This post was originally published on this site
https://content.fortune.com/wp-content/uploads/2023/06/GettyImages-1146124693-e1686156907969.jpg?w=2048Ken Griffin, the hedge fund billionaire, is simply one of the biggest names in finance. His Citadel hedge fund had a record-breaking $16 billion return last year, when many were wrong-footed by a brutal bear market. Now A.I. has the market soaring, but Griffin is again twisting where others are turning. “I do think the A.I. community is making a terrible mistake by being full of hype on the near-term implications of generative A.I.,” he said at an internship kickoff event in Fort Lauderdale on Tuesday. “I think they’re actually doing everybody a huge disservice with the level of hype they are creating.”
The hype is clear to see. Marc Andreessen, a billionaire just as prominent as Griffin but in the venture capital space, wrote a 7,000-word manifesto the same day as Griffin’s comments, about how A.I. will “save the world.” In fact, he said it could be greater than sliced bread. “A.I. is quite possibly the most important—and best—thing our civilization has ever created, certainly on par with electricity and microchips, and probably beyond those.”
For his part, Griffin addressed a popular prediction about generative A.I. these days: that it will make a ton of jobs, especially of the white-collar variety, obsolete. He said the threat of white-collar job culling is exaggerated, because ChatGPT and its ilk are still not accurate enough to do most professional work.
“If you listen to the CEOs of tech companies, it’s going to eliminate millions of white-collar jobs,” Griffin said. “I say, ‘Not that fast.’ Some professions are accepting of errors, but you have to be really accurate in finance. You have to be really accurate as a lawyer.” Griffin may have been alluding to the recent instance of a lawyer submitting a brief that he prepared with the help of A.I.—only to find (in court) that it cited nonexistent case law.
The hallucination problem
To be sure, Andreessen may well be right that A.I is the most significant technological breakthrough since the internet—or maybe ever—but to Griffin’s point, it has major issues. A.I. commonly fabricates information and presents it as real, something that techies refer to as “hallucinations.” Generative A.I. is a text predictor, which puts together words via a sophisticated algorithm trained on internet content. It has no way of ensuring the veracity of its statements, which is why it can’t compete with many real professionals right now, according to Griffin.
So people shouldn’t panic about losing their jobs, Griffin said—unless they’re coders. Professions that require high attention to detail and accuracy are shielded by A.I.’s current fickleness, but tech jobs are at high risk, Griffin said. A turning point for computer work is fast approaching, and young coders would be wise to adapt accordingly, he added.
“Programming is going to be a big target for generative A.I.,” Griffin said, when asked by an intern about A.I.’s impact on Citadel. “You want to make sure that if you are a software engineer, you are putting yourself really close to the domain problems that need to be solved. Your career path will be defined by your ability to solve problems. The days of ‘I’m a good programmer’ are becoming numbered.”
Griffin’s Tuesday comments come on the heels of a June 1 report that nearly 4,000 tech workers lost their jobs to A.I. in May. The sector this year has had the most year-to-date layoffs since the 2001 dotcom bubble burst.
IBM is already planning to pause hiring for non-customer-facing roles that can be automated. “I could easily see 30% of that getting replaced by A.I. and automation over a five-year period,” the tech company’s CEO Arvind Krishna said in a May 1 interview.
Nvidia’s CEO Jensen Huang predicted that A.I. would cause a “rebirth of the computing industry,” and like Griffin said that early career workers could avoid layoff by integrating the tech into their work as quickly as possible.
Citadel is currently negotiating to get a companywide ChatGPT license, and Griffin has said in the past that the tool will automate many tasks at the hedge fund.
Peng Zhao, CEO of Citadel Securities, the market maker that parallels Griffin’s Citadel hedge fund, has also expressed plans to integrate A.I. into his company’s workflow.
“Being liberated from the repetitive, mental laboring aspect of what we have to do as humans, that allows humans to focus upon the highest-value portion of planning,” Zhao said at a conference in May.
This article has been updated to correct multiple references to Citadel Securities to Citadel. The entities are separate and the latter is the hedge fund Griffin founded.