Reclaiming the Narrative: What BioPharma Must Learn From Tech Before It's Too Late
Our story is slipping. Tech knows how to tell one. Here's what we need to adopt to tell our story.
Have you noticed we have more tech people in Biotech now? Over the past few years there has been an influx of people from the software world becoming interested in our space. You have the group of longevity-curious tech CEOs bankrolling anti-aging companies (e.g. Bezos, Altman, Armstrong), large tech companies focusing some of their R&D on drug applications (Nvidia, Microsoft, Google), former software people becoming biotech founders (InSitro, XtalPi, InSilico), and traditionally tech/generalist-focused funds funding drug companies.
Why Tech Capital Is Fleeing Software (and Coming to BioPharma)
You don’t need an oracle to tell you that AI has played a major role in this trend. But why now?
The main reason I see is that the trajectory of AI has shown us that things like software—the field these individuals typically come from—is quickly collapsing into a world that will be commoditized. We hear stories every day, discussing how AI can be applied in the hands of their own engineers. A “super engineer” with agentic capabilities that can now deploy hundreds of AI-based engineers and orchestrate development across a number of different coding tasks simultaneously. Companies in the tech sector praise how many tokens their engineers use, asking them to continually use more and more.
On the flip side, we see many stories about how tech companies are now contracting their workforces—getting rid of lower-level engineers, QA, and QC—unnecessary functions that they anticipate will no longer be required in a more fully AI-integrated world.
This was always going to be the end state. AI was invented on the computer and trained on the internet, which in some respects is the most complete example of human thinking that we have. All of these models were coded in languages created by humans. In a sense, coding is perhaps one of the most familiar things to AI models because it is exactly what they are built on. If AI was a car, humans would be the end-to-end manufacturer of it. We built every part, bolted them together, and made it run.
Scale the software future up to its final destination. Why would a large multinational corporation buy enterprise software from someone like Workday or Salesforce when in the very near future—and perhaps even now—they could deploy one engineer to create custom software for their entire organization with the help of an army of AI bots? Following this thread, you can quickly see that software is going to be a commodity in the new world. Businesses (and consumers) will be able to easily and cheaply build the software tools they need.
Perhaps the same could be said of things that we use on a day-to-day basis in biotech strategy: Excel models, slides, curated databases, and other complex documents. It is becoming ever more clear that the value of an individual producing those versus an AI is shrinking considerably. The future of deliverable-based work in our space is going to be more about an individual orchestrating several AI-based “analysts” to produce the necessary deliverables that will feed into a strategic synthesis by the human orchestrator.
So if software is going to be a commodity, it makes sense for tech capital to flee traditional software-based companies and reallocate elsewhere. And where might that be? Well, it’s in areas like biotech, where the rules are far more complicated and there still remains a ton of room for AI to add value. Unlike software, where humans created the underlying language and understand its intricacies well, in the field of biology, we were not born with instruction manuals. By some measures, we perhaps only know 10% of the rules that govern our bodies. Moreover, those rules are exposed to the pressure of evolution; they are dynamic and will likely change over the course of our species. In the car analogy, we are even less than mechanics. In some sense, we are still trying to figure out what we’re actually driving.
As we know, in biotech, it’s really hard to develop novel drugs, another reason why we are seeing a whole host of new companies spin up that are well-funded but focusing on pre-existing targets (aka “well-validated biology”). Areas like biotech, where there are still incredible amounts of unknown, are the perfect places for new AI-focused capital to come into. We’re already seeing the biggest companies in our space discuss how AI is accelerating some of the more low-hanging operational fruit in their businesses.
But quickly we’re seeing the potential of how AI can be used to speed up drug discovery, document writing, finding optimal clinical trial sites, and a whole host of other blocking and tackling type tasks that go into drug development. Don’t get me wrong, we’re not at the era yet where AI can develop a drug “soup to nuts” on its own. That 10% of known biology is still a really big barrier. There’s no replacement for doing true discovery work and pushing the boundaries of what we know about biology to discover new rules that govern our cells, organs, and systems.
There is also no replacement for running clinical trials. We need to test more novel biology in human subjects to get a truer understanding, and more importantly, better data that we can feed back into the multitude of specialized AI models that are coming out. Only then will we get closer and closer to “simulating human biology”.
A Story About a Dog that Illustrated a Narrative Gap
Lately, we've seen some interesting stories pop up in the mainstream news about people in the tech world applying basic AI tools to develop new medicines. The one that stands out the most to me is the story of the Australian tech entrepreneur, Paul Conygham, who used AI tools—including large language models and protein structure prediction software—to help design a custom DNA cancer vaccine tailored to his dog Rosie's tumor. Thankfully, Rosie's health has significantly improved after receiving this bespoke cancer vaccine.
However, the responding commentary on social media was somewhat triggering for me. There was a group of people saying, "Why can't we do this for all cancers?" or "Look, this guy just cured cancer using AI; this completely changes everything and the pharmaceutical industry is upended." Even worse, some took this case as evidence that drug companies are intentionally withholding cures or not applying technology in a patient-centric way.
As insiders, we know that perspective is misguided. Even in Rosie's case, she didn't have a complete response or total eradication of all tumors, and we need longer follow-up to see how effective the therapy was. While it was a remarkable feat, it lacks the context that this work is already being done at a much larger scale in human patients. A number of cancer vaccine studies are already in clinical trials, most notably led by mRNA manufacturers like BioNTech and Moderna, which are in phase 3 studies with promising phase 2 data to support them.
This situation made it clear to me that in biotech, we have an issue communicating the promise of our technologies to a wider audience, whereas tech leaders are able to do this much more expertly. Solving this problem is incredibly critical, especially as we see more tech people entering biotech riding the wave of AI innovation.
What We Can Learn From the Tech Bros
I think it’s important to say that we shouldn’t shun tech folks from coming into biotech. There is much more we can gain by melding the best of our expertise and cultures together.
Tech does well at maintaining a healthy risk appetite, platform thinking, iteration speed, and communicating the value of their innovations. Where some struggle is in shortcutting checks & balances, “growth-at-all-costs” mentality, overconfidence in new domains, and “fake it until you make it” attitude. We don’t want a replay of the Theranos story.
Biotech does well at maintaining high evidentiary standards, putting our customers’ (patients) safety first, and embracing regulatory competence (Vinay Prasad aside). We struggle at maintaining consistent tolerance for high-risk (it’s more seasonal), rethinking how to accelerate “standardized” processes, adopting novel operational technologies, and most importantly, controlling our industry’s narrative.
BioPharma speak tends to change based on the audience we are speaking to. On Linkedin, we see a bunch of virtue signaling and attaboys for our teammates. With financial-facing audiences we are all about EPS, EBITDA, revenue growth, deals, market share, and “guidance”. When most of the leading figures in our space find their way onto mainstream media, they speak from a defensive position, trying their absolute best to provide non-answers and deflect the public’s gaze from the “evil Pharma CEO”.
Tech leaders have mastered a way of communicating the value of their innovations, often painting a grandiose vision for how it will change humanity or solve a problem we didn’t know we had. Twitter isn’t just another social media platform, it’s the “digital town square where matters vital to the future of humanity are debated ”. AirBnB isn’t just a short-term rental marketplace, it’s framed as creating “belonging anywhere” and “ending strangers”. Uber isn’t framing itself as a ride-sharing service, it is delivering “transportation as reliable as running water”.
I think most of us can read those statements on the surface or in reflection of using these products and know in our bones that there is a lot of fluff and BS built into them. I am not saying Biopharma needs to oversell the promise of our innovations. In fact, I worry about losing control of that with the ever-increasing influence of the mainstream tech crowd coming into our sector. We already have many of the AI leaders claiming all diseases will be cured within 5-10 years (link), a cancer vaccine can be developed in just 48 hours (link), and that AI will be able to simulate biology within 3-5 years (link). Those of us working in the space know that those predictions feel very aggressive. What AI is doing today is truly amazing, but it’s playing in a sandbox of known predictable variables. Life science is different. We just don’t know enough about biology yet and we don’t test enough new biology in clinical trials quickly enough to get there without significant overhaul to the regulatory and funding environment.
If I gave you a chess board and only told you what a couple of the pieces did, then asked you to figure out the rest of the game, but every time you got something wrong you had to give me a dollar, I would be hitting the sports book the next day with my new bankroll. How much of the of the game could you figure out if you knew 50% of the rules? Certainly your predictions would get better, but I would guess I am still leaving with a healthy stack of dollar bills to load up into FanDuel account and hit some bets. Now if you knew 90% of the rules, maybe I’d only have enough to buy us some drinks. That’s the duality of AI for drug development. Without great data about how the rules of our bodies, cells, organs, etc. work, it’s not going to be all that useful. But as we furnish it with better data, there may come a tipping point where it gets exceptionally good at predicting the rest.
Polls from groups like Gallup consistently show that the Pharma industry sits lowest or second lowest in terms of favorability ratings. Overpromising and underdelivering isn’t the answer here, but our industry’s reputation isn’t going to change unless we start acting differently.
We are actually building things that tangibly change the direction of humanity, but we need to do a much better job painting that vision for the public. Certainly, that won’t fix everything. We have a lot more on the fronts of drug pricing and conspiracy theories to dispel, but painting a grand, but importantly, a realistic vision, for our medicines’ impacts will go a long way towards rebuilding our image.
GLP-1s are changing the way we think about metabolic disease writ-large. Pretty much anyone who wants to lose weight to improve their health can do so. They aren’t just weight loss drugs, they are “rewriting the metabolic fate of humanity”. Long-acting HIV injectables are bigger than just a convenience upgrade for HIV patients, they “give future generations a chance at a world without HIV”.
Shaping a Symbiotic Future Between Tech and BioPharma
Look - the convergence is inevitable, because it is already happening. Tech and BioPharma will need to be symbiotic. Eventually AI tools will be integrated into every step of the drug development process, and we will need the help of our tech colleagues to get to the next set of transformational medicines. We should embrace their penchant for inverting slow “established” paradigms, but do so in a manner that proceeds with caution. “Move fast and break things” is fine when it’s zeroes and ones, but in our world that could cost patients their lives.
We should adopt their willingness to take on a higher risk threshold at a baseline. I talk to too many investors and company leaders who tell me that the folks who control the flow of money in our industry are afraid of novel biology, new platforms, or taking a fresh-look at modalities that were once written-off. I think if Adam Neumann, the much maligned founder of WeWork, is allowed to raise $100M for another real-estate company, our funding systems should be unafraid to take bolder, more asymmetric, and more contrarian bets whether macroeconomic seas are calm or stormy.
Most of all, we need to continue putting patients first, but also speak to humanity with a reality-grounded message that allows the species-level impact of our innovation to resonate. It’s ironic that technology, at its best, is supposed to serve humanity, but there is fundamental threat that it’s latest and greatest iteration, AI, could instead replace or devalue humanity to great degree. Our industry, BioPharma, is different. It is the most pro-humanity industry there is. It is incredibly essential to the present and future health of our species. We need to do a better job showing our fellow humans why.
