Generative AI and Fast, Cheap, or Good–pick 2

There’s a project management saying that goes Fast, Cheap, or Good, pick two. It’s referred to as the project management triangle, but I don’t think it is a triangle as much as it is an equation. The good or the quality of something is a result of the effort and cost. Now there may be cases where people are ok with a lower quality, but in most cases there is a certain level of quality that must be delivered. The base equation then is Time x Cost = Quality. If you decrease cost or time you are going to most likely negatively affect the quality.

Time and Cost are factors that can be affected by things like technological improvements. When technology has an effect on improving time or cost then there won’t be an impact on quality. For example a new technology can enable things to be done faster without losing any quality. Alternatively, technology might decrease the cost without a loss in quality. And of course sometimes technologies may decrease both time and cost without a corresponding decrease in quality.

This brings us to generative AI. Most folks believe that generative AI is a technology that will change this equation. In some ways it already has when it comes to writing. You can input a prompt and usually with a few refinements get a well structured response. It’s also looking like it can generate images which have improved significantly over the last year.

Unlike many others, I do not feel this is going to destroy the design industry. It still takes people to review and refine images. This just cuts down the time it takes to create concepts. It still takes creativity to come up with ideas for prompts and then judgement to select something that really represents or communicates the intent.

Outside of something like generative AI and even now that it begins to change how things are done to save time, it likely isn’t going to save costs. First the new generative AI models are all requiring subscriptions for their more advanced models and features. While seemingly nominal at this time, there is still a huge question around the data used to train the models and copyright. The outcome is there will likely be some increase in costs as they have to pay into the copyright system in some capacity.

So in the end the equation: fast, cheap, or good, pick two will continue to apply. While it may take less time to generate concepts, most organizations will still need to employ people to generate those ideas and refine them.

Comments are closed.