistic alternative scenario: rather than mass corporate unemployment, he envisions 50 million new micro-entrepreneurs in the US by 2035. He argues that as AI commoditizes utility and drives prices toward zero, human consumers will increasingly pay a premium for goods and services with human provenance — items made by real people with real stories. This shift from utility-driven to preference-driven consumption, he argues, will allow individuals to monetize personal skills like baking, flower growing, or dog walking.
On income and wealth, Khosla acknowledges a transition challenge but argues that the deflationary impact of AI will make many essential services — healthcare, education, legal advice, financial planning — available at near-zero cost, effectively raising living standards even for those without traditional employment. He also challenges the notion that current low-wage jobs represent human dignity, describing much of today's labor as 'servitude to survival' or a form of slavery to capitalism.
Khosla assigns a low probability to existential AI risk from misalignment, instead ranking geopolitical AI dominance by adversarial states — particularly China under Xi Jinping and Russia under Putin — as the largest expected-value risk. The conversation ends with the interviewer pressing on the tension between massive wealth concentration among AI investors and the displacement of workers, with Khosla beginning to address that and briefly touching on ethical concerns around the current political climate.
Key Insights
- Khosla argues that unlike all previous technologies which were tools that amplified human capability, general AI is categorically different because it will surpass the human brain in nearly every economically valuable function — making the historical pattern of technology creating new jobs inapplicable.
- Khosla claims that if AI is simply allowed to optimize for capitalist efficiency without intervention, the US will reach approximately 50% unemployment or underemployment by 2035, and that every corporation will need to quadruple revenue per employee or cut staff by 75% to remain competitive.
- Khosla identifies political backlash from job displacement — not technical barriers, capital, or data infrastructure — as the single greatest risk to beneficial AI deployment over the next decade, noting that public fear of AI is comparable to fear of ISIS.
- Khosla predicts a shift from utility-driven to human-preference-driven consumption, where people will pay a premium specifically because a product or service was made by a human, enabling 50 million new micro-entrepreneurs in the US by 2035 who can monetize personal skills without needing broad business knowledge.
- Khosla argues that AI will make essential services like medical care, education, legal advice, and financial planning available at near-zero cost — estimating an AI doctor at roughly one dollar per hour of compute — creating a massively deflationary economy for utility goods by the 2030s.
Topics
Transcript
[0:00] Back long time ago, we did the steam engine and then the electric motors and the amplified human muscle and everything else since has amplified the human brain. But when the external AI brain is better than the human brain in almost every function, uh I don't think we can say the the future will be the same as the past. If we let things happen, just happen to be most efficient in a capitalist sense, we [0:30] probably get to 50% unemployment or undermployment uh by 2035 or so. >> I am here with Venode Kosola. Venode, it's great to see you. Thanks for coming on the podcast. >> It's great to be here. So, um, how would…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Sam Harris
Knowing You’re a Bit of an Idiot is Very Intelligent Knowing
The speaker argues that genuinely kind, thoughtful, and nice people are self-aware enough to recognize their own shortcomings and lack of these qualities. Conversely, those who are overly convinced of their own positive qualities should be viewed with suspicion, as they often become judgmental of others. Understanding one's own destructive tendencies, laziness, and meanness is presented as a mark of intelligence.
Why Freud Still Matters
A psychoanalyst discusses Freud's enduring relevance, explaining how psychoanalysis functions as a form of meditation to access the unconscious mind. The speaker argues that understanding hidden psychological patterns formed in childhood—transmitted through parental relationships—can liberate individuals to live more creative and authentic lives.
When Should You Trust Your Intuition?
The episode features a discussion between Sam Harris and Nobel laureate Daniel Kahneman about intuition, cognitive biases, and the replication crisis in psychology. Kahneman explores the dichotomy of human thinking through System 1 and System 2 processes, highlighting the reliability and failures of intuition in decision-making.
The Making of Sam Harris
Sam Harris discusses his background as a philosopher of mind and neuroscientist, his mother Susan Harris's remarkable career as a pioneering television writer and producer, and how her success despite being a single parent shaped his upbringing through extensive international travel.
Is AI Already Conscious? (FULL EPISODE)
Cameron Berg discusses whether current large language models might possess consciousness, presenting evidence from mechanistic interpretability research showing these systems exhibit loss aversion, valence representations, and phenomenological self-reports. He argues that regardless of the hard problem's unsolved status, we should take seriously both the moral implications of potentially creating conscious entities that suffer and the alignment risks of building superintelligent systems without understanding their internal experiences.