The SPECULATE engine and the doctrine for Structured Speculation were co-authored by GPT-3 from OpenAI. But what is GPT-3 and how does it work?
Can Computers REALLY Talk? Or Are They Faking It? from PBS Otherwords does a great job of describing how GPT-3 works. Predicting answers which are likely to answer whatever it’s asked based what it’s learned from ingesting billions of pages of data.
GPT-3 helped to craft the foundational methodology behind the Structured Speculation platform.
The fundamental principle of Structured Speculation is declaring a goal completed in the future, and then examining the conditions that would make that goal real.
With GPT-3 as a research and writing partner, it was possible to say:
I have an idea about a new way the people can come together to collaborate, here’s what I’ve got so far; what do you think? What would make this more clear? What if I wanted to make it more secure?
GPT-3 was able to consider everything that we’d written together about Structured Speculation, and infer the answers to my questions based on all of the knowledge it has ingested.
Humans’ ability to reason through complex tasks has long been one of the hallmarks of our species.
In episode three of the brilliant series Mindfield (The Stilwell Brain) Michael Stevens creates a giant model of how the brain processes information – out of hundreds of people from his home town in Ohio who volunteered to be neurons on a football field.
The human brain has 100 billion individual neurons. Working together, they allow humans to make sense of the world around us. With enough neurons, thought and inference about new information becomes possible by working together in a very logical way.
What if we could multiply our intelligence by providing the tools to seamlessly collaborate and attract experts to refine any process imaginable?
Structured Speculation extends what’s possible, allowing any individual with a goal to tap into the collective body of published knowledge and expertise in a scale never before available.