Google saw neat idea, built cool software, users happy. Replace cars to generate map data with phones for utmost accuracy. Trends of maximizing profit from cool technology makes users sad and services suck like other popular sites. Host begs Google not to make maps suck next.

Every free tech service can be (notionally) graded on how much value it offers for how much of your data it sucks up and monetizes. Most of the time we see an initial high value that gradually decreases as the developer gets greedy.

Google maps has managed to become Google "best" service, but it's important to be aware of the scale of problem we'll see if they start squeezing it for increasing profits, since it sucks up so much of our data and manages to be so useful.

That's the bulk of it, but there were other points being made about solving for edge cases that seem less central to the point.

Google has a massive leg up in its technology and use cases so everyone uses it. The guy described it as a "1%" problem (as in the last percent of use cases that are hard to add to the app) and Google doesn't have as many of those.