Adam Judelson's Linkedin Analytics

To thrive in the age of generative AI, we need to think in three different ways. From working alongside some of the most ambitious product teams in emerging tech. 1. Think for yourself Start from first principles. Use your own judgment, your lived experience, your instincts. Ask: What do I actually believe here? What’s really going on? 2. Ask AI to think for you This isn’t about collaboration. It’s about input and output. Take your specific question. Try it with different models. See what AI returns. Evaluate the responses on their own terms. 3. Think with the AI Engage in a full conversation. You ask a question, share context, get a response, then respond again. Refine your thinking as the exchange evolves. Some of the work is yours, some of it is AI’s. “If we do all three — we get novel insights, better execution, better plans.” But skip one, and something gets missed. So remember: → Think for yourself → Ask AI to think for you → Think with the AI

Big companies don’t die because they’re stupid. They die protecting the thing that made them big. You can’t innovate inside the same division responsible for the cash cow. Too much to lose. Too much risk. So nothing new or risky gets built. But the problem is: – The cash cow paid for 100,000 employees – It allowed for market-leading compensation and stock – It has to be protected — or so they think So they protect the moat at all costs. And the cost is: They get outcompeted by startups doing something new. Google has already lost to Perplexity in intelligent Internet search. The only reason their market cap hasn’t tanked is that most people haven’t tried Perplexity yet. Perplexity is huge with hundreds of millions of searches, but only has about 0.25% of the search volume of Google right now. So how do big companies survive? ✅ They protect the cash cow ✅ But they also build autonomous & independent innovation teams ✅ And when the moat is under threat, they don’t send in the MBAs They respond with better technology. New diagnoses of old problems. Disruptive thinking. Because the part of the business trained to protect the past. Will always be culturally incapable of inventing the future.

80% of product value comes from 20% of features. But how do you find the right 20%? You don’t. Not at first. “If we're talking about day one — how do I know which features will matter? You don't.” You take your best bets. You do product discovery. You release features experimentally. You make users choose between options. You watch what they actually do. “The real question is: Are you paying attention?” Because once you’ve shipped, the signal is usually obvious. → People vote with their feet. → Metrics show you where the value lives. → That 20% starts lighting up. There’s a harder edge case too: “What about the 10% of the product that only 10% of users touch… but it drives a huge chunk of revenue?” You won’t find that in your daily active count. But it still matters — a lot. And if you’re not paying attention? You end up like the Winchester Mystery House. (The heiress to the Winchester rifle fortune believed she was cursed by the ghosts of those killed by her family’s guns. So she kept building — adding 160 rooms, staircases to nowhere, doors that open into walls — believing it would keep the spirits at bay.) “You just keep building and building — but none of it is really additive.” That’s the real risk: Not that you pick the wrong features — But you never stop to see which ones worked.

The reason you’re not retaining customers isn’t loyalty. It’s that your product never became a habit. Retention doesn’t start with branding or NPS. It starts with how (and how often) people actually use what you’ve built. One simple test is: Would someone use this every single day? Maybe even multiple times per day? Most people assume that’s just for consumer apps. It’s not. “If I’m prioritizing features, and one of them gives the customer a reason to log in every day — I always push for that one.” That’s the mindset. So what makes that kind of depth possible? → Ask: What’s immediately adjacent to this? Superhuman does email. Calendars are adjacent. So they added a feature that previews your calendar inside an invite. A small thing — but a deep unlock. → Find the natural path. At Palantir, early on it was maps adjacent to graphs of connections. In synthetic biology, it’s microscope integrations. Not because they’re flashy. But because they’re the next logical step for someone already in the product. → Earn the right to ask deeper questions. You don’t get to the good stuff in the first user interview. “Once they show you they’re using the product, you earn permission to ask: What’s your next question?” That’s where the unlocks are.

2022: ChatGPT has no app, just a web page and a basic language model. 2025: 64.3M app downloads in a single month. A 276% YoY jump. What happened? It didn’t go viral because of enterprise integrations. It didn’t grow because of evaluation scores. It exploded because of a vision — and a product that kept getting better. I was coaching a founder recently who asked: “How do we get people to care about what we’re building?” Here’s what I told them: Start by giving them something they care about. Something your users actually feel drawn to — delight, curiosity, utility, or just fun. The teams behind breakout products ship early and learn fast. They watch what lights people up — and build from there. Or they start with a massive vision that feels unachievable and yet somehow defeat the technical obstacles to deliver it, like OpenAI. → From a text box on a web page to 64 million monthly downloads. → From a clever demo to a global habit. From curiosity to indispensability. That’s ChatGPT.

When should a product pivot vs. persevere? Here are 3 signals to watch for — and a story from a founder who needed to hear it. Most teams wait too long to ask this question. They keep building. Keep tweaking. All without truly pressure-testing the vision. Here’s what to watch: 1. People hear the price and walk away. If your sales conversations die as soon as people hear the cost — and this happens again and again — the market is rejecting the product value, not the packaging. That’s a pivot moment. 2. You’re getting bites, just not enough of them. If the vision resonates but usage or activation is low, the foundation is likely solid. The ramp just needs to be easier. That’s a persevere moment. 3. You’re being ignored by the real buyer. If your product is for the CIO and you’ve only spoken to their reports? You haven’t even gotten real feedback yet. That’s a visibility gap. You need to get access. Persevere. – If you’re early-stage and stuck, I’m happy to talk about your product.

The market sees the shift before you do. And that’s how big companies die. They’re not stupid. It’s not because they don’t have great people. They’re just too busy protecting their moat to notice it’s crumbling. Kodak invented the digital camera. Xerox invented the GUI. Blockbuster had the chance to buy Netflix. Nokia built the first smartphone. They all saw the future and dominated for a sustained period. But eventually… They just couldn’t let go of the past. And while they hesitated, the market moved on. What happens inside the building to let this kind of thing happen? – MBAs run around acquiring or suing competitors. – Execs double down on the thing that worked ten years ago. – And internally, people are still arguing about the homepage layout, brand guidelines, and font colors. What happens outside the building? Everyone’s already using the better thing. Here’s the hard truth: The people best equipped to protect the current cash cow Are the least equipped to build the next world-changing product.

Have you seen Apple’s new Siri function? It’s a masterclass in what happens when you stop copying — and start asking: What’s the most intuitive, elegant way for this to exist in the world? Apple just reminded us how first principles thinking shows up in design. The new Siri interaction doesn’t scream innovation: it’s a subtle blue wave at the edge of the screen when you activate it. But that’s the point. It’s rethinking the interaction itself, not just the aesthetics. We’ve come to expect this kind of thinking from Apple. But let’s be honest — it’s been a while since they showed it. This move is Apple back to its best. Not chasing competitors. Not iterating on trends. Quietly redefining how tech should feel. It’s in moments like this, where design choices feel obvious, but only after someone brave enough rethinks them from scratch, that you see that real product magic isn’t in features… It’s in design.

Zuckerberg said every product team at Meta is now expected to experiment with AI. That sounds like a tech shift. But it’s something deeper: a quiet redefinition of the roles themselves. PMs are getting pulled into model design conversations. Engineers are learning how to debug behavior, not just code. We’re seeing this everywhere — in teams building agentic systems, multimodal tools, and evaluation-heavy workflows. What used to be edge-case thinking is now the day-to-day. If you’re building AI-native products, this is your edge: Hire people who thrive in the grey. Who ask sharper questions. Who treat clarity as a skill, not a given. Who see no boundaries in their job. Shipping features gets you started. But shaping systems that learn and adapt and owning the outcomes? That’s what separates you.

Jennifer Pahlka, author of Recoding America, rightly celebrated in a recent post what San Francisco City Attorney David Chiu is doing — partnering with Stanford AI experts to untangle the municipal code. Because San Francisco is buried under code (the other kind!), it doesn’t even use. In the Politico article Jennifer shared, one example really stuck with me. City staff are still filing compliance reports for hardware that hasn’t existed in years. That’s more than just inefficiency: that’s a systems failure. And while it might sound futuristic, cities already have a name for this process: recodification. The job of going line by line through outdated or overlapping laws and making sense of them. But most cities don’t have the capacity to do it at scale. That’s where AI steps in. At Common Sense, one of the agents we’re building is the world’s first clean API for city code — surfacing direct citations from municipal code that actually matches the intent of what you’re looking for. So if you want 10 semantically similar examples of the same legal clause, you can get them — in seconds. - - - To Jennifer, David, Christine Tsang, Dan Ho, and the Stanford team: We’re watching closely and we’d love to contribute.

The Los Angeles Palisade Fire exposed the deepest cracks in our city’s emergency infrastructure. The current system was never designed to withstand a crisis of this magnitude, which is why, for many victims, the real suffering began after the fire, during a painfully slow recovery. Determined to understand what’s taking so long, I spent a day at one of the local recovery centers. Just as a neighbor, as someone who could just as easily have been sitting on the other side of the table. After listening to their stories and hearing how much it had affected them, I had to ask some important questions: Why are people still waiting to rebuild? What’s standing in the way? What I found wasn’t indifference or bureaucracy for its own sake. Everyone was working and cared deeply, but they were working alone. - The fire department was mapping working hydrant locations. - Environmental teams were testing soil for toxins and organizing removal. - Federal agencies were collecting signatures and forms. Each step was necessary, but no one had a unified vision, so progress stalled, one small decision at a time. Each with its own legal, ethical, and human implications. At the core of the problem was ineffective alignment. No one had designed the system to hold people when the ground fell out from beneath them. And it made me even clear on our mission at Common Sense: building systems that respond as if people’s lives are at stake. Because after a disaster, no one should have to navigate endless red tape just to rebuild their life. They’ve already lost enough.

PODCAST EXCLUSIVE: What makes an AI PM actually valuable — and how close are we to vibe-coded product teams? — with Christian Marek, VP of Product at Productboard In this episode, I sat down with Christian to unpack what AI is really changing in product management… and what’s just noise. We talked about how GenAI is reshaping discovery, what junior PMs are missing, and why Christian’s team is already seeing PMs shipping PRs with help from AI coding tools. Christian’s background is rare: He’s built at Meta, boomeranged back to DocuSign to launch their first AI feature, and now leads product at one of the most trusted platforms for product teams. This conversation cuts through the hype — and gets real about what’s working, what’s still broken, and what the next 6 months of AI-native product work might look like. Some takeaways from our conversation: – Why discovery is the most valuable (and overlooked) place to apply GenAI right now – What separates an “AI PM” from a spec-writing PM — and how to level up – Why the best AI product builders are both deeply technical and relentlessly curious If you’re hiring PMs, building with AI, or trying to rethink how your product teams work — this one’s for you. Listen here >> Substack: https://lnkd.in/gRVy7B8B YouTube: https://lnkd.in/g-cm-CRb Spotify: https://lnkd.in/gEDhfjb5 Apple Podcasts: https://lnkd.in/gYizKQWi

Did you know that your voicemails aren’t encrypted? Do you understand how your information is used to route a cellphone call in America, or overseas? John Doyle, founder of Cape (and former national security lead at Palantir), is building a mobile network that does it differently. Most carriers hand off detailed user data constantly to other companies, and worse, to our adversaries Cape limits what gets shared. Minimal data. No unnecessary exposure. Privacy baked into the infrastructure itself. Because in a world where everything is connected, the system you trust matters more than ever.

“Dad, can you show me how to follow your YouTube channel so you have more subscribers?” That’s what my then-seven-year-old daughter asked me. Months after launching my podcast, our audience was growing on Substack, but it was smaller than a dinner table on YouTube. We had exactly 3 YouTube subscribers. And my daughter wanted to be the 4th. It was both humbling and grounding. Fast forward to today - and I’m proud to say we’ve grown a lot. - 20,000+ subscribers. - Latest full episode with over 67,000 viewers. - Lots of tech CEO’s pitching to feature. So what changed? 1. We reframed the audience In the beginning, we were speaking primarily to product builders but as we evolved, we started designing conversations for anyone curious about how the digital emerging technology world works and where it’s heading. 2. We asked better questions We moved beyond industry shorthand and leaned into questions shaped by lived experience, reflection, and relevance. The kind of questions people carry with them after the episode ends. 3. We paired depth with clarity We didn’t dilute our ideas but delivered with clarity so that anyone listening could follow along, think deeply, and walk away sharper than they came. We didn’t attempt to go viral or chase any trends. We simply focused on consistently improving our content, aiming to make each episode useful, thoughtful, and genuinely worth our audience’s time. - - - I want to say thank you to all the incredible guests over the last 12 months- Matthew Steckman, Karan Talati, Jacob DeWitte, Andrew Herr, Ari Rosner, Nate Hamet 🛰, Audrey Howell, Tim Raftis, and John Doyle.