That's a simple measure new startups could adopt, and I keep being surprised YC companies are not looking for remote people.
Were one cynically inclined, one might wonder if remote work is unpopular because it lets people bypass some of the status-signalling systems that tend to develop within organisations.
That is bad for team cohesion and serendipitous collaboration. Humans have evolved a number of emotions that enable us to work together even when it's not in our immediate self-interest: guilt and shame to realize when we've hurt somebody else and discourage us from doing it again, excitement to pick up on and pile into promising ideas created by others, compassion to help out members of the group who have fallen behind, trust to suspend disbelief and commit to a direction without knowing exactly where it will take you. Teams that lack this communication channel tend to spin apart into little atoms each doing their own thing. Without feeding off each others' excitement, there's little common purpose. Without compassion, it's easy for some workers to fall behind and slow everyone down. Without trust, the team can never take on a goal where the outcome is uncertain or unknown at inception time.
Status-signalling is a byproduct of this emotional communication channel. It's people gaming the system to generate the emotions that the group needs to function without actually doing the work.
I'd love to see remote work become viable too, but it needs a viable way of solving the "presence problem", the need for emotions to come through remotely as well as pure factual information. Without that, you throw the baby out with the bathwater. I looked into this when I was searching for startup ideas recently, but all technologies I could think of were easily 5-10 years away from being viable.
My personal preference, as a developer, is for remote work. But when I look at what the team actually accomplishes, the level of creativity in its solutions, and how quickly it accomplishes them, a colocated team almost always beats the distributed one. There's an energy that you can get when everyone is in the same room that's impossible to replicate when they're 1000s of miles apart.
The cases where I think remote work can do fine is when you have a simple tool or interface that needs implementing, or if you're just trying to optimize one number (eg. performance of a library). These are the same situations you'd use a contractor for.
There are big companies that work the same way. Cisco actively encourages employees to work from home. However, Cisco's product-development strategy is "buy startups that are already succeeding in the market", and so if you work at Cisco your job is generally bug-fixing and incremental improvements to existing products. If this is the sort of work you want to do, there are a bunch of options that will let you work remotely. I suspect that much of HN wants to develop new products though, and for them, empirically you need the team to be colocated in one location.