How can we REALLY open up OPEN SOURCE machine learning?

[MarkBruns]

We really NEED [hundreds of] thousands of viable well collaborated open source machine learning communities ... each led by some heroic benevolent dictators for life well as solid corps of commitizens ready to step up to a temporary BDFL.

We ALSO need entrepreneurial open sourcers to turn the corner from working in a development community to launch solid no-bs MLaaS products which provides a legit intelligence gathering service to customers ... so that customers support the product to allow the MLaaS companies to make money necessary to pays somewhat decent salaries and give working folks some reason to expect that their invested hard-earned retirement savings will offer a reasonable return.

What is required is RADICAL OPENNESS ... or else, the whole game of VCs, funds and institutional hypesters starts sounding like some other PT Barnum hoaxster Dunning-Kruger sucker-harvesting abboir, eg like cryptocurrencies have become in the last five years ... it is the nature of things that even good MLaaS companies will behave like any other get-rich-quick casino option when the get-rich-quick plungers 🪠 show up to build new casinos.

Yes ... suckers are born every minute ... and jealous, envious greedy people are perhaps the biggest suckers ... but it truly is unquestionably great to see the funding success now experienced by richly-deserving group like the people of HuggingFace ... BECAUSE of what their example represents ... we should remember that their success is not quick but rather the result of a long history of building discipline and getting smarter and then progressively more discipline and recursively more intelligent in many dimensions of intelligence.