Time will tell, but my theory on these last mile businesses that are springing up is that they are cramming into the market at precisely the places where there is the least money to be made, but offers the most convenience to customers. In other words, VC's are subsidizing luxuries that customers themselves have shown they won't pay for. This would be a great business if customers were willing to pay $20 to have food delivered, but the truth is that they aren't. At that point they will look for businesses that have delivery baked into the business model (ie, Dominos), or just go pick it up themselves.
There seems to be an inherent assumption somewhere that these businesses that don't offer delivery are out of touch or not tech-savvy, but the article touched on an important point with this quote: “There were many pages of orders, not just one or two,” he says. Finally, when spotting this phenomenon at a Palo Alto macaroon store on day, Xu asked why these orders weren’t being delivered. The answer was that the owner has opened his bakery to bake, and not deliver stuff. He got similar answers from people in other businesses, ranging from toys to furniture, to restaurants.
Maybe those businesses aren't interested in delivery for a reason. Just because you can build software for a smartphone doesn't mean the fundamental laws of P&L don't apply.
More detail can be found here: http://www.wsj.com/articles/u-s-mail-does-the-trick-for-fede...
Its not going away because while members of Congress who would have liked to have killed it could sell quasi-privatization, various accounting mandates, and other hostile policies to their constituents as efficiency measures, they can't take the political backlash they'd expect from killing it outright.
It doesn't take a Constitutional Amendment to kill it, just Congress deciding that they want to badly enough to actually do it and take the political heat.
Are startup investors going back to profitability as their key measure? Some might hope so but I think traction and marketshare are far more visible right now...
The remaining question is how soon. A company trying to make use of a self driving vehicle for most hours of the day across a wide customer/usage base will find it far easier to spread the capital costs, than single purpose or single customer vehicles. There are also big advantages to having a larger fleet (economies of scale). Consequently the moment the self driving vehicles are somewhat practical, I expect massive attempts to corner the market for using them. (Human driven vehicles can still be used for trickier locations and journeys until the self driving tech matures.)
Regulation is going to be the thing holding everything back. I'm not talking about the disparate and protectionist taxi regulations that Uber just ignored until they were successful; I'm talking about real safety issues like air traffic control when everyone and their mother is flying things over populated areas. The regulatory environment around how all these drones will communicate needs to be developed and standardized -- that alone will take at least a decade because adequate bureaucracies need to be built up to develop, steward and enforce the regulations. Self-driving cars are even further off because they're going to be carrying human passengers.
We're at a point where the technology has overrun our ability to regulate it, and there really are things that need to be regulated -- Who is liable if two drones crash over your house and damage your roof? Do you force drone operators to carry insurance to cover damages in the event of mechanical failure? How do drones report their position to each other to avoid collisions? What data formats, units and auditing is required when communicating with other drones? There are thousands of questions, and none of them have even made it into the sphere of political discourse, much less legislative action. Current politicians don't even understand this stuff, so we have to wait for people who do to become old enough to get elected.
I'd say we're at least 15 years off from this future, if not 30. We'll probably see it happen elsewhere in the world much sooner than it will in the US and Europe. So any business starting up today will not be able to rely on self-driving cars or autonomous delivery to prop up their business model: it has to work with human delivery or it won't survive long enough to make it to the drone era. And for the ones who do, it'll become a race to the bottom with razor-thin margins.
Mark my words: most of the "on-demand" startups that are relying on contract workers are not going to be able to defeat the misclassification lawsuits they're facing. Some might have the funds to delay the inevitable by strategically settling the first cases they're hit with, but the laws and legal precedents in this area are quite straightforward.