Singularity University and XPRIZE Founder: Many Ask What AI Will Do for Us, Instead of Asking What We Will Do with It
Singularity University and XPRIZE founder Peter Diamandis stated: Many are asking what AI will do for us, rather than asking what we should do with it. This reflects his consistent framing of tool selection as a life fork.
In recent years, he has reiterated this stance in communications: The cost of GPT-3.5 level queries has dropped from about $20 per million tokens in November 2022 to about $0.07 in October 2024, approximately 280 times cheaper; a Pew survey shows that about 65% of American workers rarely or never use AI in their work. He advocates for "picking up tools to build, write, and teach" and, along with Alexander Wissner-Gross, promotes a "Solve Everything" abundance timeline, identifying food, energy, education, and health as four pillars that autonomous systems can reduce costs for by 2035. He cites DeepMind's Demis Hassabis and Anthropic's Dario Amodei as supporting evidence for overcoming chronic diseases and extending lifespan within the next decade.
Institutional experience supports this optimistic engineering: XPRIZE buys results with prizes, Singularity University teaches exponential technologies as courses, and Zero-G and A360 sell experiences and communities. In his communications, he cites the annual growth of semiconductors (Micron, Intel, AMD, etc.) as evidence of "inner circle" investments and views the valuation leaps of OpenAI, Anthropic, and xAI as proof of computational power turning into equity. Concurrent programs continue to interview Robinhood, United Therapeutics, etc., incorporating tokenization, organ regeneration, and agent settlement into the same abundance narrative.
Who is buying, who is selling: Buyers are purchasing the agency of "the subject is still human," while sellers are trading attention stuck in victim narratives. The event is driven by a stance, with funding flowing towards the entrepreneurial circle willing to subscribe to a ten-year timeline, rather than new model training budgets. Beneficiaries are those selling courses, competitions, and communities, while those under pressure are regulatory discourses treating safety discussions as the only legitimate issue. Asking "What will it do for you" positions people as objects; asking "What will you do with it" positions people as principals.
The public post lacks new experimental data and does not disclose his personal model usage.
Source: Public information
ABAB AI Insight
Diamandis transforms fear-based narratives into tool-based narratives, encapsulating the XPRIZE logic: first reward results, then let participants find their own means. In his view, artificial intelligence is not the subject but an outsourced intelligence that costs as little as coffee. The 65% of workers who do not use it are portrayed as contestants who have yet to sign up.
The capital pathway is selling exponential stories to those who want to be principals. Competitions, summits, communications, and interviews form a funnel, collecting attention at one end and pointing towards semiconductors, model companies, and longevity biotech at the other. Valuation leaps are used as teaching molds: those who control the inner loop of turning electricity into intelligence rewrite the boundaries of nations and companies. He acknowledges that superintelligence will redraw the social contract, but the prescription remains for individuals to pick up the query box rather than to halt training first.
Benchmarks include Kurzweil's singularity timeline, Thiel's "competition is for losers," and Yudkowsky's calls for halting training. Diamandis chooses abundance over pause, viewing underutilization as the greatest risk. The industry phase shifts from laboratory safety papers to widespread tool adoption: after cost collapses, the fork is no longer about whether to create, but whether to use.
Structural judgments belong to technological substitution. The mechanism is: when the price of reasoning drops by orders of magnitude, an object mindset wastes already amortized computational power; a subject mindset transforms the same computational power into personal leverage. Safety issues have not disappeared; they have merely been reframed from "how machines will deal with humans" to "whether humans are present to direct machines." Those who are not present will become optimized variables in others' workflows.
ABAB News · Cognitive Laws
Objects ask about fate, subjects ask about usage
When tools are cheap, not using them is the greatest risk
Competition thinking first buys results, then debates whether machines are benevolent