To evaluate whether the current boom in artificial intelligence will result in a disastrous collapse or a transformative revolution, historical technology bubbles can be analyzed through three interdependent mechanisms: the Capability Clock, the Profitability Clock, and the Adoption Clock. While today’s massive investments assume rapid, widespread monetization, the divergence among these three clocks reveals the true risks and long-term trajectory of AI.

The Three Clocks Framework

Every major technological paradigm shift is governed by three distinct timelines:

• The Capability Clock (The Airship Dilemma): A technology must continuously improve to survive. Airships failed because their underlying physics hit a ceiling while airplanes rapidly improved. AI, particularly large language models (LLMs), has seen rapid capability gains in structured domains like code and mathematics, but faces flattening scaling laws, data exhaustion, and persistent hallucinations that may require entirely new architectures to overcome.
• The Profitability Clock (The Concorde Dilemma): Technological superiority does not guarantee commercial viability. The Concorde was an engineering marvel, but its exorbitant operational costs and niche customer base doomed it, while the mass-market Boeing 747 triumphed. Today, many enterprises find generative AI too expensive, error-prone in high-stakes environments, and prone to creating costly technical debt. Furthermore, leading AI labs continue to operate at steep net losses, subsidizing inference costs that must eventually rise.
• The Adoption Clock (The Railroad Dilemma): Even capable, profitable technologies can trigger devastating bubbles if capital outpaces deployment. In the 1840s, British railroads bankrupted investors because tracks were laid decades before towns and industries matured to use them, even though the infrastructure eventually reshaped the global economy. AI exhibits rapid superficial adoption (e.g., individual chatbot usage), but deep integrated adoption requires restructuring business workflows, retraining workers, and overcoming public and regulatory pushback—a process known as the Productivity Paradox that often spans decades.

The Pressures of Impatient Capital

Big tech firms and investors are projected to pour hundreds of billions—surpassing $1 trillion annually—into AI infrastructure. Much of this funding is tied to impatient capital seeking near-term returns. Because the Adoption and Profitability Clocks move significantly slower than investor expectations, an economic correction or market pullback over the next few years is highly probable.

Beyond Railroads: AI as Electricity

Despite the likelihood of an investor correction, AI is unlikely to end up like the abandoned airship. Instead of a single-purpose invention, AI functions as a foundational platform technology, much like electricity. Electrification went through multiple booms, busts, and decades of trial and error before reshaping manufacturing, consumer goods, and communications. Similarly, current LLM limitations will not mark the end of AI; they are merely the first wave in a generational platform evolution. The ultimate value of AI will not come from simply automating jobs to cut costs, but from creating entirely new capabilities and collaborative problem-solving systems.

Mentoring question

In your own organization or industry, are you seeing ‘superficial adoption’ or true ‘integrated adoption’ of AI, and what workflow barriers must be addressed to ensure its usage is genuinely profitable?

Source: https://youtube.com/watch?v=KJxfSIvv920&is=1cw17ojPm0hUnV7U