Machine Learning Rapidly Improves Waste Sorting To Environmental & Economic Benefit CleanTechnica

Humans have been building machines to separate waste into different streams of different value requiring differing processes for decades. Until recently, we were mostly failing to do it well enough to be worth the investment. Instead, millions of people globally manually sort trash, sometimes with developed country workplace safety standards, sometimes living in developing country trash fields and scraping a living out of them. In London in the 1850s, when the population was roughly 3 million, a thousand rag and bone men plied their trade, greasy bags over their shoulders or slung on rough carts, picking through the detritus of the city to find enough items of value to allow them to pay for their lodging and food. In 1988, the World Bank estimated that 1-2% of the global population made most or all of its living picking through waste.