I've worked at startups that have collectively made great decisions for a period of a few years, and people left or changed their minds, and some of those decisions didn't hold up. Some of those assumptions shaped code that too difficult in the short term to change, and there was no political will to push it through. Short term leadership was great, long term was less than great but still good. For the particular company I'm thinking of, new engineering leaders joined and things seem to be trending back in the positive.
Their CEO came to us personally to look at the product we built for them, and set with me on the PC. He was an engineer, and asked very interesting questions. Later this product was used to win the next five Formula 1 championships. Unfortunately Honda had to sell it to Daimler.