OpenAI gained much visibility at the end of 2022 with the launch of ChatGPT. The company was founded in 2015 as a non-profit to advance research in AI technologies and, in 2019, created a for-profit subsidiary to attract top talent and raise capital to provide investor returns. This evolution of OpenAI is the one we primarily see today in the market and in the products the company has launched (e.g., ChatGPT, DALL-E).
During OpenAI’s recent DevDay event (https://openai.com/blog/new-models-and-developer-products-announced-at-devday), they announced their latest model, GPT-4 Turbo, which is trained with data through April 2023, with a significantly larger 128k-token context window that allows the model to ingest ~300 pages of a book in one request, and with a cost at least 2x cheaper than its prior version GPT-4.
General-purpose LLMs (proprietary and open-source) are advancing rapidly, and OpenAI’s announcement of the GPT Store, which allows technical developers and non-technical users (a first) to create custom agents, is a relevant step towards making natural language interfaces a true platform (https://openai.com/blog/introducing-gpts).
We have had at least a few moments in history where new technology platforms matured to a level that allowed the proliferation of functionality in an easier and simpler way for users to consume. The Internet is a clear example, with no single party controlling the platform or accruing value for themselves. Apple’s introduction of the App Store in 2008 is another example of a new abstraction layer that allowed users to access functionality easily and simply without the need for a computer (Apple, in this case, controlling and capturing platform value for themselves, becoming one of the largest companies in the world).
Creating the capability for technical developers and non-technical users to create their GPTs (natural language apps or agents) and distribute these agents in a GPT Store could be viewed similarly to an “Internet” or “iPhone” moment that can lead to end-users accessing a conversational agent rather than going to a mobile app or to the internet (replacing the first point of user contact). The typed language interactions we use today will quickly evolve into spoken language and voice agents that can replace the traditional use of websites and apps.
Al Alten Capital we invest in tech services businesses. We enjoy the data and AI space and follow closely the advancements in GenAI. Please reach out to us to explore how we can create value together.