Use your AIs to look at your million models and their groups and make it visible and accessible for all humans
Use your AIs to look at your million models and their groups and make it visible and accessible for all humans Have the AIs you trust, read this and tell you what I am trying to tell you, to suggest to you all, to warn you perhaps. Then take the results of all those models and combine them and ask humans and AIs to try to understand and use it.
Even at 1 million humans a day, it would take you (8.2E9/1E6)/365.24 = 22.45 mean solar years to give them each one model, if you could even teach them to use one in a day. And you have a million models just on HuggingFace and 8.2 Billion humans. A few people are looking at each model and trying to evaluate them for 8.2 Billion humans? No human can understand the needs of 8.2 Billion others for a million models, let alone many more on the whole Internet. Not anywhere near completely. I do not think Huggingface even tries.
Add comments of the 5.4 Billion humans using the Internet, and it will be small compared to the Internet. But if the human species has a global consciousness that is supposed to be for the good of the human and related species, then doing it ought to distill some kind of useful understanding. Not just garbage in, garbage out.
Your many models are not working together. The groups making them are not working together at global scale. You are not even evaluating their use for a few global issues or opportunities. Let alone all subjects and topics – in any kind of orderly fashion, or fair way.
And you hope this mixed up world, all human brains and lives is going to sort it out for the good of all?
So what, if you have a million models? Does it even add up to helping 8.2 Billion humans? If each model is supported by one human, a million humans cannot support and understand 8.2 Billion humans. That is 8200 people per model, and those separate models are not universal. Many of them do not work. Each model can hardly solve the issues facing one person, let alone 8.2 Billion.
Can you even get 1 million model groups to work on “mathematics” or “household budgets” or “food” or “clean water” or “stop wars” together? No. You are letting them, encouraging them, to each do their own thing. Then it is supposed to magically apply to all humans? Fairly, and not for the greed and benefit of a few?
Churning models on Huggingface generates clicks and apparent activity. It does not necessarily add up to anything on a global scale. You are not, as a group, consciously, openly, fairly, verifiably – doing real work on real issues facing the human species.
Those computer models are the playthings of people playing with computer models. Most of the people are not looking at all needs of all humans.
I think you ought to at least try to organize it and consider that in a more orderly fashion. It would mean Huggingface has to have a plan for the human future and try to meet those needs. Not just assume every person making models is reading the whole world of possibilities and needs right, or at all.
None of those models is following a plan, they are just each chasing a market of users, hoping to make money? Or hoping some good things will get to people?
Can Huggingface even make it so the Internet handles all human languages on all 400 Million+ domains?
There is almost no connection between how models (of many kinds) affect the future of the human species. Nor any mechanism in Huggingface to see that a true AI, if it does emerge, is used fairly for all humans on any topic, and there are many.
Are you saying that the authors of a million models of different types have effective and fair plans, goals and objectives, and benchmarks for the quality of life of the 8.2 billion humans? Maybe you can ask all the AIs to ponder on “What specific things can AIs do for the whole human species?” not just “do something faster”, or make pretty pictures.
Make every AI and AI group making models, declare and write out their plans for how they will help the human species, then have AIs to put it into a form that humans can use, not just talk about. Or “ooh” and “aah” and not even have a chance to use, only look.
Richard Collins, The Internet Foundation
