Fast Facts
- Success can create a blind spot, making organizations and models stick to outdated assumptions, which hampers adaptation when circumstances change.
- Both Kodak and Moody’s illustrate how systems can continue to adapt superficially while underlying understanding of the problem remains static.
- In data science, repeatedly updating models without revisiting the core problem framing risks embedding unexamined assumptions that no longer hold.
- To stay truly adaptive, we must regularly question the foundational assumptions behind our methods and models, not just their technical performance.
Knowing When to Revisit Our Assumptions
We often trust ideas that have worked before. When a model performs well, it feels safe to rely on it. However, things in the world change quickly. Populations shift, new technologies emerge, and markets evolve. Because of this, it’s important to ask: When should we test our assumptions again? If we don’t, we risk trying to fit new reality into outdated beliefs. Regularly questioning our assumptions helps ensure our models and methods stay relevant. It’s like checking the map before moving forward — assumptions need updating just as much as the tools we use.
Examples from Business and Data Science
Historically successful companies show us why this matters. One company, famous for film photography, struggled because it clung to outdated ideas. Digital photography challenged its core assumptions about what customers wanted. Meanwhile, a financial ratings agency relied on models based on past market conditions. When those conditions changed dramatically, their assumptions no longer held true. On the data side, models learn from historical data. But if the environment shifts, those relationships may no longer apply. Accepting change requires we reexamine the assumptions behind these models, not just update results.
How We Can Keep Assumptions in Check
To stay adaptive, organizations should regularly challenge what they believe is true. This means asking: What must change for our current approach to no longer work? It’s not enough to monitor performance. We must also evaluate whether the core ideas behind our methods are still valid. For example, updating a model is helpful, but understanding whether the problem itself has shifted is crucial. When success makes assumptions invisible, it’s easy to lose sight of which ideas are still trustworthy. By fostering a culture of curiosity and continuous questioning, we can better navigate change and keep our systems resilient.
Discover More Technology Insights
Explore the future of technology with our detailed insights on Artificial Intelligence.
Discover archived knowledge and digital history on the Internet Archive.
AITechV1
