Kevin Kelly believes this is a good time to enter the field of artificial intelligence—but not because of the hype over ChatGPT, the A.I. chatbot that gained more than 100 million monthly active users in just over 60 days after its November launch.
Instead, Kelly—a futurist, author, and the founding executive editor of Wired magazine—thinks that from the perspective of several decades hence, there are no A.I. experts today. That means those entering the field now, or in the near future, will have plenty of time to make their mark.
“What we tend to call A.I., will not be considered A.I. years from now,” Kelly said in an interview shared this week in Noahpinion, an Substack blog from economics commentator Noah Smith. “One useful corollary of this is that from the perspective of looking back 30 years hence, there are no A.I. experts today. This is good news for anyone starting out right now, because you have as much chance as anyone else of making breakthroughs and becoming the reigning experts.”
Today, millions of people are starting to use “generative A.I.” tools such as ChatGPT and Dall-E 2, which can produce remarkable results from simple text prompts. The latter can produce sophisticated images, while the former can succinctly answer questions, pen essays in a particular writer’s style, or generate computer code in various programming languages, among other tasks.
This first round of “primitive A.I. agents” are “best thought of as universal interns,” Kelly said. People are using them to write rough drafts, suggest code, summarize research, brainstorm, and so on.
A.I. whisperers
Even so, getting high-quality results from them isn’t necessarily easy.
“It takes an extremely close intimacy to get your intern A.I. to help you produce great work,” Kelly said. “Some people are 10x and 100x better than others with these tools. They have become A.I. whisperers.”
Indeed, one way to get started in A.I. is as a “prompt engineer.” (The Washington Post recently profiled one.) These individuals devise prompts—or chains of prompts—to get A.I. systems to produce a desired result for a wide variety of purposes and industries.
It’s hardly an established field. One startup advertising for a prompt engineer, Anthropic, noted in its job listing: “Given that the field of prompt-engineering is arguably less than 2 years old, this position is a bit hard to hire for.” The salary range is $175,000 to $335,000.
The company, which makes ChatGPT rival Claude and last month lured a $300 million investment from Google, asked applicants to demonstrate their skills by “spending some time experimenting with Claude…and showing that you’ve managed to get complex behaviors from a series of well crafted prompts.”
The company described large language models—on which A.I. chatbots like ChatGPT and Claude are based—as “a new type of intelligence” and said “the art of instructing them in a way that delivers the best results is still in its infancy.”
Artificial intelligence overall is in its infancy—deeply so, Kelly believes.
“The long-term effects of A.I. will affect our society to a greater degree than electricity and fire,” he says, “but its full effects will take centuries to play out.”
