[petra]

>> And the same process that created cheap and easy tools for software, the same process that dropped the cost of starting a SaaS business 10x will happen for robotics/ai/IoT.

So OK, we've got tools for the IOT(mbed/arduino, etc). The key question is - how do you protect your product from being rapidly copied by big competitors(or worse, china),when they are using those same tools ?

In web development, we did it by getting tons of users really really fast and using that as a tool. But hardware markets are much slower.

That leaves us having unique, well protect IP(if you want a unicorn). And maybe this can be created by a startups, but a totally different type of startup: Maybe based on long university research. Maybe based of on a group of exceptional multi-disciplinary people with unique skills working together - and not a bunch of kids with an idea.

[willholloway]

I think the key to network monopolies there is going to be iot device + backend. The backend/network is where the lock in/value will come from.

[petra]

I don't understand, could you please expand?

[tdaltonc]

It's about iTunes, not the iPod.

It's about the app store, not the iPhone.

It's about the phone contract, not the phone hardware.

It's about the K-cups, not the Kurig.

It's about the cable subscription (and the ads), not the set top box.

[vorador]

It's a pretty simple idea actually --- tie your IoT device to a backend service and never worry about competitors stealing your UI/design.

Case in point: the Nest Dropcam uses backend services to do image detection and alerting, two things which require skills hardware manufacturers often don't have.

[willholloway]

Heres one idea for such a business. Design and build some type of intelligent, networked sensor that gathers X valuable data.

Find a way to get enough devices installed in the right places by the right people, or better yet take to seas.

Sell your sensor network data, which is analyzed with some type of machine intelligence to make it more valuable.

The ocean is a rich place to find such data.

Low power computing + motors + wires + rapidly prototyped printed, or machined or laser cut parts + low cost sensors + web backends + newly accessible ai libraries = infinite diversity in infinite combinations.

We won't run out of startups like that until we're living on a fully networked intelligent planet.