That aside, 99.??% of startups will need to market heavily to compete successfully.
Edit: And I just noticed the (Software) parentheses in the title. I guess you can replace self-driving cars with advanced AI, say the first AI to pass the Turing Test.
For instance, I worked with an early stage hardware startup whose customers were developers (you couldn't actually do anything with the device if you weren't a developer. It was a SDK not a finished product).
They generated tons of press because hardware is much easier to understand, show and write stories about than software. But despite being in really high profile publications (millions and millions of people seeing it) - it failed to move the needle sales wise.
- Most developers aren't reading these publications
- Most press articles don't mention the developer angle
- You're just one of many many stories that day in a big publication.
- In some ways is _really hurt_ because it made it seem like it was a $100 thing you could buy at Target than a complicated SDK. This resulted in lots of non-devs buying, being wildly disappointed and wanting to return it.
But when they get to your website, they have to know why your car is better than the Toyota being released a month later, and that they won't miss their current manually operated car. Chances are, your LIDAR guy isn't up to the task.
I was co-founder at a project that has received 10+ newspaper articles (progressively better placed) and growth kept declining over the months only to become negative.
However, I obviously understand that one can never look at these factors in isolation and that there are multiple variables on which growth depends. I would also not consider press to be a part of marketing really.
This is definitely a product that needs media management, because the key question is not so much if it works but if it's safe. That has to be a key part of any self-driving-car strategy. And you also need a legal/lobbying team, as the comma.ai guy found out.
That is why while I appreciate what FOSS has become, as someone with coding experience since the mid-80's in home computers, I don't have any qualms with commercial software, specially since I also have the experience of what it meant failing to do pure FOSS business.
I know I used to feel this way before I really tried to launch a product. Now I know that there is no such thing as a product that is so good that it sells itself. In fact the best technical solutions I've seen have failed because the market doesn't care about perfection, it cares about mind share.
There is still a stallmanesque anarchist quality to some engineers that view commercial success as tainted. I think I can appreciate that in some philosophical senses, but let's not get too high on our horses.