2 ms·
New google moonshot to do "molecular recycling", but their first pilots are for using ML and sensors to better sort waste streams. This is useful tech, but the
by datadrivenangel 2y ago
New google moonshot to do "molecular recycling", but their first pilots are for using ML and sensors to better sort waste streams.
This is useful tech, but the solution to most waste is to burn it for power with great emission controls.
- semi-extrinsic 2y agoIf we're eventually going to close the carbon cycle, it would be extremely stupid to go e.g. polyethylene -> CO2 capture from flue gas -> ethylene synthesis with H2 -> polyethylene. Sorting out the plastic and doing stuff like anaerobic thermochemical conversion is way more efficient. No need to get the oxygen involved.
- xnx 2y agoIs pyrolysis a good option?
- calmbonsai 2y agoWe already have companies using ML for recycling https://ampsortation.com/en-us https://ampsortation.com/en-us and this includes real-time plastics identification, but they don't do it to the classification level this Google subsidiary claims. Burning isn't viable atm for many waste streams due to costs, but eventually we'll be forced into it. Long before then, we'll have an i18n set of industry standards and enforcement mandating the equivalent of a UV-barcode that precisely identifies all plastics. This tech is merely a stop-gap, but a necessary one.