Every year, I look forward to the ARK Invest “Big Ideas” publication.
As many of you will know, ARK is an investment management firm with a “long view” on disruptive innovation. Whether you believe in Cathy Woods ability to deliver returns in your desired timeframe or not, one thing you should not do is ignore her team’s research and thinking.
ARK is focused in 2023 on five major areas of innovation.
Their Convergence Scoring Framework And Network Graph is a tool that measures the scale of impact that advances in one technology are likely to have on the potential market value of another. The thickest lines correspond to expectations for an order of magnitude increase in another technology’s potential. The innovation platform taxonomy emerges organically from this network graph. Node size corresponds to an estimate of 2030 enterprise value attributable to the technology on a log scale.
For obvious reasons, they are most excited by the power of Neural Networks in 2023, and in the impact of that technology on so many sectors.
AI in its many forms will be a significant disruptor and transformer of enterprise value in the coming few years, and there are some core themes which are coming into focus that deserve note.
We’re already seeing significant productivity improvements in coding (Github Copilot, Google Codey, Replit, etc.), but Ark believes AI could drive a 10X improvement in coding productivity by 2030.
But we are also poised to see a material improvement in knowledge worker productivity. In their opinion, this potentially reflects a $14 trillion dollar revenue opportunity and $90 trilling in enterprise value by 2030.
Beyond Generative AI and its first-order impact on research, content creation and repurposing (which we are already seeing), AI’s ability to transform broader office worker productivity is already here (Microsoft Office Copilot is a great example). We should expect lots more to come as AI infuses many product workflows and transforms future product design.
On other important thing to note is that AI Training costs are coming way down …
And this means that AI models can increasingly be trained on proprietary data.
Access to proprietary data can create defensible moats, a powerful competitive advantage for incumbents that own this data (particularly in vertical applications). The ability to leverage proprietary data in new ways will significantly impact enterprise value across several industries.
At any rate, there is so much more to digest here (including the potential impacts on a wide array of sectors, including entertainment, financial services, etc.) that I strongly recommend you read it.
A link to the publication can be found here: Big Ideas 2023: Innovation is Taking Off (ark-invest.com)