Reid Hoffman Thinks Doctors Should Give AI a Second Opinion

After three years work led by some of Silicon Valley’s most powerful companies — cofounding LinkedIn and sitting on the PayPal and OpenAI boards—Reid Hoffman recently started thinking about medicine.

Hoffman’s startup, Manas AI, is building an AI engine that aims to speed up the slow process. getting medicine of various cancers. Inspired by a dinner with renowned cancer doctor Siddhartha Mukherjee, the company’s founder and CEO, mission statement and “changing the drug discovery process from a decade to a multi-year one.”

But Hoffman’s interest in creating AI output, in particular, goes beyond what has previously been discussed with small molecules. He believes that frontier models—the most advanced, largest AI models currently available from companies such as OpenAI and Anthropic—should be the foundation of healthcare itself.

“If as a doctor, you are not using one or more borderline models as a second opinion, my belief is to go through the process of making mistakes,” said Hoffman, speaking at WIRED Health in London on April 16. “These AI machines, although many of them were not directly trained in medicine, entered a trillion-plus words of information.

Such comments would no doubt anger many doctors. Earlier this year, a great lesson concluded that dominant languages ​​are dangerous for people seeking medical advice because of their tendency to convey wrong and distorted messages.

Hoffman’s argument is that instead of giving AI models the ability to think critically, humans should use them as an additional source of information, which he believes can prevent misdiagnosis. He says he uses borderline models as a second opinion on matters related to his health and insists that his concierge doctors do the same.

“You can go, ‘No, I think you’re wrong, I think this is it,'” he told the WIRED audience. “But if you don’t use this as a second thought, you’re making a mistake, both as a doctor and as a patient.”

With the UK’s National Health Service facing waiting list and staffing challenges, including chronic shortage of doctorsHoffman believes there is a growing need for a large-scale language model that can serve as a free medical assistant for any mobile phone. He said it could serve as an early stage test for human doctors.

“We don’t have enough doctors, too many people don’t have access, and when you think, ‘How should the NHS be reformed?’ everyone should contact this health care provider,” he said.

Although he has a conflict of interest as an entrepreneur working on drug discovery, Hoffman is also eager to see AI play a greater role in assisting the FDA and other regulators in evaluating new drugs, as well as speeding up the availability of reliable drugs especially for patients.

“As a Silicon Valley guy, I would love to get to the point where the FDA is doing tests with natural brands, saying, ‘Oh, we should follow this one up quickly, because the potential for side effects is low,'” he said. “Do I think soon? Unfortunately, no.”

As for Manas AI, human judgment still plays an important role in the company’s decisions regarding which goals to pursue. Mukherjee reviews their ideas for an AI engine, Hoffman says, and filters the interested candidates from the “stupid bonkers.”

Although the company’s focus is on cancer, Hoffman believes that the potential of AI predictive engines is much greater, helping to develop cures for chronic and rare diseases that have previously been unaffordable for drug companies to research.

“I think that in 10 years, every major disease will have molecules that can make a big difference,” Hoffman said.