Hurricanes suck.
What if we could redirect them so that instead of hitting Florida, they remain in the Atlantic before dying off?
Here's how we might do it:
- Predict where/when a storm will form.
- Send a fleet of 50 planes near that area. Their exhaust will change the temperature and velocity of the air in that area.
- This change in atmosphere will deviate the storm from its initial path, saving billions of dollars of damage.
The hard part is telling the planes where to fly. We'll need the following:
- A model to predict the storm's future trajectory given the current measurements.
- A reinforcement learning algorithm (Monte Carlo tree search?) to determine the best locations to "add heat" and redirect the storm away from people.
It'd be breathtaking to see aircraft carriers launch hundreds of jets to redirect the tempest.
We may not even need planes to fly; autonomous drones could be enough. Since this is a control problem, we'll need extremely responsive agents, i.e. computers.
On the other hand, this technology may encourage more development in hazardous areas because people may falsely assume that this technology will work all the time. There are also ethical concerns. If you can redirect storms away from people, you certainly have the means to direct them toward people…
Concerns aside, I need to find out if this fiction is the future. I hope it is because it feels like a fun problem to solve. I hope you have a great week!
- Curtis
After thinking about this some more, I believe that a more effective way to alter the temperature in localized ocean regions is to spray aerosols reflecting sunlight. This way, a single plane can have a meaningful and (moderately lasting) impact on that part of the atmosphere.
#97: How Exploding Furnaces Saved Me Hours of Debugging | #99: Grammarly for Cold Emails