Science-Based Entrepreneurship

Recently, I had the opportunity to share my learnings and insights from "Launching Products Globally" with an amazing audience at Plug and Play Tech Center with the presence of global audience including entrepreneurs from HKSTP - Hong Kong Science and Technology Parks Corporation. Here are a few learnings and insights from the evening: 1) You need to "localize" your product & go-to-market strategy: This doesn't only mean just translating or localizing your product. It's a lot more than that. You need to localize your "go-to-market" motion as well. You may have product-market-fit (PMF) locally, in the first country/region you launched, but that doesn't mean you can take the same product and go-to-market strategy to launch in a new country/region. As an example at Fitbit, we learned how the French think about fitness (they count walking to a restaurant to get a glass of wine as their "fitness") is very different than how Americans define workout and fitness. So all our marketing and go-to-market strategies had to align with the way locals will see benefits in our products. 2) Having boots on the ground is essential for successful global expansion: You need to have boots on the ground who truly understand the nuances of how to go-to-market, how to sell, and how to deliver your value proposition to customers in different regions. There are a lot of nuances of how to do business locally that will take outsiders to any market a long time to learn. At Cleo, where we had global customers like Salesforce, Redbull, Pepsi, and Uber, we had to have local health Guides to deliver our services with an intimate understanding of customers needs and approaches in that region. 3) Understanding local, cultural, and social aspects is critical to a global expansion success: Even though at the surface things may seem similar in each region, there are a lot of nuances that make your go-to-market strategy and the way you deliver your services resonate with the local customers or not. At Teladoc, we've learned that people in different countries think about their mental health and how to get support for that "very differently" than each other. Huge thank you to my hosts Rahim Amidi, Dr. Yahya Tabesh, Amir Amidi, Ahmadreza Masrour, and Akvile Gustaite, and HKSTP leaders, Albert Wong & Pheona Kan, who are interested in continuing these conversations. It was awesome to meet great entrepreneurs and see old friends: Reza Moghtaderi Esfahani, Daniel Lo, Houman Homayoun, Wayne Chang, Golnaz (Naz) Moeini. #product #gotomarket #globallaunch #globalbusiness

Last week, I promised to answer your top questions about leadership in the age of AI. So, here goes! I’ll start with a foundational topic: What mindset shifts do leaders need to make during times of huge change? For me, it comes down to this — we need to go from being “map readers” to “explorers.” Map-readers rely on past routes and like knowing the destination. Explorers enjoy shifting terrain and thrive in not knowing the destination. They run experiments, stay close to the work and their teams, and earn trust by being present and being human.  They succeed because they are curious enough to learn. “Exploration is really the essence of the human spirit.” – Frank Borman (Apollo 8 astronaut) Minimize change → Ride the change Why it matters Change is not a phase. “Back to normal” isn't coming. Success is building resilience and helping teams thrive in turbulence. The mindset sets the tone: it has to be “let’s do this” versus “oh no, change”. What should leaders do Communicate with clarity relentlessly - what’s known, what’s unknown, and how you are making decisions. Make calls with incomplete information: run tests, adjust fast.    2. Certainty mindset → Scientist mindset Why it matters When so much is changing, doing what worked before won’t work. A scientist mindset means you have curiosity over certainty. You look for reasons you might be wrong, not just reasons you must be right and you surround yourself with people who challenge you. What should leaders do Set hypotheses and run experiments (more about this next week). Iterate, and learn as much from being wrong as from being right. Be a “learn-it-all,” not a “know-it-all.” 3. Manage from above → Get close to work Why it matters When you are exploring new paths, you need to stay close to the ground. You need to be a master of your craft Managing with decks and dashboards is not enough. What should leaders do Write prompts, embed within your team, get close to your team's processes. Triangulate with feedback from customers, partners and team members and don't rely on filtered reports. 4. Drive with control → Enable with context Why it matters The simple definition of context: it is what enables great work. Humans and AI both need it to deliver. It is the shared frame that makes the next action obvious and lets teams move with confidence and speed. What should leaders do Start with the “why” and “why now” behind strategies, pivots and decisions. Communicate it on repeat. Don’t dilute the message as it cascades down. Own it. 5. Me → We Why it matters No single leader can solve challenges alone, and being a lone explorer will lead to burnout. Choosing “we over me” puts team wins ahead of ego. And that’s how we win. What should leaders do Stay humble and recognize you may not have all the answers. Listen deeply across the business. Coach and help others grow. Let’s keep the conversation going in the comments, and I’ll share my second post in this series next week.

The most valuable leadership lesson I ever learned came from watching a science teacher, not a CEO. Last week, I observed a classroom where students were completely captivated for 90 straight minutes—no phones, no distractions, just pure engagement. The teacher wasn't using advanced technology or expensive materials. She was simply applying engagement principles that most corporate leaders completely miss. 𝗧𝗵𝗿𝗲𝗲 𝗹𝗲𝗮𝗱𝗲𝗿𝘀𝗵𝗶𝗽 𝗹𝗲𝘀𝘀𝗼𝗻𝘀 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝗰𝗹𝗮𝘀𝘀𝗿𝗼𝗼𝗺 𝘁𝗵𝗮𝘁 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗲𝗱 𝗺𝘆 𝗮𝗽𝗽𝗿𝗼𝗮𝗰𝗵: 𝗟𝗲𝘀𝘀𝗼𝗻 #𝟭: 𝗗𝗲𝗺𝗼𝗻𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻 𝗼𝘃𝗲𝗿 𝗱𝗲𝗰𝗹𝗮𝗿𝗮𝘁𝗶𝗼𝗻 Instead of telling students about static electricity, the teacher created a visual demonstration that made the concept impossible to ignore. The impact was immediate and undeniable. Application: Stop telling your team about strategic priorities—show them through concrete examples that make abstract concepts tangible. 𝗟𝗲𝘀𝘀𝗼𝗻 #𝟮: 𝗘𝗺𝗼𝘁𝗶𝗼𝗻 𝗯𝗲𝗳𝗼𝗿𝗲 𝗶𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 The teacher created wonder and curiosity first, then delivered the scientific explanation. By leading with emotion, she ensured students were primed to receive information. Application: Create emotional investment before data dumps. The quarterly numbers matter only when people care about why they matter. 𝗟𝗲𝘀𝘀𝗼𝗻 #𝟯: 𝗦𝗲𝗾𝘂𝗲𝗻𝗰𝗶𝗻𝗴 𝗰𝗿𝗲𝗮𝘁𝗲𝘀 𝘀𝘁𝗼𝗿𝘆𝘁𝗲𝗹𝗹𝗶𝗻𝗴 Each experiment built upon the previous one, creating a narrative arc that pulled everyone forward. No random collection of facts—a deliberate journey. Application: Structure your meetings, presentations, and communications as chapters in a compelling story, not isolated information packets. I immediately implemented these principles in my leadership approach: • Replaced our text-heavy quarterly updates with live demonstrations of our impact • Restructured team meetings to start with customer stories before metrics • Created a narrative arc for our strategic initiatives that builds momentum The results? Team engagement scores up 32%, meeting participation increased by 47%, and most importantly, our strategic priorities are now universally understood and embraced. The best leadership insights often come from unexpected sources. What non-business context has provided your most valuable leadership lesson? ✍️ Your insights can make a difference! ♻️ Share this post if it speaks to you, and follow me for more.

Asia is now the epicenter of global biopharmaceutical progress. Accounting for 43 percent of the global innovative biopharma pipeline, the region now surpasses both the US and Europe in scale. New research by my colleagues Anirudh Roy Popli, Fangning Zhang, and Jay Park with Franck Le Deu explores how Asia leapfrogged in biopharma from fast follower to frontrunner across discovery, development speed, and next-gen modalities. What stands out is how the region offers a portfolio of complementary capabilities rather than a single entry point: Japan for basic science and translation, China for scaled and cost-efficient clinical development, South Korea for advanced biologics and manufacturing, and Singapore for early-stage innovation and regional deployment. Organizations that can architect models drawing on these complementary strengths, beyond traditional partnerships, can drive faster innovation and more resilient systems. More here: mck.co/asiabioinnovation

“𝐁𝐲 𝐦𝐚𝐤𝐢𝐧𝐠 𝐭𝐡𝐢𝐧𝐠𝐬 𝐬𝐦𝐚𝐥𝐥𝐞𝐫, 𝐞𝐯𝐞𝐫𝐲𝐭𝐡𝐢𝐧𝐠 𝐠𝐞𝐭𝐬 𝐛𝐞𝐭𝐭𝐞𝐫 𝐚𝐭 𝐭𝐡𝐞 𝐬𝐚𝐦𝐞 𝐭𝐢𝐦𝐞," Gordon Moore told me in a 2014 interview. And because Moore’s Law turns 🎉61🎉 this week, let’s ponder: how is this an empirical law of economics? "The transistors get faster. The reliability goes up. The cost goes down." Scaling works wonders. If you can do it, it's still the best path to smaller, faster, better. And it's the density that makes it all work. Let's illustrate with the help of Douglas Adams. 🤓 📖 Imagine printing The Hitchhiker's Guide to the Galaxy. 227 pages at font size 14. Now shrink all text to font size 7. Same content. Half the pages. What just happened? ✅ You doubled the information density per page. ✅ You reduced the cost to print the book. ✅ You reduced the time to print the book. 🔬 That's traditional semiconductor scaling in a nutshell. When chipmakers shrink transistor dimensions, they don't just fit more transistors onto the same silicon area. The transistors switch faster because electrons travel shorter distances. They consume less power because lower capacitance means less charge to switch. And the cost per transistor drops because you're packing more function into the same area. One change, smaller features, improves performance, power, and cost simultaneously. That's what made Moore's Law an economic engine, not just a technical observation. 💡 Of course, the analogy has limits. Shrink that font too far and the ink bleeds, the letters blur, the printer can't resolve the detail. Chipmakers face the same wall: at some point (many points actually!), physics pushes back. Leakage currents rise. Variability increases. Going smaller requires relentless innovation across the entire chain. Chipmakers like Intel, TSMC and Micron Technology push process technology forward node by node, while equipment makers like ASML, KLA and Lam Research invent the tools that make each new node physically possible. That tension between the enormous payoff of scaling and the rising cost of continuing it is the central drama of the semiconductor industry today.

Most healthtech founders don’t get a “no” from investors. But that’s a BAD thing. Because they don’t get the money either. They get silence. Endless follow-ups. Polite interest… followed by nothing. Not because the idea is bad. But because the pitch lacks proof. And that’s the only currency that counts in healthtech. You’re often raising: BEFORE clinical trials BEFORE commercial scale BEFORE regulatory approvals Here’s how you can bridge the credibility gap and raise funding early: 1. Prove the problem is real (and costly) Use real-world stats from peer-reviewed studies or government data. The more specific and costly the problem, the more investable your solution feels. Example: “20% of diagnostic errors stem from image misinterpretation.” 2. Use early pilots to prove outcomes, not just usage A working prototype isn’t enough. Even pre-commercial, you can run: -Clinical pilots (10–50 patients) for health outcomes -Operational pilots to track time saved, error reduction, throughout gains 3. Validate from the front lines, not just the lab A quote from a nurse who uses your product 10x a day beats a generic testimonial from a CIO. Clinicians, procurement heads, and payers understand ROI and adoption - their support carries real weight. 4. Build a clear path to scale before the money arrives Investors want to see how you’ll grow from 1 to 100. Define your regulatory strategy, GTM plan, and business model early. Show the infrastructure and systems you’ll need to scale - don’t wait for the money to figure it out. 5. Know what “proof” looks like at your stage At pre-seed: expert interviews + MVP + LOI At seed: pilot results + early revenue + FDA submission At Series A: revenue growth + multi-site usage + clinical publications Reverse-engineer your path and build with intention. You don’t need millions to show traction - just the right proof. In healthtech, evidence is everything, where lives, compliance, and capital are on the line. Having been on both sides of the table, I can tell you: the pitch that wins isn’t the flashiest - it’s the one with undeniable proof. Not sure if your pitch deck makes the cut? For a limited time, I’m providing free pitch-deck reviews to Early stage healthtech entrepreneurs. To get access, DM me ‘REVIEW’. #investment #founders #startups #healthtech

🌍 Reporting from New York during the UN General Assembly — where conversations about innovation, technology, and humanitarian response give us real hope for the future. At the World Food Programme, we’re not just talking about innovation — we’re scaling it. Three examples: 🔹 Building Blocks: Our blockchain-based system, now used by 55 agencies in Ukraine, has saved $270M in the last three years while reducing unintended assistance overlaps. 🔹 SCOUT: An AI solution for supply chain optimization — already saving $2M in its very first pilot. 🔹 Food Security Forecasting: Using AI, we can now forecast food security up to 90 days ahead, helping us prepare and respond faster to crises. This is exactly what humanitarian innovation should do: make us more efficient, effective — ensuring that every donated dollar reaches the people who need it most. #Innovation #Humanitarian #AI #Blockchain #UNGA WFP Innovation Accelerator

How UoA Gets Ideas Off the Shelf and Into the World New Zealand’s new IP policy announced yesterday is a bold signal: researchers are being trusted to take a greater lead in commercialising their own inventions. It represents a significant shift in emphasis, one that has attracted attention for both its ambition and the complexity that will accompany its implementation. It is worth examining institutions that have already done the hard work of developing systems that genuinely help ideas transition from the lab to the market. The University of Auckland was singled out in the Government’s science system review for its IP approach. This article explains what sits behind that recognition. What stands out isn’t policy. It is the ecosystem around it: clear processes, attention to mātauranga Māori, practical support for researchers, and an evergreen inventors’ fund that lowers the barrier to taking the first commercial steps. UoA's experience shows that IP reform is not simply about who owns what. It is about whether we have the capability, culture and infrastructure to help good ideas grow. As New Zealand reshapes the rules around publicly funded research, these insights matter. They remind us that IP policy is only one part of the innovation equation. To realise the potential of the Government’s announcement, researchers, universities and the wider system will all need to build the confidence, skills and support structures that turn ownership into impact. The article is well worth a read. It captures what thoughtful, practical IP management looks like in a university setting and what New Zealand will need more of if the reform is to deliver on its promise. #ResearchCommercialisation #InnovationEcosystem #IntellectualProperty #TechTransfer #Universities 👉 https://lnkd.in/g_GkYTbc

Exciting Breakthrough in Recommendation Systems: Climber Framework Achieves 12.19% Performance Gain with Resource Efficiency I just came across a groundbreaking paper from NetEase Cloud Music researchers that addresses one of the biggest challenges in recommendation systems: scaling up models without prohibitive computational costs. The team introduces Climber, a resource-efficient recommendation framework that combines algorithmic innovation with engineering optimization through two key components: 1. ASTRO (Adaptive Scalable Transformer for RecOmmendation) - A novel model architecture that:   - Implements multi-scale sequence partitioning to reduce attention complexity   - Uses dynamic temperature modulation to adaptively adjust attention scores for multi-scenario and multi-behavior interactions   - Applies bit-wise gating fusion to integrate information across subsequences 2. TURBO (Two-stage Unified Ranking with Batched Output) - An acceleration framework featuring:   - Gradient-aware feature compression to optimize training data   - Memory-efficient Key-Value caching with block-wise parallelism   - A unified approach that bridges the gap between training and inference What's particularly impressive is that Climber achieved a 12.19% improvement in online metrics while requiring only moderate increases in computational resources. The framework has been successfully deployed at NetEase Cloud Music, serving tens of millions of users daily. The researchers conducted extensive experiments showing that balanced scaling—alternating between sequence length and model depth expansions—yields the best performance gains. They also found that different combinations of these factors under equivalent FLOPs can lead to significant variations in model performance. This work represents the first industrial-scale study of scaling laws in recommendation systems from a computational resource-constrained perspective. It demonstrates that coordinated algorithmic-engineering innovation, rather than brute-force scaling, can unlock sustainable performance growth under real-world constraints.

 Why Scaling Cell Therapy Demands More Than Science Alone? 👇 Cell therapy has emerged as a beacon of hope for patients battling cancers, genetic disorders, and degenerative diseases. Yet, as we stand on the cusp of a medical revolution, the challenge of scaling these therapies threatens to delay their life-changing potential. The recent news from Bluebird Bio—a pioneer in gene therapy—underscores this tension. While their FDA-approved treatment for sickle cell disease, Lyfgenia, marks a scientific triumph, its rollout has been hampered by manufacturing complexities, cost barriers, and logistical hurdles. For patients and families already enduring immense physical and emotional strain, these delays aren’t just frustrating—they’re heartbreaking. Why Scaling is So Hard Cell therapies are not pills rolling off an assembly line. Each dose is a living medicine, tailored to individual patients through painstaking processes. Scaling requires pristine manufacturing facilities, specialized talent, and rigorous quality control—all of which demand time and investment. Even Bluebird, with decades of expertise, faces bottlenecks in production and reimbursement challenges that limit patient access. The result? Therapies that could save lives remain out of reach for many who need them most. A Path Forward: Collaboration & Innovation The good news is, we’re not powerless. Here’s how we can bridge the gap: 1️⃣ Invest in Smarter Manufacturing: Automating cell cultivation and leveraging AI for quality assurance could slash costs and errors. 2️⃣ Streamline Regulations: Partnerships between companies and regulators can fast-track approvals without compromising safety. 3️⃣ Rethink Payment Models: Innovative financing, like installment-based pricing or outcomes-linked payments, can ease the burden on healthcare systems. 4️⃣ Empower Healthcare Networks: Training centers globally to deliver therapies ensures expertise isn’t siloed but shared. Bluebird’s journey teaches us that breakthroughs alone aren’t enough—we need systems that match the scale of human needs. Let’s channel our collective urgency into action: for the parent praying for a cure, the patient clinging to hope, and the communities waiting for equity in access. The road ahead is steep, but by uniting science, policy, and empathy, we can turn the promise of cell therapy into a reality for all. Together, let’s build a future where no one is left waiting. 💙 What steps inspire you most? Share your thoughts—this is a conversation we need to have. #celltherapy #biotech #invest #investor #innovation #science #market #cost Spencer Knight Benjamin McLeod