Eli Lilly CEO says he has ‘at least 1 or 2 AIs running’ during every meeting he’s in

• Eli Lilly CEO David Ricks says AI helps him stay up to date on the latest science.

• Ricks said OpenAI’s ChatGPT is “too verbal” for those types of queries.

• He wants something that is “more terse” and more reliable with its references.

Eli Lilly CEO David Ricks says he uses AI in every meeting he attends.

“I read a lot of medical journals. I go to conferences where data is presented,” Ricks told Stripe cofounder John Collison during a recent episode of Collison’s “Cheeky Pint” podcast. “I spend time with our scientists to stay curious. Yeah, now I have at least one or two AIs running every minute of every meeting I’m in, and I just am asking science questions.”

Ricks said he doesn’t like OpenAI’s ChatGPT for science-related questions — “It’s too verbal,” he said. Instead, he prefers Anthropic’s Claude and xAI’s Grok. Still, he has to be careful to watch for halcunications, an issue the frontier model companies are still trying to tamp down.

“I find it more terse and the references actually check out more often,” he said. “Sometimes the AIs produce references, and they’re actually not the thing that it said, and that takes too much work to go cross-reference.”

xAI CEO Elon Musk quickly took notice of the praise.

“Cool that David Ricks uses @Grok as his daily AI advisor,” Musk wrote on X.

Ricks is just the latest big-name CEO to reveal his AI diet. Microsoft’s Satya Nadella said he uses Copilot to summarize his Outlook and Teams messages after reaching Microsoft’s Washington headquarters. Nvidia CEO Jensen Huang said he uses AI as a tutor.

Lily has been on a tear this past year, with shares up roughly 31%. The drugmaker has capitalized on sales of its GLP-1 weight-loss drug Zepbound and diabetes treatment Mounjaro.

AI still has a way to go when it comes to helping drug development, Ricks said.

“Probably we need to create the equivalent of what got created with human language, which is a more complete repository of biological knowledge to train against before the machines get a lot better,” he said. “And today, I don’t know, I would estimate we might know 10 to 15% of human biology, so the machine is not going to be good at all until we get way above 50%.”

To even reach that point, Ricks said there would need to be a significant investment in robotics to create the training data needed to teach AI.

“That probably requires robotic 24/7 experiments just to create training data sets and this kind of big lift effort, the kind of thing actually NIH should be doing right now, I would think,” he said. “But that effort’s not ongoing, at least in our country.”

Read the original article on Business Insider