CDOQ

A few weeks ago, I finished reading Elon Musk by Walter Isaacson . The book provides a fascinating look into Musk’s life and how he has been able to make industry-changing contributions in business and technology. It also raises an uncomfortable question: can you get rockets to orbit and accelerate the transition to electric vehicles without accepting some of the chaos that comes with Musk? Isaacson captures this tension well in the final paragraph: "Could you get the rockets to orbit or the transition to electric vehicles without accepting all aspects of him, hinged and unhinged? Sometimes great innovators are risk-seeking man-children who resist potty training. They can be reckless, cringeworthy, sometimes even toxic. They can also be crazy. Crazy enough to think they can change the world."  One of the most useful parts of the book is Musk’s “algorithm” for building deep-tech hardware products:  "1.  Question every requirement.  Each should come with name of the pe...

Most product managers don’t fail because they lack ideas; they fail because they don’t make the right trade-offs. Over the course of my career, I’ve led product organizations building everything from 0→1 products to businesses generating over $600M in revenue spanning hardware, software, AI, and crypto. What I’ve found is that while the domains evolve, the craft of product management remains fundamentally the same. In the fast-paced world of product development, the job of a product manager is not to optimize for a single objective but to balance competing forces. I’ve found that good product managers consistently operate across four objectives:      A. User Experience      B. Profit (Unit Economics)       C. Competitive Advantage      D. Technology Advancement Read more about the four objectives here . The most common mistake product managers make is optimizing for one objective at the expense of the others. These objecti...

The fundamental flaw in the idea that AGI (Artificial General Intelligence) can be more intelligent than humans is the assumption that language itself represents intelligence.

My agent often answers better than I do—not because it’s more intelligent, but because it never forgets and can hold more context at once.

Technology evolves more through human belief than through natural laws.

Business school teaches you the theory of business — the frameworks, the financial models, the case studies. But real understanding comes from seeing how ideas collide with the messy, unpredictable world of people, markets, and technology. Over the years, I’ve found a handful of books that open your mind beyond the classroom. They explore the mental models, systems thinking, and creative problem-solving that shape great companies and great decision-makers. Here are eleven that I think business students will find useful:  Charlie Munger 1. Poor Charlie's Almanac   - Timeless lessons on decision-making and mental models from Charlie Munger. 2. Making things work   - Insight into the systems-thinking approach that underlies modern technology and organizations. 3. The Snowball  - A deep dive into Warren Buffett’s philosophy and long-term business perspective. 4. Fooled by randomness  - A brilliant exploration of luck, probability, and human bias...

Imagine teaching a lion to speak perfect English. Every word pronounced flawlessly, grammar impeccable. Yet, a conversation with it would still be impossible. Its frames of reference are so alien to ours that true understanding cannot occur—just as Ludwig Wittgenstein noted in Philosophical Investigations : ‘If a lion could speak, we could not understand him.’ Now imagine a counterpart trained on the vast body of human knowledge: GenAI. Here, conversation is not only possible but also delightful, because its frame of reference aligns with ours. GenAI can reason, explain, and answer across countless domains. On the surface, it seems like a mind without limits. Conversations But there’s a catch. Unlike a lion, GenAI can confidently deliver falsehoods. Without deep understanding of the topic in discussion, you may never notice. Knowledge alone isn’t enough. Context, experience, and critical thinking remain irreplaceable—even when conversing with a machine that seems to know everything.

If you’re not consistently making things better, they don’t stay the same—they get worse. 

Driving down Hwy 101 in Silicon Valley yesterday, I spotted this billboard (See image). It reminded me that while the Bay Area overflows with technical brilliance, tech companies often stumble when it comes to marketing and communication. The ad boldly declares: “DEPLOY BRAVELY.” Deploy Bravely? In software, deployment is the last place you want bravery. A good deployment should be boring—built on testing, quality assurance, monitoring, and the ability to roll back instantly if things go sideways. Sonos once deployed bravely and left some of its speakers effectively unusable.  The bravery resulted in the company’s market cap declining by half. I have no idea what Prisma Airs is—and the ad doesn’t make me interested in finding out more. Can you tell what it’s about?

I met a software engineer yesterday at a conference — let’s call him Ted (not his real name). He has worked at well-known Silicon Valley companies like Square, Google, and Yelp. Ted told me that culture varies widely across companies. I asked, “But how does that affect your job? You’re a software engineer — you just have to code to the requirements.” He smiled and said, “Culture defines how good the requirements are, how good my code is, and how happy the customer will be with the application.” Then Ted shared a practical test for culture: measure how much time you spend convincing others. If most of your energy goes into persuasion instead of building, the culture is working against you. In a healthy culture, things just flow — and you spend your day building, not convincing. In other words, in good cultures, it takes less energy to get things done.

Both Charlie Munger and Warren Buffett have an extraordinary gift for simplifying the complicated. They taught me that the most powerful ideas are often the clearest. I have since carried that lesson into my own work. At Moki, a company that provides software for remotely managing hardware in retail environments, we faced a common challenge: overcoming CIOs' fear of losing control. Instead of delving into the technical depth and cybersecurity of our product, we focused on a single, compelling promise: " Total Control of their devices." This simple framing cut through the complexity and directly addressed their primary concern. It worked.   An often-cited (but unverified) framework, attributed to Einstein, outlines five levels of intellect: 1. Smart: Possesses vast knowledge (e.g., a McKinsey consultant; whatever you think of them, they do know a lot). 2. Intelligent: Applies existing knowledge effectively (e.g., Michael Dell, who adapted the direct-to-consumer model for...

In just two years, ChatGPT knows me better than Google and Meta do after two decades.

This morning I finished reading Apple in China , a fascinating book about the interdependence between Apple and Chinese manufacturing. The book makes clear that the company might never have reached a $3T market cap without the contributions of Foxconn and support from the Chinese government. In turn, China might not have developed into a sophisticated manufacturing powerhouse without Apple’s training, investment, and partnership. Having worked with OEMs (Original Equipment Manufacturers) and ODMs (Original Design Manufacturers) in China, I found the book cast light on aspects of manufacturing I hadn’t fully appreciated. Apple Pyramid (Image credit: ChatGPT) Iconic products like the iPhone come to life at Apple through the push and pull of four teams. At the top, the Industrial Design (ID) team defines the look and feel. Next, the Product Design (PD) team ensures the product works as intended. The Manufacturing Design (MD) team then decides which components and assembly partners will br...

Companies are experimenting with AI because they don’t want to be left behind. But little thought has gone into which AI, or which implementation, will actually strengthen their competitive position. You can use AI—but if your competitor is using better AI, you may still be left behind.

The blind once led the blind. Now, the hallucinated lead the clear-eyed. 

Fears that GenAI (Generative AI like ChatGPT, Gemini, etc.) will make us dumber are greatly exaggerated. Did calculators make us dumber? Did computers? Sure, every new tool shifts what we practice, but history suggests we adapt, and often end up smarter in different ways. Brain damage? (image credit: Gemini) A commonly cited worry is that ChatGPT takes away independent thinking. Who has that anyway 😀? It’s true that you can lose elements of critical thinking if you hand over all the mental work to GenAI. Just as using a calculator without understanding math fundamentals can be a problem, using GenAI without grasping the basics of writing or coding could lead to skill loss. But for most adults, that’s not what’s happening. Those who outsource all their writing or coding to ChatGPT usually didn’t enjoy those tasks in the first place. Generally, repetition doesn’t magically make us love a task we dislike, nor does it guarantee we get much better at it. I’ve been cleaning for decades,...

We have lost the fragrance in roses. Scent did not scale.  Quantity over quality (image credit: Google Gemini)

Under U.S. law, a corporation is considered a legal person, but it has no inherent moral obligations. So why do we expect the artificial intelligence (AI) products created by corporations to behave morally?

A major societal shift in the past few decades is that most primary research in artificial intelligence (AI) is now happening within corporations, rather than in academia or through government-funded programs. Details here.  

Earlier this week, at a roundtable hosted by Chargebee , a group of GenAI (AI) founders compared notes on a shared learning: building and selling AI products to businesses is not just harder than selling SaaS — it is "fundamentally different'. Many entered the market expecting to ride the SaaS (Software as a Service) playbook to success. They are now discovering it’s written in a different language. Here are five reasons why: 1. The Vanishing Customer Profile In SaaS, identifying the Ideal Customer Profile (ICP) is often straightforward: a clear job to be done, a known budget holder, and measurable ROI. In GenAI, that clarity dissolves. Everyone wants to “experiment with AI.” Few want to commit. Interest is high, intent is elusive, and trial usage often masks the absence of a real buyer. 2. The False Promise of PLG Product-Led Growth — the darling of modern SaaS — is stumbling in the AI world. While PLG drives traffic and trials, it fails to convert at scale. AI products often...