Earlier quoted context omitted.
Data centers use a lot of power. Make data centers more efficient, for example by improving GCC or the Linux Kernel, and you can save a chunk. Don't write your billion-user apps in Python, use a more efficient language like Rust or Go. Don't make 10MB webpages where 100kb would also suffice. But I think useful change can only be made using political means. Write your representatives and tell them how you feel. Vote f…
I was walking one night and I noticed streetlights are always on. Why aren't they proximity detecting/linked to each other so they'd anticipate you moving to the next one. Also same with stop/go lights. I realize infrastructure. Edit: I was thinking about lights, I don't know far away sensors maybe? Not helpful if the lights are always being used (lots of cars). Liquid salt thorium reactor haha
- proximity detector is another potential point of failure; including it raises both installation and maintenance costs
- when you're walking at night, you'd like to see further than the distance between two lampposts; this would require the proximity sensor to trigger several lamp posts arond you (totally doable, I guess, but again extra installation and maintenance costs for having the lamps wired together in the right way)
- in the past, streetlights were using sodium-vapor lamps, which have a significant warmup time
- now lamps are switching to LEDs so warmup time is not a concern, but I wonder how constantly turning them on/off impacts their longevity
Wearing my optimist hat, I think city planners are comparing costs of the issues above with potential savings. But pessimist in me tells me that it's mostly because electricity is cheap and incentives to optimize streetlights aren't strong.