technology

We're not actually very good at weather manipulation, either in theory or in practice. Maybe counterintuitively, we have a much better handle on what we'd need to do for climate manipulation (especially via aerosol injection), and there's definitely a robust research program investigating that, though it's relatively new. We started systematically studying geoengineering proposals as part of CMIP6 in 2015 (I was actually part of the inaugural working group!) and it's a pretty significant part of the overall effort now.

Our understanding of weather (as opposed to climate) manipulation is much shakier. The most high profile attempt to engage in it was probably in China before the Beijing summer Olympics, and there's not even a widespread consensus if it was successful. They attempted some significant cloud seeding to try to keep it from raining on the games, and it didn't rain, but we're not very confident that it was the cloud seeding that did it. Part of the reason to prefer physically grounded models, though, is that it's relatively easy to incorporate this stuff. The model doesn't care if cloud condensation nuclei are injected or naturally occurring. Would this have long-term butterfly type effects? Yeah, definitely, but the weather system is so chaotic that it honestly wouldn't really matter much. There's already a pretty hard time-horizon of about two weeks beyond which we might as well just be throwing darts to make predictions, and forecasts are really only reliable for 5-7 days out. Introducing deliberate weather manipulation wouldn't put us in a significantly worse position, and there are pretty hard to overcome mathematical reasons why it's challenging to improve forecasts beyond that timeframe already (at least for weather--obviously climate forecasting is different).