cc: Life Science Podcast

• Jeremy Elser - Biology and LLMs Meet in the Lab

  At the Advanced Lateral Flow Conference, I spoke with Jeremy Elser, Head of Science Operations at Palantir and founder of Ship of Theseus, a biotech company tackling longevity and regenerative medicine. The name refers to the Greek thought experiment about replacing every plank on a ship over time, similar to Jeremy’s vision to keep rebuilding the human body, replacing the cellular “planks” so it stays functional indefinitely.He’s focusing on restoring the body’s natural ability to regenerate using resident stem cells. Damage accumulates linearly throughout life, but aging accelerates when our capacity to replace that damage falters. His company aims to “re-up” that regenerative capacity, thus the metaphor of the Ship of Theseus .Jeremy also spoke at the conference about using AI and large language models (LLMs) to break down complex scientific questions into smaller, solvable ones. This conversation was fascinating to me in regards to both the biology and the LLMs, discovering what’s possible with both.Jeremy compared an LLM to an eager intern—smart, well-informed, but needing structure and direction. You can’t just hand it a huge problem like “design a new drug protocol” and expect perfection. But if you break that into smaller, ordered tasks like “find existing injury models,” “suggest positive controls,” “compare published protocols”, the system can produce remarkably intelligent, end-to-end workflows.That approach mirrors how good scientists think. Start with clear purpose, choose the right model for the goal, and use well-established methods when you need confidence or novel ones when you want to show something better. It’s part strategy, part rigor, driven by intention. Using an LLM to see where your FDA submission meets (or doesn’t) guidelines seems a relevant example.With respect to biology, Jeremy’s team applies that rigor to wound-healing research involving Hox genes, a class of master regulators that pattern the body during development. He explained how HoxA3, in particular, seems tailor-made for wound repair. It repolarizes macrophages from their inflammatory “angry” state to a regenerative one, promotes vascular growth, and helps skin cells migrate to close the wound. In his words, it “hits wounds in three different ways.” The same gene that once told your embryo where to put your head or feet can later tell adult cells how to heal. I find this phenomenon somewhat magical and hope to someday learn how that works at a molecular level.On the AI front, Jeremy’s biggest insight was about preserving scientific context. He’s using AI to capture and structure what scientists actually do in the lab so knowledge doesn’t walk out the door when people leave. Instead of asking scientists to fill endless forms, the AI reads what they write, asks clarifying questions, and turns messy notes into structured data. The AI will generate every possible graph or chart based on the data, something most scientists would rather avoid. They can then find the ones that are interesting and discard the rest. Jeremy says,  Yeah, that’s my bribe to the scientists ‘cause we enforce a little bit of structure. They have to obey the LLM when it asks for more information. So we try to compensate for that time by doing some of the grunt work that they don’t enjoy doing, like producing a bunch of charts.Fair enough.Jeremy wants AI not just to help scientists think faster but to help us see how it thinks so we can decide what to trust. His view is that LLMs already resemble a kind of brain: opaque, pattern-driven, capable of reasoning, but not always able to explain why. It turns out humans are no different. Jeremy shared an interesting example. You’ll have to listen to the episode for that. Beyond the fascinating biology for me the takeaway (and in line with my own experience so far) is that the usefulness of LLMs goes way beyond answering questions or producing content. As I learned from Jeremy Utley of Stanford, using them as a teammate or collaborator is where the value lies. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cclifescience.substack.com

  --------  

  20:16

  --------

  20:16
• ALFC Double Feature - Making Lateral Flow Accessible Everywhere

  This episode is a double from my visit to the Advanced Lateral Flow Conference. Usability is Innovation: Atomo DiagnosticsAtomo Diagnostics set out more than a decade ago to solve a surprisingly human problem in diagnostics: complexity. Founder John Kelly describes how even the best rapid tests—validated in pristine lab environments—often fail when they reach the real world, where people have no training, and shaky instructions. That gap between laboratory precision and real-world usability has huge implications for reliability, trust, and ultimately regulatory approval.Atomo’s core insight is simple: most errors in point-of-care testing aren’t biological—they’re behavioral. The accessories people use in the field (cheap pipettes, dropper bottles, uncalibrated parts) invite mistakes, and the more steps required, the higher the failure rate. Kelly and his team approached the problem the way a designer might: observe how real users behave, then engineer around human nature instead of fighting it.To validate their approach, they went straight to the source—literally to the community—conducting studies in Africa with low-literacy users who received only picture-based instructions. “If it needs a lot of explanation, it’s probably not obvious,” Kelly notes. The goal: build a device that is self-explanatory and self-correcting.Their solution, the Pascal platform, integrates every accessory needed to run a test—lancet, blood collection, and buffer reagent—directly into one cartridge. Instead of multiple steps and parts, users simply collect, press, and go. Each step is interlocked to prevent mistakes; for instance, the reagent button won’t activate until blood is correctly loaded. It’s engineering that enforces proper sequence, eliminating user doubt and waste.Kelly describes how this design delivers the right volume, in the right order, every time—removing the “what if I did it wrong?” anxiety that undermines confidence in results. It’s the difference between a reliable diagnostic and a false sense of security.Atomo’s HIV self-test—registered with the World Health Organization and distributed across Australia, Europe, and the UK—has demonstrated greater than 99% concordance between trained and untrained users. The company also supports a blood-based pregnancy test (approved in Europe and Brazil) that detects earlier than urine tests, and they’re now developing the world’s first active syphilis test, capable of distinguishing between current and previously treated infections.What’s equally smart is their business model flexibility. Recognizing that many manufacturers already have validated lateral flow cassettes on the market, Atomo developed a “clip-on” usability upgrade that integrates their collection and buffer technology without requiring full retooling or revalidation—a bridge between old workflows and modern design.Beyond infectious disease, Kelly sees growth in at-home wellness and chronic condition monitoring—everything from testosterone and thyroid tests to celiac screening. The platform’s adaptability makes it attractive for home use and clinical trials alike. One example: a pharmaceutical partner using Atomo’s device to monitor liver toxicity in patients remotely, reducing clinic visits from three times a week to “only when needed.” It’s better for patients, cheaper for healthcare systems, and faster for research.The bigger story here is that usability is innovation. Kelly’s approach turns workflow design into a driver of impact. Instead of chasing exotic chemistry, Atomo focused on reliability and trust—two things that ultimately decide whether a test makes it into people’s hands.As diagnostics and healthcare move increasingly into the home, Atomo’s design philosophy feels ahead of its time. If the pandemic taught us anything, it’s that people can and will take responsibility for their health—if we give them tools that make sense.Pitch Competition Finalist: EAZEBIOI also sat down with Ying Chen, founder of EAZEBIO, one of the Innovation Award finalists. Her company’s portable strip-based diagnostic platform combines CRISPR and AI to bring precision health to everyone, especially in low-resource settings.The Problem: Reactive HealthcareYing opens by explaining the fundamental flaw she sees in today’s healthcare system—it’s reactive. We wait for symptoms to become severe before acting. EAZEBIO’s mission is to shift the paradigm toward proactive, precision healthcare, emphasizing early detection and personalized intervention. Her team focuses on diseases often overlooked at the root-cause level—metabolic, autoimmune, and cardiovascular conditions.Their aim is to bridge the gap between scientific breakthroughs and universal access, translating biomarker data into actionable health insights. As Ying puts it, “We hope proactive, personalized care can provide health equity for everyone, no matter where they live.”Ying’s background is a blend of pediatrics, research science, and business—she holds both a PhD and an MBA. Her experience inspired her to adapt the power of CRISPR from the lab to the home.In their prototype for sepsis detection, EAZYBIO’s system uses CRISPR to identify antimicrobial resistance genes—the genetic clues that reveal which pathogen is causing an infection. The test also detects human protein biomarkers, providing a two-layered view of infection and host response.Here’s how it works:* The CRISPR complex acts like a molecular “scissor,” recognizing and cutting specific DNA or RNA sequences associated with infection.* These sequences are tagged with a cortisol-based reporter. When the CRISPR cut happens, cortisol is released.* The released cortisol binds to split reporter proteins, generating a visible signal on a lateral flow strip.* An AI-powered app then reads and interprets the signal into a semi-quantitative result.This approach achieves roughly 300x signal amplification compared to conventional lateral flow assays—crucial for fast, reliable results.Sepsis is notoriously time-sensitive; treatment delays of more than three hours can dramatically increase mortality. Ying emphasizes that EAZEBIO’s platform could enable clinicians to identify pathogens and select the correct antibiotic within one hour—a potentially life-saving improvement.While sepsis is their initial target, the underlying platform is modular and scalable, enabling future multiplexing for 3–5 pathogens per test. Beyond acute disease, the same technology could support early cancer detection and wellness testing, making high-quality diagnostics as easy as a home pregnancy test.Ying speaks with humility about being a finalist at ALFC, but it’s clear the recognition validates EAZEBIO’s bold vision. The conference gave her valuable exposure to peers across R&D and manufacturing, as well as insights into where diagnostics are heading over the next decade.Her takeaway? Collaboration and accessibility matter just as much as innovation. “It’s not just technology—it’s about bringing care to everyone, whether they live in a big city or a rural village.” This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cclifescience.substack.com

  --------  

  29:06

  --------

  29:06
• The Future of Science Marketing: AI, Upskilling, and Human Connection

  The Future of Science Marketing: AI, Upskilling, and Human ConnectionI recently had a conversation with Isabel Verniers, partner at the Marketing Technology and Innovation Institute (MTI2) in Belgium, about how life science companies are adapting to rapid technological change while maintaining their scientific rigor.One of the interesting threads we explored was the tension between scientific and commercial mindsets. Scientists are trained to seek certainty and perfection before moving forward but in a fast-moving market, that perfectionist approach can become a liability.Isabel explained that the key is helping R&D teams become comfortable with a different kind of rigor that embraces uncertainty through assumption testing and rapid iteration. It’s about applying scientific principles differently in a commercial context.“The innovation story is that the R&D part and the commercial part need to nicely blend and avoid that it becomes this valley of death for innovations,” Isabelle noted. The goal isn’t to turn scientists into salespeople, but to help them expand their considerable expertise into market-facing activities.This is where the concept of “minimum viable products” often creates friction. For engineers and scientists, anything “minimal” can feel uncomfortably imperfect. The solution is to focus on assumptions and validation. By mapping out which assumptions are truly critical (requiring extensive testing) versus those that can be validated quickly, R&D teams can maintain their rigor while operating at market speed.We also explored how AI is reshaping market research through “synthetic personas” - AI-generated archetypes built from vast datasets that can help validate ideas earlier in the development process. While some companies eagerly embrace these tools, others remain skeptical. The divide often comes down to that same comfort level with uncertainty.What fascinates me is how AI is becoming less a replacement for human insight and more an amplifier of it. As Isabel pointed out, tasks can be automated, but human skills like critical thinking, empathy, and pattern recognition are becoming more valuable, not less. She reframes these as “power skills” rather than soft skills.The tools get more sophisticated, but the core challenge remains: How do we help brilliant technical minds connect with the market in ways that feel authentic to their training and values?A few key takeaways for science marketers:1. Build “commercial acumen” through small steps that respect scientific rigor while expanding comfort with market-facing activities2. Use structured assumption mapping to help R&D teams engage earlier without feeling they’re compromising standards3. Think of AI as augmentation rather than automation - it’s most powerful when amplifying human insight and creativity4. Focus on developing “power skills” that machines can’t replicate - deep listening, empathy, critical thinking, and pattern recognition5. Create regular “drumbeat” rhythms for market engagement rather than one-off initiativesThe conversation reminded me that while tools and technologies evolve rapidly, the fundamentals of human connection remain surprisingly constant. Our job as science marketers isn’t to strip away scientific rigor - it’s to help translate it into market impact through better storytelling and engagement.As Isabel put it, “ It’s about smart validation. It’s not about quick and dirty.” That’s something I think all of us in life science marketing can rally around.Let’s keep exploring how we can blend scientific precision with commercial adaptability. The companies that figure this out will be the ones that not only survive but thrive in bringing breakthrough innovations to market.And check out Isabel’s book (along with Nuno Camacho): https://thetalentadvantage.org/What’s your experience with this balance between scientific rigor and commercial agility? How are you using new tools like AI while maintaining the human element? I’d love to hear your thoughts in the comments. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cclifescience.substack.com

  --------  

  38:19

  --------

  38:19
• The Art of Protein Engineering

  Carter Mitchell, Chief Scientific Officer at Kemp Proteins, brings scientific rigor and an artist’s imagination to the world of protein design and production. In this episode, recorded at the Advanced Lateral Flow Conference, we explore how his company is pushing the boundaries of protein expression, quality, and analysis using tools that merge machine learning, automation, and human creativity.A company reborn through precision and innovationKemp Proteins has deep roots in recombinant protein production, tracing back over 30 years to a company that began with insect-cell expression systems. After a rocky acquisition phase, the company was revived with renewed focus under CEO Mike Keefe, this time with a modern quality management system and new emphasis on antibodies and engineering solutions for diagnostics, therapeutics, and vaccines.Carter, a self described protein nerd, joined around that time, bringing expertise in structural biology, protein engineering, and quantitative analytics and a mission to integrate AI into the company’s core processes.Why insect cells still matterI knew that people used insect cells but I didn’t know why. Mitchell explains how insect cells, long used in protein production, still offer unique advantages. Unlike E. coli, insect cells can perform post-translational modifications, such as glycosylation—key for producing proteins that resemble their natural human counterparts. While mammalian systems like HEK293 have since made expression “paint-by-numbers” simple, Carter notes that insect systems still excel when complexity and authenticity matter. “It’s about having multiple expression capabilities,” he says, “so you can choose the right one for the problem at hand.”Four questions that guide every projectCarter’s approach to solving client challenges starts with four questions:* What is the protein?* What information is available?* What’s the intended use?* What’s the scale?From there, the team tailors both the process and the system to ensure reproducibility and regulatory readiness, whether the goal is a diagnostic reagent or a therapeutic protein. As an aside, manufacturing kilograms of protein still blows my mind.As Carter puts it: “Regulators don’t want to see a smear on an SDS page. We think like regulators, anticipate their questions, and design out variability before it becomes a problem.”From data lake to digital expert: ProtIQThe centerpiece of Carter’s innovation is ProtIQ, an internal expert system that combines structured data, AI models, and domain expertise into a 200–300-page report for every target protein. Initially, these reports were for experts, but Carter’s team is now transforming them into an interactive chatbot interface so anyone on the team can query the data conversationally.“If a technician can ask, ‘What’s the isoelectric point?’ or ‘Does it have a secretory tag?’ and get an immediate answer, they’re empowered,” he says.It’s part of a broader effort to turn technicians into scientists, helping them engage more deeply with data, notice anomalies early, and contribute to process improvement.Predicting protein liabilities before they happenUsing sequence analysis and AI-assisted visualization, Kemp Proteins can predict potential degradation sites or stability issues before production even begins. Carter’s team also models how viral variants like influenza strains might evolve over time, identifying changes in glycosylation patterns that could impact diagnostic binding. “We’re actually collaborating with the FDA on this,” he adds.When science meets artCarter looks at protein structure like art. A lifelong painter and flamenco guitarist, he traces his fascination with structure to his mother’s art studio and his childhood encounters with crystals in Texas soil. That visual mindset drives how he thinks about molecules: “Art flattens multi-dimensional space to describe motion. That’s what we do in AI and machine learning, flattening complexity into something interpretable.” This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cclifescience.substack.com

  --------  

  20:35

  --------

  20:35
• Storytelling vs. Math: How Narratives Shape Biotech Investment Decisions

  When it comes to biotech innovation, data and logic should drive decision-making, but sometimes stories take the wheel. This week, I spoke to Julien Willard, a board advisor and strategist for biopharma leaders. He had written a post on LinkedIn about how biologics may be favored over small molecules when logic suggests otherwise.Despite perceptions, small molecules still make up 73% of all FDA approvals from 2009 to 2023. They’re faster and cheaper to develop with median costs around $2.1 billion vs. $3 billion for biologics and easier to manufacture and scale. Yet, they’re increasingly overlooked.Why? Three systemic distortions drive this imbalance:* Regulatory bias. The Inflation Reduction Act grants 13 years of market exclusivity for biologics but only 9 years for small molecules — the so-called “pill penalty.”* Narrative premium. Investors are captivated by stories that sound futuristic. Saying “we’re reprogramming immune cells to fight cancer” sounds far more thrilling than “we designed a molecule that blocks a kinase.”* Flawed valuation models. Risk-adjusted NPV calculations rely heavily on peak sales assumptions and exclusivity duration, favoring high-priced biologics even when they serve smaller populations.The result is a market that systemically favors expensive therapies and leaves affordable innovation underfunded.Julien’s critique is a call for narrative accountability. He’s seen investor rooms go silent not when presented with data, but when shown the story of an 8-year-old patient whose tumor vanished. Emotion drives attention. “We’ve created an industry where the best storytellers get funded, not necessarily the best science.”This bias ties to a deeper cognitive flaw, biotech’s “narcissism of outliers.” Every founder believes they’ll beat the odds. Despite historical data showing 60% of Phase I drugs fail, CEOs say, “Ours is different.” Investors, too, often prefer that hero narrative because it promises 100× returns.So how do we rebalance the equation?Julien points to portfolio biotechs, including firms like BridgeBio that apply modern portfolio theory to drug development. Instead of betting on a single compound, they fund multiple related programs. Ten projects with 20% success odds each yield a 90% chance of at least one success, diversifying to reduce risk. This requires different incentives: rewarding teams for portfolio success instead of single-asset hype. “Right now,” Julien says, “we’re training biotech CEOs to optimize for the biggest headline, not the best outcome.”I feel like this shift also runs against America’s “blockbuster mindset.”As Julien notes from his European perspective, U.S. business culture celebrates the lone hero, not the collective success. That mentality seeps into funding, regulation, and storytelling alike.We shifted gears to antibiotics, a perfect example of market failure that will require cooperation for success. Developing them makes no financial sense: they’re used sparingly to avoid resistance, and priced cheaply against generics. Yet antibiotic resistance is a national security threat. Programs like BARDA and CARB-X aim to fix this through “decoupled incentives”: guaranteed government purchases and funding for early-stage platforms. Julien sees this as a template for fixing other market failures, from rare diseases to mental health.Could AI restore the balance? Julien argues that the best use of AI in drug development is reducing bias and noise in decision-making.Today’s biotech investing, he says, is “embarrassingly primitive”. Billion-dollar bets are made off 30-page decks and gut instinct. AI could process far more variables- molecular data, patient subsets, regulatory shifts to make more rational decisions.But we’re using it wrong: optimizing existing systems instead of reimagining them. Real innovation will come when AI enables adaptive trials that learn and evolve in real time to detect human bias rather than amplify it.I have some thoughts on how storytelling can still win: Let’s amplify the narrative that there is a better way to do things that could actually deliver better results for everyone:* Lower risk for investors through diversified portfolios* More accessible treatments for patients* Better alignment between scientific progress and public health needsNext week (Oct 13-15), I’ll be moderating a panel at the Advanced Lateral Flow Conference in La Jolla, CA. Use code LSMR25 to save 25% on your registration.. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit cclifescience.substack.com

  --------  

  37:40

  --------

  37:40

More Science podcasts

Trending Science podcasts

About cc: Life Science Podcast

cc: Life Science Podcast: Podcasts in Family