Maxim number one: Don't judge a user's intent by his/her device. The basic error of a lot of mobile design is taking the canonical user to be someone in a hurry, on the bus, or only looking at the phone for a second. Designers limit the mobile user to things that are local and instant, which leads to condescending design and making the mobile app a lite version of the real thing. When you do a media query, all you get is the device context, not the user context. Do authors ship a hardcover version and a paperback version? No.
Maxim number two: Consider the laptop. And desktop. And phablet. (Sorry) The idea behind mobile first should be to front-load all your development pain, not to make something simple and then add features on the desktop. The design should be holistic, and even if you aren't shipping to them now, you should start considering the wider world of desk-nots: kiosks, watches, thermometers and other things that aren't desktops, but aren't phones, and may not really be mobile in any conventional sense.
Maxim number three: If there's a scent, users will chase it. There's a perception that the only way to get people to pay attention is to shove everything on the screen at once, but they are actually willing to interact with you, as long as you can give them confidence that you can deliver.
Maxim number four: Keep the originals Image selection should be server-side. Get really high-res originals, then downgrade them for smaller devices, and sniff the user agent on the server (or in js) to pick which one to serve. (This is what magazines do, for example. They store the raw version of everything and the printing process is responsible for making it bad enough to print small and cheaply. So when Time does retrospectives, it can print everything better today than it was thirty years ago, because they have great originals and better printing tech.)
[/caption]
I spent much of this week scurrying between my office and the