Steele's comments on open sharing of knowledge being the key to creation of vast amounts of wealth and prosperity rings very true to me. I think that this is at the crux of the impact of many 20th century innovations. Concepts like the transistor that can be relatively easily understood replicated and used in greatly enhancing the wealth and convenience of society at large.<p>The problem comes when the requirements for a technology entail a grand scale or immense expense, such that it's natural for those already resource rich to fence in 'innovation' and use it for the benefit of the few. I fear this will happen with AI if it requires a quantity and quality of data that only large corporations have access to for effective operation. Ensuring AI is truly open should be one of society's key objectives over the next 50 years. We need recognition that our data belongs to us collectively, and to disallow privatization of common value as has been the case with oil extraction and carbon emissions.