Making Open Data Work for Open Source Software

It's telling that big companies like google, facebook, microsoft are now open sourcing their machine learning frameworks - these things represent considerable investments in money and brain-power, so why make it available to potential competitors for free? Answer: because these frameworks are worthless without data to work from, and that's where the companies now hold their advantage and value.

I would love to see some open data foundation to work on this kind of thing, and maybe some kind of github-for-data too. Come to think of it, starting with some kind of git-for-data would be good start.

Smart people think that being smart is the most important character trait (frankly it's not in the top 5) and coders think code is the most important thing. Meanwhile artists, musicians, photographers and writers continue to be marginalized and mined by technologists and tech/media companies for big ad dollars but little cultural benefit.

I always felt torn between the various open source sects, but Creative Commons always felt right, more inclusive and opt-in, and somehow more important. Perhaps that's my artistic bias. Or maybe I just saw the future while everyone was worried about open source PDF viewers and reverse-engineering the latest half-broken audio/video codec.

It's the music, the photos, the papers, the art that matters. I abhor the word "content" but by all means lump in "data" if that helps move things forward.

https://creativecommons.org/about/history/

Block chains are one solution to this problem. They are open source data repositories. Currently, most block chain data is transactional in nature, covering transfers of currency from one account to another. As more distributed applications emerge that depend on block chain technology this may give rise to additional data sets.

If you want that to happen you ought to take some of the money you make programming operating systems and put it to work.