The author emphasizes that true enterprise adoption of AI is still many quarters or years away, as large enterprises are still trying to understand what AI means for them. Despite this, the author remains optimistic about the future of AI, citing Google’s MedPaLM2 model as an example of the potential impact of AI on industries such as healthcare. The author concludes by stating that we are only in the early days of AI, with both peak hype and peak impact still in the future.
Key takeaways:
- The author argues that the rise of new AI architectures, particularly transformer-based and diffusion-model based approaches, marks a significant shift in the field of technology, akin to the invention of the airplane after cars.
- There are likely at least four waves of AI adoption, starting with GenAI native companies and ending with the first big enterprise adopters around 2024/2025.
- The launch of GPT-3 and GPT-4 marked significant advancements in AI, with GPT-3.5 solidifying the perception of transformer-based models as the future.
- Despite the advancements, true enterprise adoption of AI is still many quarters/years away, as large enterprises are still trying to understand what "AI" means for them.