To mitigate that, Uber and Lyft heavily subsidize launches in new cities by pre-signing up drivers well ahead of initial demand, to ensure low minute order fulfillment. This is done by giving drivers activity bonuses in the initial new-city market formation stage, no matter the actual orders. This ensures order low latency fulfillment for early adopters in new locales, which leads to customer satisfaction, word-of-mouth advertisement and soon enough full marketplace formation (and lock-in).
If you were trying to compete with Uber in New York City, the main problem you would have is that you'd need a lot of capital to have hundreds/thousands of drivers to ensure order fulfillment on average under 5 minutes by having a driver as close as possible to any new order being placed.
Convincing people (using money) to stay idle despite lack of initial order demand (until the marketplace is formed) takes more than $5 per driver.
People aren't desperate enough to hustle for $5. Even $200 is low. I'd consider that the minimum to even have a program worth running.