--
This decision shows me that Meyer doesn't respect or understand engineering culture. She's bought into the management BS "accidental collaboration" rationalization for industrial age butts-in-seats ideology.
Engineering culture comes, to a great degree, from the way you treat engineers and the process of engineering.
Treating engineers like cubicle bunnies who just can't wait to get interrupted by their Pointy Haired Boss is not conducive to building a good engineering culture.
In fact, requiring people to be in the office shows an anti-engineering mentality, because engineering, an effort of the mind, requires situations that are best for the mind.
Two key things enable good engineering: Collaboration (which requires communication) and coherent thought (which requires silence or peace or the isolation from interruption necessary to do it.)
This means that even if every engineer is in the same room, they're going to start "working remotely" by isolating each other via the use of headphones, and a preference for non-interruptive working (Eg: send email, or an IM rather than walk over and tap the engineer on the shoulder.)
It's true that in an office getting together in a conference room to has something out is easier and more convenient, but the tradeoff is that even with all the isolation people try to put into effect interruption creep is a real thing- eg: meetings, etc.
Working remotely prevents these interruptions at the slight cost of a higher level of effort needed to have a "meeting" (using a virtual whiteboard or just a phone call or whatever.)
So, if you spend most of your time in meetings, then you need everyone together.
If you value engineering and spend most of your time engineering, then whether people are together or apart physically, they are all isolating each other and effectively "Working remotely".
IM, Email and other collaboration tools that allow engineer isolation work as well whether the engineer is in the office or across the country.
Plus, lets not forget the minimum 2 hours of lost productivity that comes form requiring people to go to an office- either the commute (and the resulting need to get into work)-or the long lunches at those free cafeterias, and the endless cycle of distractions that are accepted non-work in offices. A "15 minute coffee break" at the office really has a 20-40 minute work interruption, because it often involves other people, while that same break working remotely can easily be exactly 15 minutes, and likely will be shorter because 10 minutes is enough to get the same level of relaxation from the day.
Almost everything in an office is designed to distract you from engineering, and the cost of this overhead is significant.