-
Jensen Huang said that a $500,000 engineer should use at least $250,000 in tokens or risk setting off alarm bells.
-
“If that person says $5,000, I’ll imitate something else,” the Nvidia CEO added.
-
Huang suggested giving engineers AI tokens worth about half of their annual salary.
Jensen Huang raises a new red flag for top talent: Not enough AI tokens are being used.
The Nvidia CEO said on an episode of “All-In Podcast” released Thursday that he would be “deeply shocked” if one of the chip giant’s top engineers spent too little on artificial intelligence.
“I would be shocked if that $500,000 engineer didn’t consume at least $250,000 worth of tokens,” Huang said.
“At the end of the year, that engineer is worth $500,000, I’m going to ask them how much did you spend on the token? If that person says $5,000, I’m going to imitate something else,” he added.
Asked whether Nvidia would spend $2 billion on tokens for its engineering team, Huang said: “We’re working on it.”
“It’s no different than one of our chip designers saying, ‘Guess what? I just use paper and pencil,'” he said, referring to top engineers who underutilize AI tokens.
Earlier this week, Jen-Hsun Huang said at the GPU Technology Conference that tokens could become part of his engineer recruiting strategy.
“They will make hundreds of thousands of dollars a year, which is their base salary,” Huang said of the engineers. “I’d probably give them another half as tokens so they can amplify 10x.”
“It’s now one of the recruiting tools in Silicon Valley: How many tokens will my job bring?” Huang added. “The reason is very clear, because every engineer who has access to the token will be more efficient.”
Tokens are the basic units used by artificial intelligence systems to process text. The more text the AI reads or generates, the more tokens are consumed, which is why companies typically charge based on usage per thousand or million tokens.
Huang isn’t the only one arguing that engineers need to make heavy use of AI computing, and that companies should be willing to pay for it.
Business Insider’s Alistair Barr reported earlier this month that tech companies may be trying a new way to compete for talent: offering AI reasoning capabilities alongside salaries, bonuses and equity.
Tomasz Tunguz of Theory Ventures told Barr that tokens are a potential “fourth component” of compensation.
Peter Gostev, head of AI capabilities at Arena, a startup that measures model performance, recommended that OpenAI and Anthropic “create job boards where clients can post job ads, listing a nominal budget for the position, as well as a salary range.”
Thibault Sottiaux, director of engineering on the OpenAI Codex team, wrote on X that candidates are increasingly asking how much computing power they will receive.
OpenAI CEO Sam Altman has previously said that AI tokens could become a form of universal basic income.
“I wonder if the future is more like Universal Basic Computing, rather than Universal Basic Income, where everyone can share in the calculations of GPT-7,” Altman said on “All-In Podcast” in May 2024. “They can use it, they can resell it, they can donate it to someone for cancer research, but instead of the dollars you get it, like a slice, you own a piece of the productivity,”
Read the original article on Business Insider