|
| ▲ | idiotsecant 6 days ago | parent | next [-] |
| The techniques will get more efficient, but the quantity of training will increase monotonically. We aren't going to use less energy overall. The ratepayers are absolutely the ones who will lose out on this. |
|
| ▲ | dorkypunk 6 days ago | parent | prev | next [-] |
| https://en.wikipedia.org/wiki/Jevons_paradox |
| |
| ▲ | nitwit005 6 days ago | parent [-] | | People using consumer generative AIs are already using it for free, or very cheaply. It may be hard for falling costs to drive more demand. |
|
|
| ▲ | ericmcer 6 days ago | parent | prev [-] |
| The demand for energy will never go down, the more we can produce the more we will use. |
| |
| ▲ | lotsofpulp 6 days ago | parent [-] | | The article says > Electricity demand in the U.S. held steady for 15 years but, last year, it increased by 3%— marking the fifth-highest rise this century. More jumps are projected for years to come. https://www.eia.gov/energyexplained/electricity/electricity-... Total electricity generated has been relatively flat for a couple decades. | | |
| ▲ | 0cf8612b2e1e 6 days ago | parent [-] | | Surely EVs would change the equation. Also increasing installation of heat pumps vs gas heating. | | |
| ▲ | lotsofpulp 6 days ago | parent [-] | | Perhaps, my point was an electricity supplier that invested a ton of money in the early 2000s assuming that aggregate demand will keep growing forever would have been in for a rough time. A variety of factors may or may not make a future where aggregate electricity demand would increases, or stagnates, or even declined. |
|
|
|