Mapnificent

Just came across Mapnificent, a service on top of Google Maps to aid in using public transit to get around.  It's exceptionally easy to use and quite intuitive, and to say the least, after 30 seconds I was already impressed with its ability to display the estimated area that you can cover within a given amount of time using a configurable combination of public transit, walking, and biking.

Plotting 15 minute reachability zones from two separate locations in night mode.

If public transit options in vanilla Google Maps are a top-down planning method (you know where you want to go, and GMaps tells you how to get there), Mapnificent is decidedly bottom-up (you decide how much time you have, and it tells you where you can go in that time).  Handy for finding a new place to go on your lunch break.
Just in time, too.  GMaps generates what I consider to be, in my case, a rather pessimistic route, considering that they had to take into account average walking speeds.  I don't walk, and I don't run slowly.  When it comes to making connections, there's a whole world of optimization that could be made to the Maps interface to allow the user to improve the quality of generated routes; Google usually wants me to wait 13 minutes to make a connection after a 10 minute walk.  An quality route for my case might end up being a 3 minute wait after a 5 minute run).  Not that I would have gotten around to it, but I had been pondering writing a quick and dirty python script make a local index of RTD's (Denver Metro area's transit agency) schedules, and generate more realistic plan.
Mapnificent gives you a slider to adjust the maximum amount of time you want to spend en route.  I'm hoping they'll follow suit and use more sliders to let you specify your average foot and bicycle travelling speed (on level ground).  Furthermore, it seems entirely reasonable that they should let you specify the minimum time spent in transit (with a bike or walk exclusive option--check all that apply) so that you can reinvent your exercise/recreation routine, and further specify if those times are for one-way travel or also include the return trip, including time spent at destination (thereby solidifying it as a lunch-finding tool, if it were also adapted to use GMaps standard car-direction services).  Heck, taking a simple lesson learned from Wii Fit, add options to plug in user height and weight, and it can estimate the number of calories you'll burn on any given trip.
This is one service that has a lot of potential.  I hope to be seeing further developments from these guys soon.