Forget AI as an all thinking omniscient being. AI is simply a single, or series of, mathematical equations. Check out our development pipeline below.

The steps we take to built a mode or an AI are the same. We figure out what question we want to answer and create some hypotheses.

We then collect data relevant to the question and we build the model. The model can be built in a number of different ways depending on the question we are answering and can involve statistics, mathematics and operational research as well as sociology, psychology, philosophy and a number of other disciplines.

If the model is domain dependent it might be a military model or a government model and would need subject matter experts to advise on the development cycle.

We then run the model and perform testing and then validation and verification so we can be sure that the model does what it says it will do. We then check the output.


This is potentially a lot more boring than the idea of sky net but it is much much safer. The model or ‘AI’ is not thinking its is just a processor.

Even large language models are simply based on probability theory in statistics and models that look like they are sentient are just regurgitating large amounts of data to you.

Watch the FRACTAI tik tok on this