Facebook’s open-sourcing of AI hardware is the start of the deep-learning revolution | Ars Technica

A good article describing the logic and also, to a degree, logistics of open source collaboration. Pretty much a normal situation where the goal is lofty and vast and the resources needed to produce works that can be commercialised are not clear; and no one is interested in a one-off, or unique result, as that, in biological terms, would be tantamount to producing a mule at best.

All of which is to point out the obvious, at least to those of us who are involved. Open source is… normal. But, evidently, the term, if not the actuality, still gives pause to most business people and even public sector officials. And it should. Any way of making and distributing a product should force the stakeholders into considering the process and its consequences, and “open source,” as a still novel, broad, even vague constellation of license and method and even unjustifiably uncertain legal consequence stands out, in much the same way as a pitch to unionise might. It’s not that it’s “disruptive,” it’s that there remain still too many uncertainties.

I suppose that situation is to my benefit. After all, as a consultant who resolves anxieties and sets up communities that work, I indirectly benefit from open source anxiety. But, there ought, at this point to be no more anxiety surrounding open source than any other production and distribution strategy. They’re all fraught, they’re all modes of risky business.

 

Collaboration is key to building the machine-learning boat and getting it afloat.

Source: Facebook’s open-sourcing of AI hardware is the start of the deep-learning revolution | Ars Technica