The Raspberry Pi is a hobbyist play thing but you'll be surprised how close it gets.
It's a general purpose computer with 64gb SD card, passively cooled 1.2ghz 64bit quad core CPU, no moving parts (vibration resilient), hdmi, ethernet, WiFi, ultra low power requirements (3-5W) in a tiny form factor (45g).
Looking good so far until we look at operating temperature rages. The Pi FAQ says the main SoC is good from 0-70 C. Unsure about the solder/PCB -- probably as good as my 760 radio solders which seem to keep breaking every year. I read one test which says it worked at 50C but started freezing at 55C. You're right - it's just too amateur. But that's not what people use.
Ok let's knock it up a notch to the Nvidia Jetson TX2 board. Expect about $600 retail. (US $399 for a thousand of them).
https://devblogs.nvidia.com/parallel...lligence-edge/
Firstly the thermal is -25-80C. That's more like it. 8-15W power usage (still tiny). But the specs are also a lot bigger: 2ghz quad core CPU, very powerful graphics chip (I'll get to this in a second). 8gb high bandwidth ram, 4K video encoding/decoding, 85g weight, dual CAN bus controller, high bandwidth video/sensor/data links, gigabit Ethernet, 802.11ac 2x2 wifi, blah blah etc etc. It's awesome.
Ok so this is what people mean by consumer hardware disruption: it's specifically designed for things like self driving cars, drones and so on.
Now the fun: the gpu isn't designed for playing games. It's for machine learning, which is how cars can drive themselves in real time. The chip in this thing (Pascal architecture) is one of the most advanced things in the world right now, save custom asic's developed by Google/Facebook for specific machine learning applications in data centres. Once trained from bucket loads of real world data, it can take in high bandwidth sensor data and output smart decisions like where to steer/fly.
Watch this and let me know if you think the regulator can keep up with this kind of technology:
https://youtu.be/BLlwm5Dq7Is
Self driving cars have already forced the hand of most worldwide regulators to accept them. The public get what they want.
There are challenges to short range flying drones because we're still at the version 0.1 stage (remember when smartphone didn't even exist? 2007). That EHang is a proof of concept at best, but it is happening.
I can honestly see dozens to hundreds of self flying drones on pre-programmed routes doing 10-30min trips across congested cities carrying 1-2 pax for $10-20 a pop in the near future, if somebody can solve the hundreds of puzzle pieces.
The argument "if Richard Branson hasn't done it yet so it must be Snake Oil" is false. Often we must wait for multiple technologies to simultaneously mature & reduce in cost to economically unlock opportunities. For self flying cars it has not existed until now. That is cheapish Lion batteries (sub $500/kWh), tiny cheap sensors, high performance lightweight embedded computers, machine learning algorithms & research, mobile payments and so on. Things I hope really happen soon are some battery breakthroughs in supercapacitors for ultra fast charging plus better specific energy (lighter) batteries.
10 years is a long time...How exciting for aviation! I reckon we'll see it. But it won't be Uber!!