BioHive Week's U Startup Launchpad brought together founders, researchers, entrepreneurs and students at the University of Utah for a day of hard-won wisdom, from a biotech pioneer's emotional farewell to a pair of AI founders reshaping what it means to build lean.

Salt Lake City, Utah — April 17, 2026

Part summit, part mentorship session, part community gathering, the second annual U Startup Launchpad filled the Utah Museum of Fine Arts on the University of Utah campus Thursday as the capstone morning event of BioHive Week. Hosted jointly by the University of Utah Technology Licensing Office and BioHive — the state's nonprofit life sciences coalition — the one-day program drew researchers, founders, investors, and students united by a shared conviction: that science-based companies are one of the highest-leverage things a person can build.

The sessions ran the gamut from the deeply personal to the bracingly practical. A fireside conversation with Chris Gibson, the outgoing CEO of Recursion, traced the human arc of company-building across fifteen years. A panel of AI-native founders offered a more immediate kind of instruction: how to actually compete right now, with the tools available today, in a world where the rules of the game are being rewritten in real time.

Together, the two sessions made for a morning that was equal parts inspiration and tutorial.

Chris Gibson on Building a Startup: A Most Beautiful and Terrible Voyage

Jenna Barbari, BioHive's Senior Communications Director and a former Recursion employee herself, took the stage alongside Chris Gibson for what felt less like a formal interview and more like a conversation between two people with a shared history and genuine affection for each other.

Gibson co-founded Recursion in Salt Lake City in 2013, building it from a PhD-lab idea into one of the most closely watched companies in the drug discovery space, one that uses high-throughput biology and machine learning to map the relationships between biological perturbations and human disease at a scale previously unimaginable. He has since stepped down as CEO and is moving on to his next chapter.

Barbari asked when the science became a company.

"We thought about the business on day one," Gibson said. "Even before we started the company."

The story he told was disarmingly funny. Still a PhD student in Dean Li's lab at the University of Utah, Gibson was sent by his mentor to Stanford Ignite — a summer program for scientists with startup ideas, which he described as "basically a little MBA for scientists." There over six weeks, he and Ron Alfa, M.D., Ph.D., who would become part of the Recursion founding team, began sketching what would become Recursion — then called Fourth Dimension Therapeutics. At the end of the program, participants pitched to a room of Sand Hill Road venture capitalists.

"I kid you not — literally one of them fell asleep," Gibson recalled. "And they were tech VCs. We did not tell the story well."

He still has the slide deck. The point, though, wasn't the failure — it was the attempt. "If you're building a technical company, you can't defer the thinking of the business," he said. "You can change it, you can pivot it, but it has to be part of your plan from the very beginning."

From there, the conversation turned to culture. Gibson's voice shifted. He spoke about the environment that shaped him in Dean Li's lab, one defined by an almost paradoxical combination: exceptionally high standards and exceptionally deep care for people.

"I used to go into the lab at two in the morning because I wanted to beat Dean into the office," he said. "And Dean was there at two, three, four AM every single day."

But what he remembered most wasn't the hours, it was an act of quiet humanity. A woman who worked in the grants office had a husband who suffered a heart attack. Dean Li showed up at the emergency room at two in the morning, in his full lab coat, and stayed with her through the entire experience. She was one of roughly forty or fifty people in the lab.

Gibson paused. He got emotional. He collected himself.

"That kind of humanity — alongside a high bar — that was the culture I grew up in with Dean," he said. "And so I wanted to replicate that at Recursion."

It was the thing he felt he got right. The thing he felt he got wrong was harder to admit. In the early days of the company, he said, the balance tipped too far. He let time with his family slip more than he should have. He read extensively about leadership and management, but not enough poetry. He didn't spend enough time with old friends who could ground him.

"At the end of the day, it's all about people," he said. "And one of those people is yourself."

His closing words were addressed directly to anyone in the room considering starting a company: "It is the most beautiful and terrible voyage you will ever go on. You will have some of your best moments and many of your worst and hardest moments. So don't forget to take care of not only the people who build it along with you — but also yourself."

Barbari, who has been on the receiving end of Gibson's personal check-ins during difficult times, noted that he made those calls even when Recursion had grown to five or six hundred people. The room was quiet in a way that suggested the message had landed.

Building AI-Enabled Companies: The Practical Reality

If Gibson's session was a long view of the founder's journey, the next panel was a live wire — a fast-moving conversation about what it actually looks like to build a company today, when AI tools are rewriting the cost structure of startups on a near-weekly basis.

Moderating was Penny Atkins, Associate Director of the University of Utah's One-U Responsible Artificial Intelligence Initiative. On stage with her were two founders currently doing the work.

Warren Pettine, M.D., is the co-founder and CEO of MTN (Medical Timeseries Networks), a Salt Lake City company building a clinical workflow orchestration platform for hospitals. Pettine spent over a decade in computational neuroscience — training at Harvard, Stanford, NYU, and Yale — before applying that same analytical framework to healthcare, where staff shortages are reaching crisis levels. MTN's system integrates wearable and telemetry data with electronic health records to help care teams identify where attention is needed most, without attempting to diagnose patients or replace clinicians. MTN was covered by TechBuzz in December 2025.

Carmen Kivisilid is the co-founder and CEO of Elnora AI, which builds AI tools for biomedical lab protocols — helping preclinical researchers with tasks like literature search and results analysis. A molecular neuroscientist by training, she did her PhD at UCL and a postdoc at King's College London before leaving academia, returning to her native Estonia, and eventually making her way to Utah through the Altitude Lab startup program. She had applied, in part, because she had long followed Chris Gibson's work — and was offered a fellowship by Gibson and his wife that enabled her to relocate to the US.

Both founders are running what they describe as fully AI-enabled operations, and both had stories that illustrated the stakes of making that shift — and the resistance that comes with it.

Kivisilid described taking her team from nine people to five, and then from five to two — letting go of team members who refused to use AI in their daily work. It wasn't a comfortable process.

"I had people saying, 'I went to university, I studied fifteen years of computer science — I know better than AI,'" she said. "And I was like, you're right, you do. But let's use these tools to empower you." When they wouldn't, the decision became unavoidable. "Now one hundred percent of our code is written by AI. Our contracts — AI. Our website — AI. Everything. I work harder than I've ever worked before, because AI enables me to do so many things."

Pettine framed it through the lens of cost structure — a phrase he returned to often. "It starts and ends with cost structure," he said. The companies that will struggle to compete, he argued, aren't necessarily the ones that don't build AI — it's the big incumbents that produce great AI models but can't change how they operate internally. "There's a lot of internal turmoil at some of those large companies right now, because there's a difference between producing awesome AI models and actually competing within a lower cost structure."

For a small founder, the advantage is real. Pettine described being at a conference a few weeks prior, typing feedback from conversations into his phone throughout the day, then letting coding agents run overnight to build a new demo — which he was showing to potential customers by the next morning.

Kivisilid matched that with her own example: she had ten minutes before a client meeting. She gave an agent a prompt asking it to use Figma to build presentation slides. She pressed enter nine minutes before the meeting. She reviewed the first version five minutes before walking in, gave feedback, and presented those slides. "That's all possible now," she said. "But not everyone has the skills for that."

Which is why, she was quick to add, accessibility matters. She and her collaborators are running free workshops on April 29th and May 1st for people who want to get started — beginning with something as basic as setting up a computer and opening a terminal. She described a recent workshop with a biotech client in South Texas: fourteen brilliant scientists, two hours to set up their machines, and people walking out of the room shaking their hands in frustration. "I was in shock," she said. "I needed months to get where they were trying to get in two hours. But you have to start somewhere."

Pettine compared learning AI tools to learning Excel — uncomfortable, time-consuming, and completely unavoidable. "Just suffer through it. Get to know GitHub," he said. "These tools are the new Excel, the new PowerPoint."

The conversation turned to responsibility, and both founders gave honest answers. Kivisilid was blunt about the limits of guardrails: "You can write it in a system prompt. But it's so easy to trick AI into doing bad things. Anything that works on human psychology works on these models." The solution, she argued, is less about restricting AI and more about educating humans. The same tool she built to help design lab protocols could, theoretically, be misused, and the barrier to entry for that misuse is lower than it has ever been. "We have to train humans not to do harm," she said.

Pettine took the more operational angle: data governance, he argued, becomes a company-critical priority far earlier than most founders expect. MTN uses Vanta, a platform for compliance and penetration testing, to stay ahead of security requirements. Kivisilid noted that Elnora — a two-person company — is already SOC 2 Type 1 certified, something that typically takes months and enormous resources. She credited AI agents with completing the certification process in a record six weeks. Her auditor, she said, had never seen it done so fast. "I gave our agents access via CLI," she said. "That's how we did it."

"Don't necessarily focus on selling AI," Pettine advised the audience. "Use it to enable operations in a huge greenfield, not just in new markets, but mostly in existing, boring markets. The incumbents are locked into pricing models and cost structures that make them susceptible to disruption. Just jump in."

To make his point about cost structure and AI-enabled disruption, Pettine cited MEDVi — a Los Angeles-based GLP-1 telehealth startup founded by brothers Matthew and Elliot Gallagher that generated $401 million in its first full year and is projecting $1.8 billion in 2026 revenue, with a headcount of two. A New York Times profile published April 2 held it up as a real-world proof of Sam Altman's long-discussed prediction that AI would eventually enable a one-person billion-dollar company. As Pettine summarized the lesson: don't sell AI — use it to run operations in markets where incumbents are locked into old cost structures.

Pettine's example landed cleanly in the room. However, as it turns out the MEDVi case is a bit more complicated, considering the FDA's warning letter over concerns about its marketing of compounded GLP-1 products. Investigative reporting revealed the company had used AI to generate fictitious doctor profiles, fabricated testimonials, and deceptive before-and-after images in its advertising. MEDVi's own website acknowledges that individuals appearing in its ads may be AI-generated portrayals.

Nevertheless, the company has not been found guilty of any violations, and FDA warning letters are advisory rather than final enforcement actions. The episode adds a layer of complexity to the panel. In a session where Carmen Kivisilid warned that "you can use AI to come up with a bioweapon," and where both founders stressed that guardrails ultimately depend on human ethics rather than technical controls, MEDVi ends up being a more instructive example than Pettine might have intended — for better or worse.

A Launchpad, Not Just a Conference

The U Startup Launchpad sits at an interesting moment in Utah's innovation ecosystem. BioHive — now five years old — has grown the scope of its annual week considerably, with events spanning investor summits, student programming, startup showcases, and the Best of BioHive Awards Gala. BioHive Live, the flagship conference, moved to the Loveland Living Planet Aquarium in Draper this year, as covered by TechBuzz, to accommodate a projected 900 attendees — nearly double last year's sellout crowd.

The U Startup Launchpad is more intimate by design, and it is aimed specifically at the researchers, grad students, and early-stage founders who are still deciding whether to make the leap from bench to company. What the morning's two sessions offered, side by side, was a kind of stereo signal: the human frequency from Gibson, who has traveled the full distance from PhD student to publicly traded company and back to whatever comes next; and the high-frequency pragmatism of Kivisilid and Pettine, who are deep in the trenches right now, building, adapting, and figuring it out week by week.

The message, remarkably, was the same from both directions. Build intentionally. Use every tool available. Take care of the people around you. And take care of yourself.

The U Startup Launchpad is presented by the University of Utah Technology Licensing Office in partnership with BioHive. For more on BioHive and upcoming events, visit biohive.com.