Quick question. When you leave a room, do you turn the lights off? Yeah, me too! Our parents did a great job teaching us not to waste electricity, didn’t they? Especially with the price of electricity these days, it’s good NOT to be wasteful.
Did you know that when we hear ‘it’s in the cloud,’ it’s not?
There is no magic something or other in the clouds where all of our electronic pictures and documents are stored. Here’s a fun fact. “It’s in the cloud” is a sexy way to say we’re keeping it safe for you in our mega data center. ‘In the cloud’ sounds so much better and completely obfuscates what is truly going on.
Many years ago, data centres were nondescript, unmarked buildings where big companies stored towers and towers of computers, which did predominately one of two things. Stored all the backups. Or processed the back-end operations. Y’know, the stuff that nobody pays attention to until something doesn’t work, and then we all break out in hives because we were inconvenienced in some way. For years, that was the function of a data centre. The average person didn’t know they existed.
Until Saas and Amazon became things.
SaaS (Software as a Service) means ‘Use us. We’re in the clouds!’ Essentially, if you want to use that kind of software, you can access both the software AND processing power over the internet. The best thing about SaaS is that the issue of storing software and processing power is no longer your responsibility. You pay for it, and voila. They reside, protected, with the vendor, high above, in the clouds.
When Amazon entered the AWS (Amazon Web Service) business, it established that division as a cloud company. Today, AWS is the world’s leader in cheap, cheerful, and reliable processing and storage, from the ginormous to the small fry in the clouds.
Computers have been massive electricity hogs since the beginning of time. My husband is old enough to remember the rooms where banks of computers were stored and kept at a balmy temperature of around 45 degrees. I’m old enough to remember being yelled at constantly for accidentally closing the door to the computer office room, which had to be kept no warmer than 68 degrees.
In a data center, there is very little activity. See, there is little need for a big team of people. Imagine racks and racks and racks of computers stacked like soldiers against the walls. They only need to be tended to when the power goes out. Without human activity, there is no circulation. No circulation means the heat generated by the computers can not be reused. (As in moved around to other rooms on the same floor or in the same building) Sure, fans might help, but fans use even more electricity.
Our power grid is already under stress, and many of us are worried about what electric cars will do to it.
In late 2022, we were introduced to Chat GPT,
And suddenly, the world of AI came crashing in. ChatGPT, the most famous but not the only version of AI, uses ten times the processing power of a computer compared to a simple Google search. To get the most from ChatGPT, one must learn how to do ‘prompts’ correctly. That learning curve is almost never-ending. It takes at least four different tries to get the correct prompt to start to receive one decent answer. Those four different tries use about an hour and a half of ‘leaving the lights on.’ Gulp.
Imagine just 25% of the world’s 8 billion people leaving the lights on at least once daily. That is a ton of additional computing time and electricity use for those data centers, which wastes almost 100% of their electricity usage today.
Hand Wringing
I’m not sure about you, but I’ve seen much hand-wringing regarding electric vehicles and the power grid. I am still waiting to see ONE piece of news about the effect AI is having and will have on the power grid.
This week the 5Cs Podcast begins a new series about the electric grid and AI. The solution is NOT to ask humans to forgo their ‘addiction’ to technology or even not adapt to AI. The solution is to demand that the data centers become more sustainable. My guest shares a solution with a payback of less than one year and multiple millions in cost and grid savings, offering a beacon of hope for a more sustainable future.
Please make some time to listen to it. Especially if you use your telephone to do most of your computing work, you’ve moved most of your operations to SaaS or are dabbling with any AI program. It will blow your mind.