The authors also outline six imperatives for building enduring AI-first companies. These include creating and sustaining an undeniable data advantage through the development of 'designer datasets', and recruiting and empowering AI scientists. They also highlight the importance of reinforcement learning with expert human feedback, and the potential for AI-first companies to leverage their position to serve adjacent customer segments and create new categories.
Key takeaways:
- AI-first companies are in the business of advancing AI as a science, whereas AI-enabled companies are implementation and distribution machines. The two company phenotypes establish moats at different layers.
- The impact of AI-first companies will be greater, financial returns superior, and moats more enduring than their AI-enabled counterparts.
- AI-first companies exhibit an insatiable appetite for data and employ creative means for acquiring it sustainably. They develop designer datasets that are uniquely suited to deliver high performance on specific tasks.
- Recruiting and empowering AI scientists is crucial for AI-first companies. They require deep AI research acumen, investors willing to take a long view, materially more capital, and potentially less conventional business models than AI-enabled peers.