It appears there are three letters that make Marc Benioff raise his eyebrows.
Salesforce has taken a strong, AI-first turn in recent years. Whether AI will reach “AGI” — an ill-defined milestone generally understood to be when AI can perform complex, human-level tasks — its CEO is less confident.
On the “20VC” podcast, host and investor Harry Stebbings started by referencing a recent AI talent story. Amazon’s San Francisco AGI lab head recently said that there were fewer than 1,000 people in the world that would be “extremely valuable contributors” to building a new frontier model.
Before Stebbings could finish his question, Benioff interjected: “AGI head — that sounds like an oxymoron,” he said.
“You’re talking to somebody who is extremely suspect if anybody uses those initials, ‘AGI,'” Benioff said. “I think that we have all been sold a lot of hypnosis around what’s about to happen with AI.”
The AGI debate has grown rapidly in Silicon Valley, with tech leaders debating about if and when it will come — or whether it’s already here.
Sam Altman recently said that GPT-5 was not yet AGI because it couldn’t learn “continuously.” Meanwhile, Altman’s company OpenAI has been warning investors that money may become obsolete in a post-AGI world.
Demis Hassabis, the CEO of Google DeepMind, said that a lack of consistency was keeping AI back from general intelligence.
For his part, Benioff thought that we were a while away — but he’s not ruling out that it “couldn’t happen one day.”
“We’ve all seen those movies,” Benioff said. “Peter Schwartz, who wrote ‘Minority Report’ and ‘WarGames,’ works for me. He’s our chief futurist.” (Schwartz was a consultant on both films. Scott Frank and Jon Cohen wrote “Minority Report.” Lawrence Lasker and Walter F. Parkes wrote “WarGames.”)
“But I just realize that isn’t the state of technology today,” Benioff said.
Benioff joins a number of tech moguls who are skeptical of the hype around AGI.
In a recent New York Times op-ed, former Google CEO Eric Schmidt and analyst Selina Xu wrote that Silicon Valley had “grown so enamored” with AGI that it was “alienating the general public and, worse, bypassing crucial opportunities to use the technology that already exists.”
Others have cautioned against jumping to doom-and-gloom predictions about what will happen if the milestone is reached.
David Sacks, the White House’s crypto czar, said that the AGI “apocalypse” was “overhyped.”
Benioff is still very keen on AI. He called large language models “state of the art.” He also said he understood why, when using LLMs, it can feel humanlike.
“But it’s not a person, and it’s not intelligent, and it’s not conscious,” Benioff said. “It hasn’t suffered, it doesn’t have compassion.”
Read the full article here
