Brian Gongol
Life is better for most people on the planet today than it ever was before. But improvements in the quality of life are driven by lots of little incremental changes that accumulate to really big things. Here are some of the things we need next:
- Calendars that tell us the weather.
Whether it's Palm Calendar, Microsoft Outlook, Google Calendar, or one of many other services people use to manage their daily schedules, we still haven't seen the arrival of a fully integrated calendar with obvious cues to the weather. It makes no sense for us not to have these things; we have reasonably good predictive powers for the weather (forecasts are pretty good out to about three days in the future), and it makes perfect sense that our calendars should at least tell us what the weather will probably be like over those coming three days. Weather can affect social plans, travel plans, exercise plans, and even just our moods. Seeing the weather right next to the planned "to-do" list for a day would help us make better plans. As it is, nobody really remembers the forecast very well. Why not put it directly where we're likely to need it most?
We've needed this at least since: October 2009
- A Pandora service for creating original music.
The main attraction of Pandora is that it can recognize the patterns that cause a person to like or dislike lots of different kinds of music. But what we don't have yet -- and should have -- is a system that takes those same patterns that we like and creates new, original music in real time to match our tastes. It shouldn't be all that hard; Pandora relies on the Music Genome Project to make its predictions, based upon listener preferences like a favoritism towards uptempo songs with a strong bass line or a preference for edgy guitar solos. Music, meanwhile, is a deeply mathematical construct, with predictable patterns that build upon simple relationships like 4:4 time and the pentatonic scale. A reasonably good system for artificial intelligence should not only be able to draw from a library of songs already created by humans to match those attributes preferred by a listener, but also to generate original music based upon those same attributes. There will always be room for human artists, but why not put our computers to work coming up with new material that we might enjoy? Perhaps we could even inspire our own music by setting parameters and adding constraints, just like an online fractal generator allows us to do with pictures.
We've needed this at least since: October 2009
- Self-driving cars
Once one learns to understand the value of time (and the time value of money), it's easy to understand why well-paid people have drivers and fly on private planes. But there are millions of Americans who spend lots of time driving who could be putting that time to much better use if they could let the car do the driving for them. The average American drives 13,476 miles per year. Even if those miles were all being driven at 70 miles per hour (which they aren't; most are driven at much lower speeds), that represents 192 hours per year per driver spent behind the wheel. Even if those hours were valued at the current Federal minimum wage of $7.25 an hour (which represents a massive under-estimate for people making middle-class salaries around $50,000 a year), that's a lost value of $1400 per driver per year. That's a serious under-estimate, of course, but it gives a sense of the scale of the lost value that could be recovered if only we had self-driving cars.
We've needed this at least since: April 2010