The Milken Formula and the AI Revolution: Why Context Matters

Those who forget the past are bound to repeat it.

Welcome to the first weekend edition of the VSTRAT newsletter, an experiment focused on the economics of AI and the history of innovation. If you like them, or don’t, let us know at info@vstrat.ai.

I used to hang out with a friend, a psychologist who worked in group homes when she was younger. She told me Michael Milken used to show up unannounced, take the kids out to eat, pay for everything, and patiently talk to them. She was stunned when she learned about his past. I was stunned he had any unscheduled time at all — this was a man who, at the time, ran his day in six-minute increments.

It's hard to forget about Milken, creator of securitization and eventual convicted felon (pardoned by Trump). He cast a long shadow both in terms of history and in terms of the shadow banking system he created. The man who revolutionized finance through junk bonds also quietly spent time with vulnerable kids, managing his philanthropy with the same precision that made him billions and ultimately landed him in prison.

The Phoenix Formula

Still, like a phoenix, Milken's thoughts from back then keep rising from the ashes. Before coming to INSEAD, I worked on the US foreclosure crisis. One day, a colleague introduced me to what she called the Milken prosperity formula:

P = ΣFtᵢ * (ΣHCᵢ + ΣSCᵢ + ΣRAᵢ)

That is, prosperity equals the application of financial technology times the sum of human capital, social capital, and real assets.

During the foreclosure crisis, this formula revealed itself with devastating clarity. The computers found patterns in mortgage data that turned out to be devastatingly accurate. These algorithms leveraged human capital (countless people whose lives were affected), social capital (enormous press coverage), and real assets (people's homes themselves). The result was a system that could undo rot the moral and structural wellbeing of the global financial system.

Of course, Milken originally used this framework to enable junk bonds: high-yield debt instruments that democratized corporate finance while creating systemic risk. The formula worked exactly as designed: financial technology amplified by human expertise, social networks, and underlying assets created unprecedented prosperity for some and catastrophic losses for others.

The AI Amplification Effect

Today, I'm more focused on the formula's application to generative AI, our latest form of supercharged technology. Many evaluate generative AI in isolation—measuring its capabilities through benchmarks, comparing model parameters, or marveling at individual outputs. But this misses the crucial context that makes AI truly powerful.

The real transformation happens when AI becomes the financial technology component in Milken's formula, amplified by the human capital of skilled practitioners, the social capital of trust networks and institutional relationships, and the real assets of data, infrastructure, and established businesses.

Consider how generative AI is reshaping entire industries. It's not just that ChatGPT can write code or that Midjourney can create images. It's that these capabilities, when combined with human expertise (programmers who know what to build), social networks (teams that can collaborate and clients who trust the output), and real assets (existing codebases, brand relationships, distribution channels), create value that far exceeds the sum of their parts.

A solo developer using AI to build software isn't just using a better text editor; they're leveraging decades of accumulated programming knowledge (human capital), tapping into online communities and professional networks (social capital), and building on existing platforms and libraries (real assets). The AI amplifies all of these existing resources simultaneously.

Recognizing the Pattern

Having lived through both the dot-com bubble and the housing crisis, I can assure anybody asking whether we're in an AI bubble: yes, we absolutely are. The hype cycles, the inflated valuations, the rush to slap "AI" onto every business model — it's all eerily familiar.

But here's what I learned from those previous bubbles: the underlying hype wasn't wrong, just the timeline and the initial applications. The internet really did transform everything, just not in the ways or on the schedule that 1999 investors predicted. Similarly, securitization and sophisticated mortgage instruments weren't inherently evil — they became dangerous when misapplied and poorly regulated.

The dot-com bubble burst only to be replaced by something more sustainable. Amazon survived and thrived, Google emerged from the ashes, and the internet became the foundational infrastructure we can't imagine living without. The same pattern will likely play out with AI: the bubble will pop, many companies will fail, but the underlying transformation will prove even more profound than the early evangelists predicted.

The Double-Edged Formula

The Milken formula works both ways: it can be used for tremendous good and it can enable significant harm. Like a rule of physics, it simply describes how prosperity gets created when financial technology meets human capability, social connection, and real resources. The formula itself is morally neutral: its impact depends entirely on how we choose to apply it.

When Milken used it to create junk bonds, the result was both the democratization of corporate finance and the creation of systemic risk. When applied during the foreclosure crisis and afterward, it initially enabled both more accurate risk assessment and the exploitation of vulnerable borrowers and, eventually, the partial cleanup thereof. Now, as it shapes the AI revolution, it's simultaneously enabling breakthrough medical research and sophisticated disinformation campaigns.

Our Moment of Choice

Understanding this formula gives us a framework for thinking about AI's future impact. The technology itself—impressive as it is—will succeed or fail based on how effectively it integrates with human expertise, social structures, and existing assets.

The companies that will thrive in the post-bubble world won't be those with the most impressive AI models, but those that most thoughtfully combine AI capabilities with deep human knowledge, strong relationship networks, and valuable real-world assets. They'll be the ones that recognize AI as a powerful amplifier rather than a replacement for these fundamental sources of value.

Our hope, and our responsibility, is ensuring this amplification gets directed toward prosperity that benefits everyone, not just those who control the technology. Milken's formula reminds us that technological power without wisdom, social responsibility, and ethical constraints doesn't create lasting prosperity … it creates bubbles that eventually burst.

The question isn't whether AI will transform our world: it already has. The question is whether we'll learn from history and apply this transformation more wisely than we have in the past. Time will tell if we're up to the challenge.