For any investor, the most important fact to understand is that AI is an exponential technology. The speed of its development and the implications that come with it are so gigantic that humans struggle to grasp the impact that AI will have. The difficulty in understanding exponential technologies like AI stems from a combination of cognitive biases, psychological barriers, the inherent complexity of the technology, and the mismatch between human intuition and the nature of exponential growth. We humans have a natural tendency to think linearly. We expect everything to change in steady increments.
I believe this bias is inherent in most predictions, including those from Accenture Research and McKinsey. I believe that the prevailing estimates of the extent of automation or augmentation in knowledge-intensive sectors are significantly understated. A case in point is the McKinsey Global Institute’s 2017 projection of 50% automation of knowledge workers’ working hours. In a subsequent update for 2023, this projection was revised upward to potentially 70%. I contend that such projections remain significantly conservative, and offer a more radical perspective in which I see 100% of language and knowledge work tasks eventually being fully automated, replaced by advanced generative AI.
It’s important that investors and entrepreneurs don’t get caught up in the linear thinking of an exponential technology. A new perspective can be gained by looking at AI as a general technology, like electricity.
Since the invention of electricity, it has not only brought us electric light, but has reshaped entire industries, economies, and societies. It also led to the Internet, which in turn created millions of new businesses that were not possible before. The Internet, built on electricity, enabled the emergence of today’s basic AI models, which in turn are widely applicable.
The most significant entrepreneurial opportunities in AI may not necessarily revolve around the foundational models themselves, such as GPT-4, Llama 2, Claude 2, Mixtral, or new emerging competitors. Instead, the real potential lies in using existing AI technologies as a platform to create innovative business models and ventures that were previously unattainable without the advanced capabilities of AI.
Equally important is the ability to anticipate which industries will become obsolete in the age of AI — just as the steam engine became obsolete in the age of electricity. Similarly, industries that relied on manual typewriters became obsolete with the widespread adoption of computers and word processing software. The once-thriving video rental industry declined with the advent of online streaming services like Netflix. Landline telephones became less relevant with the rise of cell phones and smartphone technology. In addition, traditional print media has faced challenges in the digital age as online news and social media platforms have gained prominence.
