Possible is an experiment in collaborative & A.I.-assisted exploration of possible ideas. In its current form, each time you hit the refresh button, Possible will present you with a combination of two “market categories”. These are meant to serve as a starting point for exploring what ideas might exist in this new combined space.
I use an AI model called GPT-3, that’s developed by OpenAI, to try to suggest possible new businesses and innovations that might exist at the intersection of these two market spaces. I use a set of prompts I selected like “Brainstorm some ideas combining X & Y”, and pass them to GPT-3 for suggestions. Depending on what prompts you give GPT-3 it’ll give you quite different responses.
For context on GPT-3, and OpenAI, you can read more here: https://openai.com/blog/openai-api/. At a high level, it’s an impressive natural language processing model developed by OpenAI that can respond with text given a text prompt.
Why did I build this?
Beyond wanting to play with GPT-3 and try out some new front-end technologies, I was also inspired by an idea I read nearly a decade ago.
In the book “Where Good Ideas Come From”, Steven Johnson argues that most good ideas come from a combination or modification of existing ideas. For example, let’s rewind to the 1990s and the “world wide web” as a new technology that came to market. Traveling back in time we can start to trace how people began to combine this new technology of the “web” with existing concepts to produce new ideas and novel market categories. You get “web commerce” which becomes renamed “eCommerce”, you get “web publishing” which has its own unique flavor and needs new vocabulary to be described: like “blogging”, and “Twitter’s micro-blogging”.
He suggests that at any time in history there is a set of available technologies and ideas we have available to us that we can combine to generate new ideas. He calls this frontier of ideas and technologies the “Adjacent Possible”, extending the concept introduced by the biologist Stuart Kaufman to describe possible paths of evolution.
You can watch his TED Talk here, and some snippets I particularly liked were: