Hi! 👋 Welcome to The Big Y!
As large language models (LLM) become more prolific, there will be a strong incentive to monetize them, just as OpenAI has begun doing. The challenge will lie in creating a user friendly (i.e., low-code or no-code) method to interact and derive value from the models.
With regards to large models, such as GPT-3 or LaMDA, I currently see two approaches to monetization. First is a monetization model where a research-oriented company does the fundamental research, publishes the model, and then monetizes the direct use of the model. The second approach is where a more product-oriented company specializes in a direct outcome based on a model e.g., a topical chatbot service derived from an LLM.
Examples of the first type of companies include what we might think of as the incumbent tech companies that have dedicated AI research teams such as Google, Microsoft, and OpenAI. These teams create either open-source tools that others can use (for free) or as we’re seeing with GPT-3 and DALL-E, they’re distributing basic access through APIs.
There are several companies popping up that fall into the second category. One example is Cohere which has built a set of tools and services for building custom language models with the purpose of enabling their customers to create NLP-based offerings that they can leverage directly in their end products.
I think we’ll see more and more companies move beyond selling the initial types of “easy AI” such as recommendation engines to selling products based on advanced LLMs and more complex algorithms. I believe there will be an entire industry that serves the purpose of bridging the gap between increasing scale and practical implementation.
Saying hi from Southern France! I hope everyone is having a great summer and making sure to take time to enjoy their favorite activities!
Thanks for reading! Share this with a friend if you think they'd like it too. Have a great week! 😁
🎙 The Big Y Podcast: Listen on Spotify, Apple Podcasts, Stitcher, Substack