Open source the ML models/data of search/recommendation

[MarkBruns]

"It's amazing what you can find [in Nature] when you bother to SEARCH." -- Sacagawea [chastising the affluent idiots she was guiding in 1805 by repeating an old expression passed down by hunter-gathers to their offspring for 100,000 yrs]

If we are really enamored by recent developments in information technology, we might believe that it is even more amazing what we can find when we automate the search process and gather intelligence to suggest recommendations.

Industries have been built on search technology but many still don't really think about fully exploiting the power of search. People still think that they can develop code faster themselves or build things themselves or limit their range of options to what is locally available. Refusing to search is an ARTIFICIAL constraint, basically a matter of religious isolationist fundamentalism.

The range of options that we find from even a brief search dramatically changes our approach to solving problems. There may be economic or industrial organization reasons why consumer-oriented search for retail might be best served by proprietary quasi-transparent-ish search, ie Google/Amazon...

... but for more genuinely transparent purposes such as scienific and research needs, we probably want to move aggressively to more OPEN frameworks such as the Lemur Project or Lucene. There are a lot of people working in this realm

What's even MORE amazing than what automated search and recommendation engines find for us is what entrepreneurs, open source developers, scientists, musicians and creative minds decide to build AFTER they understand how search and findability works.