The Early Days of AI: Building Moats for the New Era

The Early Days of AI: Building Moats for the New Era

Introduction:
The field of artificial intelligence has undergone a significant transformation in recent years. Rather than viewing LLMs, Transformers, and diffusion models as mere extensions of previous AI technologies, it is crucial to recognize them as the dawn of a new era. This discontinuity from the past has brought about a wave of new capabilities and products, marking the beginning of a technological revolution. In this article, we will explore the early days of AI, the different waves of innovation, and the importance of building moats to protect businesses in this rapidly evolving landscape.

The Waves of AI Innovation:
In the early days of AI, the value was largely concentrated among incumbents rather than startups. The limited capabilities of previous AI technologies prevented new market openings. However, with the advent of advanced AI models like ChatGPT, we witnessed the mainstream acceptance of AI's potential. Large enterprise planning cycles, coupled with the time required for prototyping and building, indicate that we are still far from reaching peak AI usage or hype. As we navigate through these early days, it is essential to consider at least four waves of AI:

Wave 1: GenAI native companies
ChatGPT, Midjourney, Character.AI, Stable Diffusion, Github Copilot, and other early launches have gained significant revenue and user traction. These companies have leveraged the new capabilities of AI to create innovative products and services.

Wave 2: Early startup adopters and fast mid-market incumbents
Startups have begun building on top of advanced AI models like GPT-3.5/4, such as Perplexity, Langchain, Harvey, and others. Additionally, a select few multi-billion dollar companies like Navan, Notion, Quora, Replit, and Zapier have quickly embraced AI-powered products, establishing themselves as early adopters of this wave.

Wave 3: Next wave of startups
The upcoming wave will witness the founding of new startups that explore AI applications in voice, video, and other formats. Natural language processing will extend to more verticals and diverse applications, accompanied by the emergence of new types of infrastructure. Companies like Eleven Labs, LFG Labs, Braintrust, and others will contribute to incremental advancements in AI experiences.

Wave 4: First big enterprise adopters (anticipated in 2024/2025)
As enterprise planning and build cycles are typically lengthy, we can expect larger companies beyond the likes of MSFT, Adobe, Google, and Meta to launch fully developed AI products in the coming years. This phase will mark the widespread adoption of AI in the enterprise sector.

Building Moats in the New Era:
In the context of AI, companies that possess the best products, talented individuals, and rapid growth are the ones that require moats the most. Moats act as barriers protecting a business's margins from competitive forces. Hamilton Helmer's book "7 Powers" identifies seven types of moats: Economies of Scale, Network Effects, Counter-Positioning, Switching Costs, Brand, Cornered Resource, and Process Power.

However, it is important to note that moats alone cannot guarantee success. Startups must first achieve Product-Market Fit (PMF) before focusing on moat-building. Uncertainty acts as a training wheel moat, keeping competition at bay until a more permanent moat is established. The level of uncertainty surrounding an idea determines the speed and depth at which moats need to be developed.

Uncertainty in the AI Landscape:
Uncertainty manifests in two forms in the AI landscape: Novelty Uncertainty and Complexity Uncertainty. Novelty Uncertainty refers to the uncertainty surrounding whether a company can actually build what it claims to build. On the other hand, Complexity Uncertainty assumes that the product can be built but questions whether there will be a profitable market for it.

The equation for determining the depth of moat needed is as follows: Depth of Moat Needed = How Obviously Good Your Idea Is - How Hard it is to Build. The easier it is for a startup to raise funds, the more immediate the need for moats becomes. Airbnb's struggle to raise funds initially allowed them the time to develop moats such as Brand and Network Effects, ultimately leading to their success in the market.

Conclusion:
As we navigate the early days of AI, it is crucial for startups and companies to recognize the transformative potential of this new era. Building moats to protect businesses from competition is essential, but only after achieving Product-Market Fit. In this rapidly evolving landscape, embracing uncertainty and strategically developing moats will enable companies to thrive. Before concluding, here are three actionable pieces of advice:

1. Prioritize achieving Product-Market Fit before focusing on moat-building.
2. Embrace uncertainty as a training wheel moat, allowing time to develop more permanent barriers.
3. Tailor the depth and speed of moat development based on the level of uncertainty surrounding your idea.

By staying ahead of the waves of AI innovation and strategically building moats, businesses can position themselves for success in this new era of technology.