I've switched to listening to NPR exclusively using podcasts. I do find I'm behind on current events though, even major ones, such as riots, but I attribute that to me not having added any current event shows, such as morning edition, which I believe also comes in podcast form.

I use Pocket Casts for android to handle automatic downloading of specific sets of podcasts when my phone is on wifi and plugged in, out of the larger set I've decided is interesting, so getting the content is really a non-issue for me (and I wholeheartedly recommend Pocket Casts!).

It's worth noting that my lack of awareness of current events persists even though I get a good dose of random pop radio in the morning while dropping my children off at school, so I don't attribute that problem as much to a lack of FM radio as I do to a lack of news radio.

You're also missing local news and much of the mid-tier state news. That doesn't always matter, but when it does, it really does.

Instead of the "directly-streamed linear programming" model, you could do radio audio broadcast the same way you do GPS: just have the local radio towers spraying some selection of podcasts on a loop (with high enough frequency+bandwidth+compression that the effective audio rate is faster-than-real-time, unlike satellite radio), and then have "radios" in cars (or wherever) which catch the files and cache them.

You can build whatever you like on top of that—a hybrid device that will siphon up local OTA content along with arbitrary podcasts when it can find 4G, and then intelligently arrange the combination into linear programming, for example. (Really, though, you're just creating a weird sort of secondary internet that only allows retrieval of files relevant to the local area. In fact, you could make the "radio" into a 4G picocell that actually re-serves said limited internet to your phone, where the OTA content then shows up in your regular podcast app.)

Effectively, this would be the over-the-air equivalent of how pay-per-view (not video-on-demand, the older one) worked, but with the client device then acting as a set-top box to turn the received broadcasts into something equivalent to video-on-demand.

This would also be equivalent to how the http://en.wikipedia.org/wiki/Satellaview received game programs.

You technically could, but wouldn't the complexity and the cost of migrating over to a system like that completely outweigh the benefit you could reap? Think of all the cars that only have FM/AM tuners, and all the stations that would lose listeners the moment they transition their programming model over.

Not to say its a totally awesome idea in terms of features and functionality. The amount of inertia it takes to move away from a proven technology wouldn't let such a system see the light of day.

This sounds like a ton of work to do what FM already does reasonably well; if anything I'd like to see the amount of funding, variety, and _bandwidth_ for public radio increased, hopefully breaking some of clear channel's grip. I know everyone likes technologically advanced solutions, but honestly for anything that isn't straight broadcast, such as your suggestion above, the amount of work needed to workably implement it on the surface appears huge; why replace what just works with something else that doesn't have a immediately distinct advantage?

That sounds super complicated compared to good old FM radio.

You're imagining FM radio as some magical simple system that turns audio files into airwaves. But guess what's in the middle? People. People making decisions about what airs, in what order, on a very limited amount of real-time audio band.

This would move that decision-making to the client device, which could follow any algorithm the user preferred. There would be no people "working" at such a radio station. It would just be a box out on a hilltop with an antenna, retrieving files from an S3 bucket and spraying them—a bit like a numbers station.

Now, there might be some curation going into what counts as "local content" for the box to download and broadcast, but it could be algorithmic or just economic. Picture something more like the "curation" that Public Access television gets: anyone who wants to show up at the station and book a booth for an hour for a fee can have a "TV show." In the same way, anyone who wanted to hand the maintainer of the box an RSS feed of their podcast could have a "local radio program" that cars could receive.

The key, here, is that there would be an unlimited number of such "local radio programs."

Don't imagine radio as it is today. Imagine driving into the vicinity of a city and "picking up" the YouTube channels of everyone who lives there.

That's not how GPS works.

I think a nice change for FM (and AM), rather than going away entirely, would be to turn over the bandwidth cheaply or freely to public radio, non-profits and other small time operators, maybe like college radio. I think it would be an excellent way to introduce more diverse and interesting programming without being totally beholden to the lowest common denominator, and kind of get away from the awfulness that is clear channel. Also having a more community focused bent would work wonders; CPR makes NPR even better around here (Northern Colorado), as it's not just national news, and you get a lot more really local stuff (including music). I think there's still a niche for FM out there, as I certainly enjoy listening to the lower end of the dial on my commute, as do many people I know.

> I'm curious, what do you think will replace FM NPR in cars?

Time-shifted listening, like e-books, podcasts, etc. The real question is, why listen to Prairie Home Companion on their schedule.

By the way, that's the killer app for FM in mobile phones: Time-shifting shows that are not available for downlaod.