[webmaven]

I'm not familiar with the market, but it sounds viable.

You seem to have missed one trick though: offering a 10-20% discount in return for adding their (labelled) data to the common training corpus.

[heshiming]

Good idea. Although most likely I'm serving a dev shop in the middle, not the end user. So I doubt they would contribute.

[webmaven]

Yeah, you would need to cut the middleman in on the deal somehow, or sideline them in an unobtrusive way (for example, set up or join industry-specific consortiums that will own the combined training data sets rather than you, but you help define the standards for that data).

You might find these interesting:

https://blog.gardeviance.org/2013/02/attack-defend-and-dark-...

https://blog.gardeviance.org/2015/12/open-source-as-weapon.h...

[heshiming]

Interesting read, thanks. Although in my life as a developer, I tend to avoid relying on seemingly open-source project that has a well-funded company behind. The "open source as a weapon" idea is probably more suitable for big corporations. And few people consider controlling an intermediary (like data) as "control". If I own the data, I'd rather choose a "real open-source project" to consume it, not something that could end up charging me a fortune in the future. Without the foundation software work, data is nothing. It can't be easy to get people onboard.