AI Is a Lot of Tools without an Infrastructure
Content Insider #916 – So Innocent
By Andy Marken – andy@markencom .com
“This is why you’re drawn to conflict without hesitation. It’s part of your training. You are not just a warrior, Alita. You’re an URM Berserker. The most advanced cyborg weapon ever created.” – Dr. Dyson Ido, Alita: Battle Angel, 20th Century Studios, 2019
The other day, we wrote something to post to one of the social media sites – a rarity for us – and as we finished, this little button appeared and asked (nicely), “Now do you want us to write it right for you?”
Okay, it didn’t really say that but that was the gist of the question.
Geez dude, if that wasn’t what we wanted to write, we would have written something else.
Nice thing was it didn’t feel insulted or hurt because as much as people increasingly interact with and get close to AI, especially digital assistants, AI has zero feelings.
It was a completely unemotional question; and if we had said okay rewrite it, the results may have been something totally different or perhaps simply expressed better.
But it wouldn’t have said, “Damn, that’s good stuff!”
More importantly, it wouldn’t be ours!
But AI simply sits there on the system and waits for the next chance to look over our shoulder and say let me rewrite … something.
As much as these autonomous agents, digital coworkers, copilots and customer support respondents do something, they don’t have a freakin’ clue.
You may count on them in a pinch, but they are…clueless!
Renowned author Ernest Hemingway said it best, “Writing is something that you can never do as well as it can be done. It is a perpetual challenge … it makes me happy when I do it well.”
We don’t even want to compare our thought process and writing with Hemingway other than it’s our thinking and working through to the end.
Why is that so important?
As messy as the process is, as frustrating as it is and as time-consuming as it is; there’s a helluva lot of satisfaction when the project is done.
Yeah, you came out the other side and survived.
You can look at it and say … I did that!
We’re not being anti-AI, we just think there are times that you need to think/act for yourself and live with the consequences – good or bad.
We have no problem going to one of our streaming services and simply asking them for something to watch.
We’re not dumb, we know they’ve all been tracking what we watch, when we watch it, how long we watch it and on the screen we use.
The same is true with our kids who live on social media – TikTok, Snapchat, Instagram, YouTube – posting and watching stuff.
So, when we go to Netflix or Disney+ or Tubi, bam they offer up a few movies/shows based on what we’ve been watching so there’s zero frustration, just relaxing entertainment.
When our daughter goes to TikTok or one of her other favs, she gets a couple of female things she’s probably interested in.
We’re both happy.
As for the technology? It doesn’t feel a d*** thing!
It’s sterile … emotionless.
When we go online to gather info/ideas for an upcoming Insider, Google serves up topic lists as well as its AI Overview on the topics that may be right, sorta right or WTF.
Yeah, hallucinations crop up everywhere that seem so right, so perfect they can’t possibly be true.
Our kids have bumped into them as well, especially deepfake photos and videos.
Deepfakes can range from bad to dumb to obscene.
They were rampant last year when 70 countries around the globe held elections when politicians and celebrities were raised from the dead to endorse causes and individuals.
People spent as much time denouncing the deepfakes as they did promoting what they would do when elected.
Half the time, we don’t even believe the stuff we really want to believe.
As one techie enthusiast said, AI makes the impossible feel possible.
Yeah … right!
Volumes – The volume of data/information created, captured, copied and stored will continue to grow exponentially in the coming years and we have to learn how to intelligently use it.
But the fact is we live in an information/data-driven world.
According to industry analysts, we created, copied and used more than 149 zettabytes of data; and over the next five years, that volume will grow to more than 394 zettabytes.
If you wonder about making the right decisions all of the time, know that it’s almost impossible to do it on your own.
Rather than putting in the time, effort and pain of working through it all, the AI providers have given us a wonderful new partner.
We’re rushing to use it … everywhere.
We’re already seeing it widely used in social media and streaming analytics.
It is increasingly used in content creation, production and delivery with writing room assistants, story boarding, scene detection, shot evaluation/selection, color correction, motion tracking, VFX/CGI, audio enhancement/editing and more.
It’s here. It’s going to stay. We’re human and we’re better/more creative than a bunch of 1s and 0s.
Our favorite application is the increased use of audio translation tools.
They’ve made it cost effective for producers/distributors to translate the content from the native language to something we can understand, including perfect to almost perfect lip sync or subtitles.
Either way it has opened the door for people to enjoy and appreciate video stories from around the world so that maybe – just maybe – we understand each other better.
And at least for the time being, humans still need to double-check the accuracy of the translation.
Okay that’s entertainment; but organizations are also rushing to implement the technology in their businesses and maybe it can have similar results.
Businesses are deploying the technology in call centers to respond to queries quickly and accurately as well as using it to enhance their marketing efforts to evaluate product and marketing message alternatives to produce the best sales at the lowest possible cost.
It’s also being used to optimize manufacturing processes and predict when systems may fail so they can be upgraded before failures cause costly production delays.
The driving force behind the growing use of AI is to increase efficiency, develop new business opportunities, improve quality, address skills shortages and augment today’s and tomorrow’s workforce.
Growing Shortage – In the coming years, the percentage of our global population that will carry the brunt of the workload will slow while older generations increase, placing more productivity pressure on the young.
The increased use of AI in all of these areas is going to be important because the world’s population is ageing (65+) faster than younger men/women (18-50) enter the workforce.
As noted in the chart above, the only countries that will have younger populations will be in Africa.
Of course, Elon Musk is doing his part to develop the workforce of tomorrow with 14 — and counting–offspring; but one horny person won’t make a big difference.
Okay, forget the sarcastic note.
The fact is we need to implement AI in almost every area and do it intelligently rather than blindly, using it anywhere and everywhere.
We can’t simply turn over all of our thinking over to AI.
We were reminded of the importance of those dangers after reading a new study carried out by Carnegie Mellon University and Microsoft. They warned that over-reliance on AI could erode people’s critical thinking skills … if you don’t use it, you lose it.
That concern isn’t an idle sci-fi scare tactic.
But it reminded us that we put aside our writing in long hand when we began using a keyboard and now we have difficulty deciphering what we’ve written when we put pen to paper.
The same is true for dealing with numbers. Why add, subtract, multiply and divide when we have a calculator on our phone?
In China, where drawing traditional characters by hand (calligraphy) is as much an art as a way of communicating, students have been encouraged to practice the technique, so it isn’t lost to history.
Putting aside our ability to think and reason could/would be even worse.
A big problem with the glazed eyes love of AI that the cheerleaders have is that they don’t know how the technology is going to act/react once it interacts with itself and decides what needs to be done.
Answering a support call and giving application advice is one thing but trusting a loved one or yourself to a robot surgeon is a different matter.
Delivering the right answer 99.99999 percent of the time is okay, but the surgery?
***** you!
The problem with AI is that it’s like mercury.
Governments and companies are trying to establish guidelines and guardrails but at the rate the technology is being developed/released, that’s impossible to do.
Ask any AI developer or CEO what their product is going to do, and they’ll proudly proclaim:
- It’s going to help you make better movies
- It’s going to help you manage your finances better
- It’s going to make driving from point A to point B more enjoyable
- It’s going to cure today’s incurable diseases
- It’s going to give everyone a better quality of life
It’s going to serve mankind but it’s up to people to ensure it doesn’t serve mankind!
All of this fantastic new technology may be great … someday.
The problem is we don’t have the infrastructure in place to support/take advantage of it.
Jensen Huang, CEO of Nvidia, has spent the past year going from country to country explaining to officials that they need a strong underpinning of advanced data centers to develop and manage tomorrow’s AI.
O.K., so he has a vested interest because his organization clearly has the GPU leadership position today and he’d love to see them build those powerful centers on top of their processors. That includes the powerful quantum systems that can deliver the processing power for advanced, complex problem solving.
The problem is we only have 11,800 data centers around the globe and to deliver all of the processing power that will be needed to deliver the LLMs (large language models) performance that is needed, Huang has suggested the computational power for sustaining AI will roughly double every 100 days.
If you think that’s a tough hurdle to get over, chew on this.
Today’s data centers use roughly two percent of the globe’s electrical production.
But don’t worry about GPUs…Nvidia and others will make more even if the tariffs are “a little” steep for you.
The power needed for data centers to process AI tasks is already increasing 26 – 36 percent annually.
To achieve a tenfold improvement in AI processing efficiency, data center power requirements will increase an estimated 10,000 times.
Solar, wind and even coal/oil power production can’t meet the demand.
Governments are once again dusting off the plans for more clean energy nuclear power production which were shelved back in the early ‘80s.
So maybe it’s time to revisit the high volume clean-energy technology again.
Oh sure, we had the Three-Mile Island accident followed by the Chernobyl and most recently the Fukushima Daiichi disaster; but if we are going produce the clean power we need to take advantage of all of AI’s capabilities/potential, we’re going to have to address the infrastructure needs now.
Of course, we can only hope we learned from our mistakes on the last go around.
But hey … what’s the worst that could happen?
Scalability of quantum computing may be achieved through some early projects being conducted called quantum teleportation.
Sounds cool but is probably years away and AI techies want their power and performance … now!
No matter what the cost.
As Dr. Dyson Ido said in Alita: Battle Angel, “You’ve been given a chance to start over with a clean slate.”
But the way the industry is rushing this stuff out right now, we tend to agree with Alita, “I do not stand by in the presence of evil!”
Maybe we should be a little more cautious because it may be impossible to correct errors … later.
Andy Marken – andy@markencom.com – is an author of more than 800 articles on management, marketing, communications, industry trends in media & entertainment, consumer electronics, software and applications. Internationally recognized marketing/communications consultant with a broad range of technical and industry expertise, especially in storage, storage management and film/video production fields. Extended range of relationships with business, industry trade press, online media and industry analysts/consultants