Are humans and technology in competition with each other? Can computers do whatever humans can do? Will AI take over the world?
According to Peter Thiel, the author of Zero to One, humans and technology don’t compete with each other—they complement each other. In the job market, computers don’t eliminate the need for human workers; they just empower human workers to be more productive. And, in the resource market, computers don’t compete with humans for resources, because computers are not consumers.
Here’s why the concern of competition is irrelevant.
Humans and Technology
In Thiel’s view, the concern that technology might compete with us for resources stems primarily from a popular misconception that computers can be trained to do anything that humans do. He acknowledges that many computer scientists, particularly in academia, have devoted significant study to the problem of teaching computers to perform tasks that would otherwise be done by humans, but he points out that if anything, these studies only highlight the fact that humans and computers excel at different kinds of tasks. Even a low-end computer can solve arithmetic problems thousands of times faster than the world’s leading mathematicians. But by the same token, even a child easily outperforms the world’s leading supercomputers at tasks like object recognition and value judgments.
Thiel argues that developing new technology is the solution to the problem of competition. He addresses the concern that technology itself might begin competing with humans for resources and offers his perspective on artificial intelligence and its capabilities.
Leveraging Complementary Capabilities
Furthermore, Thiel asserts that because humans and computers excel at different kinds of tasks, businesses can combine their respective strengths to provide unique capabilities. As an example of this, he discusses Palantir, a software business that he co-founded.
The premise of Palantir was that computers excel at filtering large amounts of data based on objective criteria, and identifying simple patterns, while human analysts are much better at figuring out what patterns in the data really mean or how they can be useful. Palantir’s software provides human analysts with a tool that can sift through vast amounts of data and flag patterns or items meeting certain criteria for human inspection.
Thiel recounts how government analysts used Palantir’s software to uncover fraud, insider trading, and child pornography rings; predict trends in the spread of food-borne disease; and even warn of insurgent attacks during the war in Afghanistan. He clarifies that the software did not identify these insights by itself. It was merely a tool that enabled human analysts to gain these insights.
Taking Advantage of Differences
According to Thiel, computer scientists in academia frequently overlook the obvious differences in human and machine capabilities: Instead of taking advantage of these differences to create computer software that will enhance people’s productivity, they struggle to create algorithms that can mimic human capabilities. The field of artificial intelligence (AI) in particular suffers from this tendency.
Extrapolating advances in AI to the extreme, some people anticipate that computers could someday become better than humans at everything and take over the world. This kind of “strong AI” could result in either a utopian society or an apocalyptic scenario, depending on how superhuman AIs decided to treat humankind.
Thiel doesn’t absolutely rule out that possibility, but he contends strong AI is so far beyond the state of the art in computer technology that it’s not worth worrying about in the 21st century. For now, we should focus on leveraging the advantages of computer technology and developing the next generation of technological breakthroughs.
———End of Preview———
Like what you just read? Read the rest of the world's best summary of Peter Thiel's "Zero To One" at Shortform .
Here's what you'll find in our full Zero To One summary :
- Why some companies genuinely move the world forward when most don't
- How to build a company that becomes a monopoly (and why monopolies aren't bad)
- Silicon Valley secrets to selling products and building rockstar teams