> to detect ebikers (easy to tell from speed vs road grade) and weight their routes more heavily in deciding the best route to recommend.
seems like an anti-pattern that screws over people who don't have e-bikes.
If users can tell us that they're using an e-bike and we give a separate option then both e-bikers and pedalers can have navigation options that work best for the specific use case, rather than shoehorning both into the same option.
We need to stop assuming things using ML and pretend like that's somehow more user-friendly than the user directly telling us what they want. (See: constant griping about how poorly non-verbatim Google search is getting)
> to detect ebikers (easy to tell from speed vs road grade) and weight their routes more heavily in deciding the best route to recommend.
seems like an anti-pattern that screws over people who don't have e-bikes.
If users can tell us that they're using an e-bike and we give a separate option then both e-bikers and pedalers can have navigation options that work best for the specific use case, rather than shoehorning both into the same option.
We need to stop assuming things using ML and pretend like that's somehow more user-friendly than the user directly telling us what they want. (See: constant griping about how poorly non-verbatim Google search is getting)