The seemingly overnight emergence of artificial intelligence (AI) at the end of 2022 into the wider public consciousness has been a long time coming.
While the technology has roots stretching back to Dartmouth University in 1956, it was the creation of the transformer model by Google in 2017 that unlocked its potential and paved the way to the recent creation of generative AI tools such as ChatGPT, Midjourney and DALL-E 2 that have captured the public’s imagination. Corporate interest, experimentation and adoption has been at least as great, if not greater than public interest. The underlying technology is triggering a profound wave of innovation which we believe will create huge, completely new markets.
Generative AI is a new class of technology that has dramatically improved the ability to identify structures and relationships in data to generate new and original content. The underlying large language models (LLMs) are trained on huge datasets in a process that facilitates extremely accurate prediction, resulting in the generation of content that can appear uncanny in its understanding and relevance.
While public tools have leant heavily towards text, image and other media creation, business usage has expanded into coding, biological sciences and other, less obvious areas of analysis. New use cases are emerging daily with huge research intensity and spend testing the limits of the technology.
Market creation through rapid disruption
Our experience investing in disruptive technology tells us that pairing breakthroughs in the underlying technology and the user interface can lead to the creation of massive markets that have never existed before. In recent history, the creation of the PC triggered multiple waves of innovation, but it was only with the Windows operating system that it became accessible to the general population. Likewise, the true value of the internet was realised by the creation of the web browser, which later lead to markets like the $570bn e-commerce or $750bn online advertising markets. Similarly, smartphones really came to prominence and wider adoption with the advent of the iPhone, from which the $6trn app economy evolved.
We also know that the rate of technology diffusion is increasing. The time from the emergence of a technology to its widespread adoption is reducing with each major wave of technological innovation, and indications are that this is true for generative AI too. The desktop internet took around 12 years from inception to reach 50% adoption; smartphones took about half this time to reach the same point, bringing the mobile internet to consumers which allowed Instagram to reach one million users in two and a half months from launch in 2010. ChatGPT reached the same number of users from launch in five days.
The reason for these analogies is our belief that the pairing of transformer models and LLMs with a simple to use, natural language interface is likely to prove the tipping point for AI, and generative AI in particular, where adoption can bloom well beyond the previous, more niche use cases. That 60% of workers are employed in occupations that did not exist in 1940 shows the power of technological innovation to create both physical and digital markets, and we expect that generative AI will again lead to the creation of enormous new markets, many of which might not yet be visible.
