高爾夫灌溉用水達 548 billion gallons:別讓偏見左右對 AI 的指控
548 Billion Gallons. Hope You Enjoyed Yourself.
美國高爾夫球場在 2020 年使用約 1.68 million acre-feet 灌溉用水,約合 548 billion gallons;文章先以批評口吻描述這筆用水,最後才揭示談的是高爾夫而非 AI。作者藉此指出,面對 AI 的環境指控也應先查證證據,不要只憑大數字和語氣定罪。
Approximately 548 billion gallons of water in a single year.
In the United States alone.
About 1.5 billion gallons a day, averaged across the year, supporting an industry people can comfortably live without.
And people will defend it because they enjoy using it.
They’ll explain that it has value. That it supports jobs. That there’s an entire economy built around it. They’ll tell you that calling it unnecessary is subjective and that you don’t get to decide which activities everyone else should find worthwhile.
Meanwhile, the water keeps flowing.
Apparently, personal enjoyment becomes a compelling argument when enough paying customers share it.
Think about the scale. Hundreds of billions of gallons to maintain commercial facilities so customers can spend part of their day doing something optional.
Nobody needs to participate. Nobody’s survival depends on having access to it. You could stop using it tomorrow and carry on with your life.
Yet somehow, asking whether it deserves those resources becomes controversial.
The customers get defensive. The industry points to efficiency improvements. Someone brings up economic contributions, as though generating revenue automatically settles the question of whether something is worth its environmental cost.
Convenient.
We ask ordinary people to think about the water they use. Turn off the faucet. Fix the leak. Be responsible. Consider the consequences of your choices.
Then an industry applies hundreds of billions of gallons of water in a year, and suddenly everyone has a carefully considered explanation for why their particular enjoyment deserves consideration.
How much consideration?
Where’s the threshold?
At what point does “people like it” stop being enough?
Because if you’re going to lecture people about unnecessary consumption, surely this qualifies for your attention.
An enormous amount of water. An optional activity. Customers who could spend their time doing something else.
You’ve probably already decided how you feel about them.
Maybe you’re picturing the facilities. Maybe you’re thinking about someone you know who uses the product constantly. Maybe you’re preparing to explain that society survived perfectly well before it existed.
You might even be wondering why anyone would defend it.
I’m talking about golf.
U.S. golf facilities applied approximately 1.68 million acre-feet of irrigation water in 2020, according to the Golf Course Superintendents Association of America. That’s roughly 548 billion gallons.
If you thought I was talking about AI, take a moment with that.
I never mentioned a chatbot. I never mentioned training a model. I never mentioned generating an image or asking software to write an email.
You supplied all of that.
I gave you a large number and an angry description. You filled in the target.
If your reaction changed when you read “golf,” the water figure had nothing to do with that change. It stayed exactly the same.
Suddenly, you might have questions. You might want context. You might think people are allowed to enjoy things, even when those things require resources.
Where was that generosity thirty seconds ago?
Before the reveal, you may have been ready to condemn strangers for participating in something you’d already decided was wasteful.
Now it’s a familiar activity. Something your neighbor does on Saturday. Something you might do yourself.
The outrage gets harder to sustain when you recognize the people you’re aiming it at.
You can care about water and still ask whether a headline has earned your anger. You can dislike AI and still demand evidence that actually supports the accusation being made.
But you have to ask those questions before the story confirms what you wanted to believe.
Otherwise, anyone who knows what you hate can point you toward a target.
All they need is a big number and the right tone.
來源:DEV Community · dev.to