Making a sandwich is something many of us do almost perfectly yet find challenging to explain how we do it.
This is an example of unconscious competence; a skill we’ve learnt so well we perform it without consciously thought.
Most of us are unconsciously competent in skills related to our professions and our lives. We know how to do them, but may not easily explain to others how we do.
When it comes to delegation and systemization of business activities, it is critical to detail these tasks so someone else performs them consistently and in a high-quality manner meeting our expectations. This is referred to as ‘conscious competence’, being aware of each step in how we perform a given task in an explainable form.
But this isn’t always the case with AI. Artificial Intelligence, or machine learning, can often leap from the need for conscious competence through learning from inputs and outputs and inferring steps in-between.
Give an AI enough examples and it learns to replicate the output. As well or better than most humans, and in a fraction of the time.
When I ask our AI how to make a sandwich, this is what it suggests, Our AI has never made a sandwich or seen a sandwich being made, however has read many accounts of sandwich-making.
Instruction to AI:
Detail how to make a sandwich step by step:
AI output:
1) Choose your bread. You can use anything from a bagel to a tortilla to sliced bread.
2) Choose your condiments. Common options include mustard, mayonnaise, ketchup, and BBQ sauce.
3) Choose your fillings. These can be anything from deli meats and cheese to veggie patties or hummus.
4) Assemble your sandwich by spreading condiments on your bread, adding fillings, and then topping with the other slice of bread.
5) Enjoy!
—-
Our AI did this in seconds. Faster than you or I could think through and document our sandwich-making steps.
Consider how AI could assist you at work by performing tasks you regularly do, but that take you hours or days.
You may not have to explain how to do these tasks, only show the AI examples of inputs and outputs you expect.
(BTW the image for this blog post was generated by the Stable Infusion AI)