To Be, or Not to Be (Disruptive)

In a recent talk at a major industry event in Bangalore, I spoke about the WHY of adopting (or foregoing) the latest trends in AI and IA, as well as the WHAT. The HOW is left to the performers' discretion and imagination.

Deciding when, where and how much to invest in major technological trends is not an easy task. On the one hand, no organization wants to be left out at the station waving goodbye to the fast trains of the latest industry trends.

On the other hand, the fear of missing out (and anxiety about not being able to run fast enough to catchup later) often results in rushing onto the wrong train headed for the cliffs. Often, Sunk Cost Fallacy results in staying the course while clearly heading for disaster.

All models are wrong, but some are sometimes useful (LLMs, or otherwise)

To navigate such "innovation dilemma" (or more colloquially, "getting on the right train"), the well-established discipline of Decision Analysis often provides useful models.

Ironically, the latest decision analytical frameworks that could be used to decide whether or not the latest "AI Spring" is going to be an eternal one, are themselves pinned on the latest AI/ML models and methodologies, creating a "generative chicken and egg" problem (susceptible to hallucinations, drift, bias, and model poisoning).

In mathematical terms, the innovation dilemma could be viewed as navigating Chaotic Attractor Regions (specific regions within the state space of decision analytical system where trajectories tend to converge over time, but with chaotic twists, some of which result in death spirals).

Back to Basics

In order to understand when, where, and how much to invest in the latest innovation trend, the organization must first recognize which innovation track it is on. Well-established incumbents can rarely get away with "good enough," and often need to invest significant amounts to meet their growth objectives and keep the premium customers from churning.

Disruptors, on the other hand, need to deliver "good enough" offerings that meet the 3 Cs of Cost Effectiveness, Core Features, and Convenience (also referred to as "Cheap, Crappy, and Convenient"). In the hands of competent sales enablement and clever marketing, duct-taped solutions meeting the 3 Cs can disrupt empires, while disruptive innovators that spend inordinate amounts on over-polishing their goods will likely miss the market, and end up bankrupt.

The Alphabet Soup

Once the organization is confident in its innovation identity, the next task is to balance the decision analytical stool on the three legs of:

• TCO (Total Cost of Ownership): What are the short-term, mid-term, and long-term costs associated with pursuing one option, while foregoing many others? Here, costs span beyond the myopic CapEx and OpEx calculations, and require a more mature analysis into the nth order effects of the decision (including goodwill and reputation, opportunity costs, etc.).
• TTV (Time to Value): How soon can the chosen option prove its value to the stakeholders (e.g., customers, partners, investors, internal teams, etc.)?
• ROI (Return on Investment): How much return (e.g., growth and/or savings) can be anticipated, and how soon can it be realized?

If at any point in time one of the legs of the stool begins to wobble, it is time to reformulate the strategic and tactical execution plans to avoid burning more resources in the money (and time) pit (e.g., jump off the moving train as safely as possible). With some luck and craft, the lessons-learned can forge a better stool in the next round.

The Decision Factory

Entities that master the art and science of producing this "decision analysis stool" for one field or industry, can easily reproduce it at scale for any other field or industry, thus creating a "decision factory."

The factory requires a few things:

• Quality Data: The sparsity and symmetry inherent in many data models makes them more susceptible to the phenomenon of "model poisoning" where <1% of 'bad data' can poison an entire model otherwise constructed on 'good data'. By way of analogy, the amalgam of 99% pure gold and 1% highly radioactive plutonium results in a completely toxic alloy unsuitable for daily wear.
• Data Intelligence Platform: A modern, data-centric ecosystem that checks all the desired functional and non-functional requirements for Data Democratization and AI Democratization.
• Domain Experts: The platform is akin to a modern race car, clean and quality data is the fuel, and the expert is the race car driver (human or machine).

Journey Into Another Automation Revolution

While no one can predict the future, we can certainly pull a thread from the many historical spirals to anticipate and proactively plan for one or more of the many future possibilities. Just as today's Intelligent Automation and Artificial Intelligence products are the results of decades of failure and success spirals, similar motions took place centuries ago, eventually leading to the ubiquitous presence of modern automobiles.

The Evolution of Automobiles: From Steam to the Model T

The history of automobiles is a fascinating journey that spans centuries, marked by innovative engineering and evolving technologies. From the earliest steam-powered contraptions to the ubiquitous internal combustion engine (ICE) vehicles we see today, the evolution of automobiles has transformed transportation and revolutionized society.

The Early Days: Steam Power

The concept of a self-propelled vehicle dates back to the late 17th century, but it wasn't until the 18th century that the first practical steam-powered cars emerged. In 1769, French inventor Nicolas-Joseph Cugnot built a steam-powered tricycle designed to haul artillery, but it proved to be cumbersome and impractical for everyday use.

Throughout the 19th century, inventors continued to refine steam-powered vehicles, with varying degrees of success. However, steam cars faced significant challenges, including long start-up times, limited range, and the need for frequent water replenishment. Despite these limitations, steam-powered cars played a crucial role in demonstrating the potential of self-propelled vehicles and paving the way for further innovation.

The Rise of Internal Combustion Engines

In the late 19th century, a new type of engine emerged that would revolutionize the automobile industry: The internal combustion engine (ICE). Invented by German engineers such as Nikolaus Otto and Gottlieb Daimler, ICEs offered greater efficiency, faster start-up times, and a more compact design compared to steam engines.

Karl Benz, another German engineer, is often credited with building the first practical automobile powered by an internal combustion engine in 1885. Benz's three-wheeled vehicle, the Benz Patent-Motorwagen, marked a significant milestone in automotive history and set the stage for the mass production of automobiles.

The Ford Model T and the Standardization of ICE

The early 20th century saw a rapid proliferation of automobile manufacturers and models, but it was Henry Ford's Model T that truly transformed the industry. Introduced in 1908, the Model T was designed to be affordable, reliable, and easy to maintain, making car ownership accessible to a wider audience. Ford's innovative assembly line production methods further reduced costs and increased efficiency, making the Model T the first truly mass-produced automobile.

The Model T's success solidified the dominance of the internal combustion engine as the standard power source for automobiles. Its simple design, ease of use, and affordability made it the most popular car of its time, with over 15 million units produced. The Model T's impact on society was immense, as it transformed transportation, spurred economic growth, and changed the way people lived and worked.

It also put the final nail in the coffins of possibly thousands of entrepreneurial ventures that bet on the wrong horse power, and it took over a century for the large-scale production of electrical vehicles to be resuscitated back to life (steam powered automobiles may not be so lucky).

Knock Knock ...

The success of ICE was despite its one very clear problem: The deafening ping echoing from the early engines, known as the ominous sound of "engine knocking." In the early days of motoring, engine knocking was a pervasive problem. This unwelcome phenomenon, where fuel combusts unevenly in the engine's cylinders, led to decreased performance, increased polluted exhaust, and even catastrophic engine damage and explosions.

A solution emerged in the form of tetraethyl lead (TEL), a compound added to gasoline to raise its octane rating and suppress knocking. American chemical engineer Thomas Midgley Jr., who was working for the U.S. corporation General Motors, was the first to discover its effectiveness as an antiknock agent in 1921, after spending several years attempting to find an additive that was both highly effective and inexpensive. Leaded gasoline quickly became the norm, touted for its ability to smooth out engine performance and boost power.

However, the victory over engine knocking came at a steep cost. Lead, as we now know, is a potent neurotoxin with devastating effects on human health, especially in children. The widespread use of leaded gasoline led to a public health crisis, prompting its gradual phase-out beginning in the 1970s. The last country in the world to officially abandon leaded gasoline was Algeria in 2021 (exactly a 100 years after TEL was discovered), but unofficial usage may remain throughout the globe.

Conclusion

The turbulent and remarkable journey of technological innovation and societal transformation of AI/IA in the last few decades is in many ways reminiscent of the evolution of automobiles from steam-powered contraptions to the iconic Ford Model T and the latest generation of vehicles, which spanned several centuries.

While steam power laid the foundation for self-propelled vehicles, the internal combustion engine's superior efficiency and practicality ultimately prevailed, shaping the automobile industry as we know it today. The Ford Model T, with its revolutionary production methods and affordability, democratized car ownership and ushered in a new era of personal mobility.

Today, the data intelligence platforms democratizing artificial intelligence and intelligent automation equally have to the potential to transform industries and ultimately, societies. And hopefully, the lessons-learned from the automotive industry can be applied to avoid spewing out a century's worth of neurotoxins into our world.