Content Insider #964 – Deathstar

By Andy Marken – andy@markencom.com

“Look, I can’t get involved. I’ve got work to do. It’s not that I like the Empire; I hate it, but there’s nothing I can do about it right now … It’s all such a long way from here.” – Luke Skywalker, Star Wars: Episode IV – A New Hope, Lucasfilm.

We admit it … we use AI the way AI companies develop/developed their “products” …we use
key words/phrases to scrape the internet to get the information we need for the subject we’re writing about.

Most of it is crap, misleading which gets thrown away so we kill more bits/bytes.

Then we think, write.

Sometimes, three or four times.

Yeah, it’s a sloppy, messy process and AI execs/proponents will say it’s a foolish waste of time; but in the end, we can honest to gawd say … we did that.

So, we’re not totally against AI but we just don’t think everyone has thought this thing through before jumping into the deep end … or they simply don’t give a crap about the consequences.

We know, it’s the “hot, new” thing that’s going to make life easier, better for … everyone.

It’s not new. Alan Turing started fiddling with machine intelligence back in 1950.

Nvidia has been producing AI chips for years … updating, enhancing them, making them more and more powerful (and expensive).

Now that they’ve really taken off ,Jensen Huang (CEO) has been busy circling the globe spreading the gospel, breaking bread with government leaders in almost every country, buying chunks of AI tech companies and juggling the thankless job of determining how AI data center builder/owner get the chips they need and still stay within the myriad of governmental guidelines/regulations.

And yes, keeping the company ahead of the competition especially AMD and a number of Chinese “quick learners.”

But hey, it’s a big business because everyone is investing in it and the potential varies all over the map with Gartner projecting $1.5T will be spent on AI this year while Yahoo Finance estimates the market will be worth $758B this year.

Yeah, it’s anyone’s guess but the fact is 32 countries have or are building AI data centers with China and the US hosting the majority (90 percent) of the 11,800 specialized data centers.

The leading hyperscalers are Amazon, Microsoft and Google while China’s Alibaba and Tencent lag behind it’s not because of a reluctance to finance bigger, better data centers but because of they can’t get bleeding edge chips so they’re waiting patiently/impatiently.

Of course, Meta is also investing heavily in buying “experts and building out data centers for next generation technology so they can milk everything possible from the world’s data.”

These aren’t your run-of-the-mill data centers.  

They use tons of GPUs, TPUs to handle the massive processing, acres of high-speed, scalable storage and yes, they have plenty of room for growth because the data centers will only get bigger and bigger and …

Oh yes, while you think your personal ChatGTP assistant and image generators are cool to use, we need to note that AI data centers demand a ridiculous amount of power – 6-14 kW pr server, 30-100+kW per rack compared to an ordinary data center rack that consumes 5-15 kW.

Standard data centers use traditional air conditioning to manage the heat the hardware generates but an AI data center runs so fast and generates so much heat they can consume up to 550,000 gallons of water per day.

While a discussion with your chatbot will only use about 500ml of water, IDC projects that by 2028, AI will require 720B gallons of water annually in the US, probably a similar amount where you live.

Much of the data center consumption will come from local community water supplies, many of which are already strained.

Iceland is probably the best location for AI data centers – plenty of renewable energy and an ocean of distillable water at their doorstep.

But then, they don’t affect your AI carbon footprint and negative environmental impact unless the industry wants to change the way they bring AI to life and use it.

Of course, the existing power grid in places like the Americas, China, EU and MEA wasn’t built to deliver this level of power AI needs day-in and day-out.

That means power grids will simply collapse, or utilities will have to double, even triple their capacity to meet the accelerated demand, which means your home power will cost more as you subsidize the infrastructure costs.

Advanced grid management technologies will also have to be developed/implemented.

The good thing when it comes to power production is we have options.

Sane countries like Australia, China, EU, Britian, CANADA, MEA and others opt for sustainable solutions like solar/wind farms and other renewable energy facilities.  

Shortsighted countries know that climate change is really a hoax, so they opt for fossil fuel power production … don’t ask.

O.K., so there are a few side effects. Jeezz, but the benefits are so cool!

And, when the time comes, AI will clean up after itself.

After all, all of the positives like instant answers, write-ups, papers, contracts, scripts, etc. far outweigh the negatives, right?

There has been an IT statement/observation that has been around since Grace Hopper first debugged a computer (and perhaps even before).

Garbage in … Garbage out.

Yeah, AI develops its answers, work product by scouring everything that it can get its circuits on … documents, books, music, movies, Post-its, reports, reviews, artwork, pictures, smart/dumb/really stupid ideas, whatever.

It really doesn’t give a s**t.  Data is data and AI lives for/lives on data!

And it’s going to filter, sieve, sort, manipulate and spit out an answer/solution/finished product … BAM!

After doing all of that “work”, if it doesn’t have everything it needs it will simply fill in the blanks and holes with … more data.

So, what if it’s “a little” biased, discriminatory, slightly off the mark or even completely wrong; it was done quickly, neatly and perfectly in line with the question/challenge/objective you gave it and the data that was supplied.

AI is a powerful tool that can be used, misused, abused.  Give it all the data you want to give it (or think it might need), play nice with it and it will always come through for you.  

.    

But what if someone asks you how you arrived at the solution, recommendation, completed project?

Yeah, you’re s****ed!

With most human solutions, ideas and recommendations there’s a filtration process you go through to arrive at the finish line.

It makes it sorta difficult to say, “Yep, found the answer, done and done.” 

Unlike AI, Spock was only half Vulcan so he only occasionally/slightly suffered the foibles and shortcomings of humans. Most of the time, he relied on the logic of data.

Kirk relied on his critical thinking skills and data/inputs–often without thinking about it too much. It’s just that as folks increasingly rely on the power and work of LLMs, critical decisions have to be made that help the individual or organization profit from the action, enhance their position in the marketplace and meet the desired objectives/goals.

Or, to put it a different way … winning (at all costs).

Don’t like the solution, results presented?

Fix it in post!

What it delivers is a final product is based on other people’s research, work, and execution.

We’re not saying AI can’t filter all of the data/information/options and arrive at the right answer, but most breakthroughs and wow solutions/products came about as an ooppss.

In 1928 Dr. Alexander Fleming went on holiday and forgot to put some bacterial cultures away properly.

When he returned, he found mold had contaminated the cultures and killed the bacteria.

Viola penicillin was discovered reducing deaths from bacterial infections and improved the success of major medical procedures.

Oh, and if you over rely on AI …

Yeah, your brain will turn to mush! 

The less you use it in your area of expertise and the less critical thinking you do the brain will become lazy and won’t be as sharp as it was … back in the day.

Don’t believe us?

Let’s say you’re 30-40 and have a busy work/family life and haven’t done some of the things you used to do all the time when you were young.

Run out and jump on your kid’s bike or skateboard. You either look like a damn fool or fall on your behind.  

There’s a reason experts say use it or lose it and the mind is no different.

Go ahead … ask AI.

And if all of this isn’t enough, we won’t sugarcoat it … jobs will be lost across every profession you can think of.

The first thing to go will be the entry level jobs because you’ve got Chat GTP and your AI assistant so there’s no need to pay someone to learn the job from the ground up.

For example, in the film/show series industry there are positions like production assistant, interns, assistant editors, junior screenwriters, researchers and similar entry-level jobs which AI can easily handle without a lot of oversight or rework.

The gap will widen between highly skilled senior managers and lesser skilled workers.

The new bridge positions will probably require specialized skills middle managers (and some senior managers) may not be able to learn/develop.

Or, to put it bluntly, a lot of folks are going to be between a rock and a hard place!

And don’t forget, AI doesn’t have a conscience, doesn’t experience emotions, heck it doesn’t even think.

But don’t worry, AI will figure it out.

As Luke Skywalker said in Star Wars:  Episode IV – A New Hope, “if there’s a bright center to the universe, you’re on the planet that it’s farthest from.” Then, in thinking about his comment he added, “I have a very 

bad feeling about this.”

We don’t agree with C-3PO when he said, “We’re doomed.”

But we do have to carefully follow Ben Obi-Wan Kenobi’s advice, “Use the Force.”

And we darn well better use it wisely or …

Andy Marken – andy@markencom.com – is an author of more than 900 articles on management, marketing, communications, industry trends in media & entertainment, consumer electronics, software and applications. He is an internationally recognized marketing/communications consultant with a broad range of technical and industry expertise, especially in storage, storage management and film/video production fields. He also has an extended range of relationships with business, industry trade press, online media and industry analysts/consultants.