How interesting could those kinds of jobs really be?
And if you really need or want world class talent, why would you limit the pool to people who want to commute into San Francisco?
The reality is typically the opposite. Most startups are technically straightforward to start with and only get really advanced with scale.
Each additional magnitude of scale also brings a new wave of engineering challenges and opportunities.
I'd be shocked if this didn't apply to Airbnb. This is just as much a teaser of what's to come as it is a shrine to what 's been built.
(I personally have don't have the SF requirement nor any comment on that - Kaggle fully supports and promotes a distributed workforce)
Right now, you essentially have to be a part time hotelier if you're renting out one or two AirBNB room and a full time hotelier if you're renting out more. The intersection of people who both have room free and are willing to be hoteliers is pretty narrow. I know people who go on vacations of 3+ months at a time with their rooms empty and leave $10,000 of revenue off the table simple because it would be too much hassle to figure out how to effectively AirBNB or craigslist the room.
What I want is a magical team of professionals who enter my house an hour after I've left for vacation to find my house in whatever messy state it's in. They clean, tidy, professionalize and convert my house into something that's rentable to other people and I don't have to think about my house while I'm away, everything is handled for me. An hour before I get home from my vacation, that same team of professionals restores my house back to exactly the way it was. I'd even be willing to split the revenue 70/30 AirBNB's way if they could handle this because the alternative is $0 for me.
Moving up the value chain and taking over more of the travel experience is also something AirBNB is uniquely capitalized to do. AirBNB is experimenting with this with it's neighbourhoods feature and tentatively trying to become a more complete travel solution. But you could imagine AirBNB being an ad platform for other travel service providers for example. aka: pre-commit to enough museum tickets/sports games/restaurant reservations etc. before you arrive in town and we'll throw in the room for free.
Alternatively, AirBNB could start owning more of the travel vertical. You could imagine AirBNB airport shuttles that drop you off at your AirBNB door (with your key and Wifi credentials provided by the shuttle) if density in a region gets high enough.
It doesn't take much to imagine how much is left to be built of AirBNB as well as pretty much any other technology company. Talk to anyone in any company and they'll tell you they're still only 10% of the way there.
The truth is that every year is way bigger (and thus harder) than the previous year. We have a graph that shows traffic over time vs response time over time, and thankfully those lines go in opposite directions, but it's only due to the continuous work of dozens of engineers as we discover new ways that large amounts of transactional traffic and data can play mischievously together. And while doing that we also have to move the needle on product: if you look at a screenshot of Airbnb two years ago and compare it to today, it's very different (especially on the native apps and mobile web). And the screenshot you'd be looking at is probably just on the guest side, which is simpler: on the host side, we've done even more that's invisible to the casual traveler, like automatic pricing recommendations based on geographic and seasonal data, major investments into fraud detection and prevention, global currency exchange, etc. Not to mention all of the tooling, monitoring, and alerting that needs to happen to support all of that growth.
None of this stuff is free and usually it requires rethinking how old systems work and interact with each other, especially as we do more complex calculations with larger amounts of data. To be honest I'd say we're generally about a year away from hitting the wall, every year. We've made a lot of improvements (when I joined almost four years ago we were maybe... two months away from hitting a wall? Super scary), but if we all just phoned it in the growth in usage would overwhelm us quickly.
Re: San Francisco — yeah, it's definitely tough to recruit out of a single geographic pool. SF is probably the best area to do it in, since there's a large cluster of really great engineers here, but there's no doubt we miss great talent that won't or can't relocate. It takes a lot of work to do remote engineering offices well: how do you make sure that the engineers in satellite offices feel connected to HQ? How do you make sure they work on meaningful projects, but also don't miss out on context and information that comes from humans naturally chatting to the people that are around them? These aren't unsolvable problems (and a number of more-established, older companies have already solved them to some degree, and even some startups that are trying the Moneyball-style approach), but as a team we haven't done it yet. We do have remote non-engineering offices, so it's not impossible to imagine some day having a distributed eng team, but today we're just in SF. Some day, maybe? It takes work but there's certainly payoff for extending your reach.
Maybe its AirBnB's primary product, but it made me think of the old Mapquest vs Google Maps "war" and how Google Maps eventually won out with continuously improving on the core product, even thinking waaay outside the box by strapping cameras on top of moving cars to construct a huge VR map of the world. I'm sure when Mapquest was new most people were thinking it was already a perfect product for what it did.