As much as most current implementations of AI are garbage and some are in fact dangerous, as ecologically & economically unsustainable as its material basis truly is in its current state and will only get worse if its evolution slows down – it nevertheless really is the technology of the future.
Full stop.
I will argue this point until the cows come home – and I started from a position of skepticism.
Barring a catastrophic event (non-zero chance tbh), both the models and the servers will continue to improve, reaching and surpassing the point where they become sustainable, and language models large and small will become as much an interwoven part of our lives as cellphones have become.
However:
Capital-driven AI gets implemented primarily to increase profit by reducing the number of salaries the company has to pay to human beings. It’s short-sighted, unethical, and foolish if you know anything about using AI.
Labor-driven AI gets implemented primarily to assist human workers in increasing their ratio of outputs to inputs, drastically augmenting rather than replacing human competence and expertise. It’s responsible, moral, and intelligent if you know anything about working with AI seriously on a daily basis.
It’s relatively easier to see how AI can empower the “white collar” worker than it is to see how it can empower the “blue collar” worker. Capital-driven AI isn’t focusing on either, and Big Labor is too caught up in Luddite politicking to seize the opportunity on behalf of the workers of the world.
We need to de-hype the conversation around AI and start looking for both the ways in which it can be improved and the ways in which it can improve the lives of workers of all stripes and their local communities, especially in the global south, where the trickle down effect of American-led technological development is very slow to arrive and sparing in its beneficence.
“On that day, uttered the Lord of hosts, each of you will call to his neighbor beneath the vine and beneath the fig tree.” – Zekharia 3:10