The Future of AI: Contrarian Theses and the Power of Large Language Models (LLMs)

The Future of AI: Contrarian Theses and the Power of Large Language Models (LLMs)

Introduction:
Artificial Intelligence (AI) has been a transformative force in various industries, revolutionizing the way we live and work. As the field of AI continues to evolve, it is essential for early-stage investors to understand the contrarian theses that challenge conventional wisdom. This article explores five contrarian AI theses for early-stage investors and delves into the potential of Large Language Models (LLMs) in shaping the future of AI.

1. Horizontal LLMs vs. Vertical LLMs:
   The prevailing belief is that LLMs will pave the way to achieving Artificial General Intelligence (AGI). However, some AI experts argue that LLMs are not the ultimate solution. Instead, they believe that the next breakthroughs in AI will come from new and superior technologies, not just an accumulation of more data. Additionally, the notion of verticalizing LLMs suggests that multiple specialized LLMs might prove more effective than a single horizontal LLM for specific applications. This challenges the traditional notion of a one-size-fits-all AI approach.

2. AI's Impact on Specific Companies:
   Unlike previous technological waves, AI is not expected to impact entire markets uniformly. Instead, AI's impact will be specific to individual companies. AI fundamentally applies intelligence and learning to different steps in the corporate value chain. Consequently, companies with similar value chain steps are more likely to benefit from AI. This insight suggests that industry boundaries might be less relevant than the optimization of value chain processes when evaluating the potential of AI for a company.

3. The Erosion of Competitive Advantage:
   The widely-held belief that companies can establish and maintain a significant competitive advantage through AI is challenged by the contrarian thesis that AI will ultimately erode most forms of competitive advantage. Google's famous "we have no moat" memo exemplifies this thinking. As AI continues to advance, the ability to leverage AI effectively will become commonplace, narrowing the competitive gaps between companies. This shift calls for a reevaluation of traditional notions of competitive advantage in the AI era.

4. The Bifurcation of the Economy:
   As AI progresses, it is predicted to divide the economy into two distinct realms: the real world and the AI world. This division will be propelled by the proliferation of AI agents, which will reshape customer acquisition channels. The increasing reliance on AI agents for various tasks may lead to a reduction in the use of software application interfaces. Companies that have adopted product-led growth (PLG) or community-driven go-to-market (GTM) strategies might need to reassess their approach as AI agents become more prevalent.

5. The Future of LLM Infrastructure:
   For applications that rely on LLMs but do not own the models themselves, questions arise regarding the long-term outlook for LLM infrastructure. Will multiple providers commoditize LLM models, or will a single cutting-edge company emerge as the gatekeeper due to its superior engineering capabilities, hardware, data, compute power, and community? Investors must consider the potential implications of LLM infrastructure development and its impact on the competitive landscape.

Actionable Advice:

1. Understand the unique steps in a company's value chain that can benefit the most from AI. Look for companies that have built workflows amenable to leveraging intelligence and learning effectively.
2. Evaluate the data moat and cost implications when considering LLM applications. Assess the feasibility of existing proof of concepts and consider alternatives to avoid dependence on a single provider.
3. Stay updated on the evolving landscape of LLM infrastructure. Monitor the potential commoditization of LLM models and the emergence of companies with superior capabilities in engineering, hardware, data, compute power, and community.

In conclusion, the future of AI presents intriguing contrarian theses that challenge conventional wisdom. While LLMs have played a significant role in AI development, the emergence of new technologies and the impact of AI on specific companies are factors that should not be overlooked. As early-stage investors, understanding these contrarian theses and their implications can help navigate the complex and ever-evolving AI landscape successfully.