Katherine Radeka’s Post

Hey, LinkedIn these AI-written articles are just . . . wrong. Not because they're AI-written but because they are representing things as fact that we product developers have debunked a long time ago. It's garbage in - garbage out. I don't want to contribute because I don't want to mislead anyone who might think the rest of the article is OK. This particular article is riddled with poor assumptions about how innovation and product development programs actually get done. It assumes, falsely, that all the work needed to develop a new product is known and predictable. That may be true for construction, which is a much better fit for the techniques described here, and I suspect that the AI engine is drawing upon sources written for this industry and others that have low uncertainty but high cost of change. But there are significant differences between building a new apartment complex and building a new physical product which has both high uncertainty and high cost of change — and even more with software which has high uncertainty but low cost of change. The key difference is that in both hardware and software product development, I'm usually trying to make the most of the schedule and resources that I have to deliver the best product I can, not trying to figure out how long something will take. This is the driving motivation behind Agile Software Development's Sprint Planning process for software. It's why physical product development teams do best with Rapid Learning Cycles in early development followed by Execution Cycles (and Integration Trains where both hardware and software components need to come together). How do you measure and improve estimation accuracy and reliability for a product development project? Nobody cares, as long as a good product gets out the door on time, with an acceptable level of quality. How do you get the right product out the door at the right time and the right price? That's the question product development leaders care about.