Do you see one winning over the other? Like Bard versus Chat GBT and other players in this space, I don’t know. I don’t think there are any others that are worthy of that level, at least not yet. There are a few other players that are also doing really interesting things, such as Anthropic and Cohere. But to the meta question of whether there is a winning strategy or a winner, it’s not dissimilar to the operating system war.
When operating systems first came out, there were many flavors of it, but in the end, Windows, Macintosh, and Linux emerged as the dominant players. Windows and Macintosh became the duopoly that ruled the private consumer market, while Linux became popular among developers and hobbyists. Similarly, in the field of artificial intelligence, I believe that some flavor of these three will eventually win out.
OpenAI, with its large language models like GPT, is of the highest quality but offers less control. On the other hand, other models may be more open and hackable, allowing for more customization. The differentiation between these models lies in their ability to ground the system in truth. While GPT may excel in creative queries, other models may be better suited for factual-based tasks.
Data will play a role in differentiating these models, but not to the extent that one will completely outperform the other. Another factor that may differentiate them is the focus on creativity versus factual accuracy. Some models may choose to prioritize creativity, while others may prioritize providing grounded and factual answers.
Additionally, companies like Facebook are exploring an open-source approach, where they train a model and make it available for others to use and customize. This allows for more hackability but less control over how the model is utilized.
In the end, the future of artificial intelligence will depend on the choices made by these companies and how they prioritize different aspects of AI development. Only time will tell which approach will prevail.
