Alexandr Wang is an MIT dropout who co-founded ScaleAI, an Amazon-backed, data-labeling AI startup valued at $14 billion. He became a billionaire at age 24 and is still the world’s youngest self-made billionaire at 27, with a net worth of $2 billion.
In a conversation with WaitWhat media CEO Jeff Berman and Intel’s Lama Nachman, Wang answered the question, “What are the skills you think young people need to develop today for the economy that’s coming?”
Wang said prompt engineering, or writing prompts for AI chatbots to respond to, was important. Much more than that though, is being able to puzzle through the problem of how AI can achieve more human-like reasoning across longer spans of time.
Related: Goldman Sachs CIO Says Coders Should Take Philosophy Classes — Here’s Why
“I think this is something that humans will always be differentiated in… we’re very good at long-form thinking and very good at thinking over very long time horizons,” Wang said.
Alexandr Wang. Photographer: David Paul Morris/Bloomberg via Getty Images
AI models are currently good at predicting the next step, but can’t process multiple steps accurately, according to Wang.
“They usually make a mistake on the third, or fourth or fifth, reasoning step or chain of thought,” he said.
Wang emphasized that humans will “always” have a meaningful advantage over AI when it comes to thought over longer periods of time and that these were “things to really dig deep on.”
Related: How to Be a Billionaire By 25, According to a College Dropout Turned CEO Worth $1.6 Billion
When it comes to disciplines that emphasize long-term thought, Wang says that math, physics, and other technical fields stand out.
“It’s classic stuff, like doing math, doing physics, doing these technical fields is very important,” he said. “I think there are a lot of fields like economics which force you to think very long-term and think through implications many steps down that are valuable.”
Nachman answered the same question and said that critical thinking and reasoning were valuable skills to have while interacting with AI, especially when AI answers are inaccurate.
Wang also said that most of the AI models out there were trained on “most of the Internet,” which aligns with comments made last month by Microsoft AI CEO Mustafa Suleyman that AI can be trained on almost all content on the Internet.
Related: Microsoft AI CEO Says Almost All Content on the Internet Is Fair Game for AI Training
