The LinkedIn co-founder and Manas AI backer says doctors who skip frontier models as a second opinion are bordering on malpractice, and he wants the FDA using AI too.
Reid Hoffman has a blunt message for doctors who aren’t yet running their toughest cases past an AI: you might be flirting with malpractice.
Speaking at WIRED Health in London on April 16, the LinkedIn co-founder and Manas AI co-founder argued that frontier large language models, the kind built by OpenAI and Anthropic have become powerful enough that ignoring them in clinical practice is itself a risk.
“If as a doctor, you’re not using one or more frontier models as a second opinion, my belief is you’re bordering on committing malpractice,” Hoffman says.
It is a provocative line, especially from someone who sat on OpenAI’s board, helped fund the lab in its early days, and now runs an AI drug discovery startup of his own. But Hoffman framed it less as a tech-industry talking point and more as a practical patient-safety argument.
“These AI systems, even though many of them are not specifically trained for medicine, have ingested trillion-plus words of information,” he says. “As a second opinion, it is bringing superpowers that no human being has.”
The case for an AI second opinion
Hoffman’s pitch is not that AI replaces the doctor. It’s that AI sits next to the doctor, quietly checking the work.
He told the WIRED Health audience that he personally uses frontier models as a second opinion for his own health, and insists that his personal concierge doctors do so as well. The doctor can disagree with the model, he says, but they should at least be in the conversation.
“You could very well go, ‘No, I think you’re wrong, I think it’s this,'” Hoffman says. “But if you’re not using this as a second opinion, you’re making a mistake, both as a doctor and as a patient.”
The framing arrives at a delicate moment. A major study earlier in the year concluded that large language models present real risks to members of the general public seeking medical advice, citing inaccurate and changeable information. Hoffman’s answer to that is essentially structural: keep the LLM in the loop, but keep the human in charge.
He extends the same logic to systems under strain. With NHS waiting lists stretching and family doctors in short supply, he sees an LLM medical assistant on every smartphone as a kind of early triage layer, a way to sort what actually needs a human appointment.
“We just don’t have enough doctors, most people don’t have access, and when you think about, ‘How should the NHS be redesigned?’ everyone should be interacting with this medical assistant,” Hoffman says.
Why Manas AI is the backdrop
Hoffman is not making the argument from the sidelines. In January 2025, he co-founded Manas AI alongside Pulitzer Prize-winning oncologist Siddhartha Mukherjee, who serves as the startup’s CEO. The company launched publicly with a $24.6 million seed round led by General Catalyst and Hoffman, with Greylock participating, according to TechCrunch.
Manas AI is building an AI engine that aims to fast-track drug discovery for various cancers, with an initial focus on aggressive cancers including triple-negative breast cancer, prostate cancer, and lymphoma. The pitch is simple: shift drug discovery from a decade-long process to one that takes a few years.
“We believe this approach can shift drug discovery from a decade-long process to one that takes a few years,” Hoffman and Mukherjee wrote in a joint blog post reported by Fierce Biotech.
The division of labor inside the company is just as telling. Mukherjee, the cancer physician, reviews the AI engine’s proposals and sifts genuinely interesting candidates from the “bonkers stupid” ones. Hoffman is, by his own framing, the AI guy. The bet is that neither side works alone.
“It isn’t just the best of science and it isn’t just the best of AI, because either of those two are insufficient,” Hoffman told CNBC. “You need to put those two together.”
Hoffman wants AI pushed further still – into the regulatory layer, where he’d like to see the FDA and other agencies use biological models to assess emerging medicines and fast-track promising drugs. He concedes that’s not coming soon.
“As a Silicon Valley person, I would love to get to a point where the FDA was also running tests with biological models, going, ‘Oh, we should fast-track this one, because the likelihood of negative consequences is lower,'” he says. “Do I think that’s anytime soon? Unfortunately, no.”
The longer arc is what Hoffman seems most interested in. Manas AI starts with cancer, but he believes AI discovery engines could eventually identify drug candidates for chronic and extremely rare diseases too.
“I think in 10 years, every major disease will have target molecules that could at least make a serious difference,” he says.
For doctors, the nearer-term challenge is smaller and more immediate. Whether to open a second tab.
