Full Circle: The Return of the Prompt

Is prompting the future or was it simply the past we forgot?
Everyone now gets all excited over ChatGPT’s ability to generate information or action via a written prompt, a simple method of typing out what it is that you want and then being delivered the result. Whether that be a written response, a generated image or even a video that matches your request. Users are jumping onboard of this new-fangled way of accessing a form of knowledge or action, similar to when a trend goes viral, and everyone is quick to throw a bucket of ice over their head on camera.

Since the publication and promotion of AI systems such as OpenAI’s ChatGPT, people have been wowed by the amazement of being able to generate an outcome simply via a short-written prompt. Off the back of this seminally new way of accessing data, people have started generating and even selling prompts that users can quickly paste into AI systems, allowing an even faster and easier outcome. But is this really a new way of working?

The forgotten past
From the inception of the electric computer in the 1940s, operators have had to access information and action via a keyboard, using a written prompt. Information or action could not really be accessed via any other method. Typing “FIND” and entering the information you wished to access was the standard for decades. However, in comes the 1980s and the mainstream use of the almighty mouse. For the previous four decades the use of keyboard and prompting meant that mainstream use of computers was limited, due to limited knowledge and training. The mouse allowed even the least learned user to be able to use a computer simply and easily, rapidly pushing forward the average person’s use and access to information and action. The introduction of the mouse of course meant that prompting initially stayed within computer engineering society and eventually resulted in it disappearing all together. 

The resulting action of adopting the mouse meant that users could primarily access information or a form of action via clicking. Clicking on a folder, clicking on a browser, clicking on webpage, clicking on download. Despite being easier for the average user to adopt, it alternately was a slower form of action. This is why throughout the time of the websites existence; we have spent decades and millions optimising websites to reduce the number of clicks before resulting in an action. The rule of three even became common place phrase, where ideally you want no more than three clicks before a user can access an action on a website.  

The path dependence feedback loop
The on-going use of the mouse and the optimisation of the click has generated what is called “the path dependence feedback loop”. This is a self-reinforcing cycle where an initial event or decision (the introduction of the mouse) has a disproportionate impact on future events or decisions. The resulting feedback loop of the mouse meant we optimised an inefficient method. Instead of continuing to optimise and advance the “prompt”, which we moved dramatically away from, until now! This is not new; it is simply a forgotten way of operating. Despite Google for over two decades allowing users to use a prompt to find what they are looking for; we did not adopt this simple prompt method into websites. Surely it is easier to land on a website and simply tell it what you want? “I wish to purchase one black handbag” and boom, it is displayed and added to checkout.

What will be the future?
AI prompting has definitely brought back the concept of accessing data via a written prompt, however, are the general public more advanced now than they were in the 80s and will this allow us to now optimise this way of operating? Or will it once again fall away because on mass people cannot adapt. Only time will tell…