Google
 

Aardvark Daily

The world's longest-running online daily news and commentary publication, now in its 30th year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.

Content copyright © 1995 - 2025 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk



Please visit the sponsor!
Please visit the sponsor!

The drug deal that is AI

1 July 2026

I really can't believe that so many companies are walking into the AI trap with their eyes wide open.

While it's true that when used appropriately, AI can offer some degree of cost savings to many businesses, those savings are not always as great as touted and they come with a huge risk.

If you build key parts of your business's front or back-end support systems around a cloud-based AI agent, you become incredibly vulnerable to any issues that service may experience -- even issues that are beyond their control such as comms outages.

We've already seen what happens when the EFTPOS network fails for any reason -- heavily dependent retail operations such as supermarkets grind to a halt, with only cash transactions possible.

Now magnify that by a factor of ten or more...

How does a company cope if their AI-based telephone and online support systems simply stop working due to an outage? Anyone attempting to ring those businesses may simply get dead air.

I'm sure you're thinking "yeah, but these cloud-based services are pretty reliable, I mean, look at how seldom Google goes down".

You'd be right of course, however we live in a world where hackers make a pretty penny by targeting service providers and demanding ransoms to call off DDOS, ransomware and other attacks.

However, the big risk doesn't come from evil little script-kiddies trying to earn some pocket money by extorting cloud service providers.

The real risk comes from the AI companies themselves.

The cost of providing AI services is huge and getting greater by the day.

The countless billions of dollars that are being poured into AI datacentres has to be recouped somehow and ultimately, that money will come from users of those services.

We've already seen that virtually every vibe-coding service has hiked their prices by a huge amount recently, in an attempt to break even or spin a profit. The cost per token has skyrocketed and as these systems get more sophisticated, those tokens don't go nearly as far as they used to.

So here's the real danger...

A company tries AI, finds that it offers some real cost savings and commits to full-scale adoption in critical roles.

Life is good, costs are down, service levels are maintained or perhaps even improved.

Then the AI provider doubles or triples its prices.

Suddenly the economics of that AI service no longer stack up. Even worse, you've probably reduced costs by cutting back on staff numbers -- perhaps even skilled workers that are not easily replaced.

Another nightmare scenario is that the AI service you're using decides to upgrade to a new model and suddenly things start falling apart. Elements of the service that worked perfectly well yesterday are now error-prone or just don't work at all.

Either way, as a business, you're stuffed and there's very little you can do about it.

Recently, Ford learned this lesson the hard way and ended up having to hire back a bunch of workers it had replaced with AI systems.

The reality is that right now, AI companies are in their drug-dealer phase. They're giving away the first taste, to get you hooked and dependent on their product. Once you are dependent, the price will rise, and rise and rise. They know full-well that companies will likely just keep paying because the cost of reverting to their previous manual systems is too high.

There might even be a third scenario and that's where the AI bubble bursts and your chosen AI provider simply goes bust. This means that suddenly, the AI service on which your business is now highly dependent simply goes dark, leaving you out of business.

My advice would be that if you're a business looking to commit to using AI in critical roles, do it in-house rather than through a cloud provider. The medium to long-term costs will he much lower, as will the risks.

The next year or two are going to be very interesting in such an environment and I expect to see quite a bit of blood spilled along the sides of the AI superhighway.

Carpe Diem folks!

Please visit the sponsor!
Please visit the sponsor!

Here is a PERMANENT link to this column


Rank This Aardvark Page

 

Change Font

Sci-Tech headlines

 


Features:

The EZ Battery Reconditioning scam

Beware The Alternative Energy Scammers

The Great "Run Your Car On Water" Scam

 

Recent Columns

Theft, by any other name
When I was a spotty-faced teenager, I'd spend a fair chunk of my pocket-money on buying music...

EVs and climate change
We're told that EVs will help reduce the rate of climate change...

Slate delivers their EV
Okay, they're not actually delivering yet but they are taking pre-orders...

That's not how the internet works
If you live in the UK (and likely the rest of the world eventually) then get ready for more "shaping" of your online experience by the government...

Malfeasance
Yesterday I started wading through the hundreds of documents that were released by the South Waikato District Council...

Snake oil merchants galore
Modern technology can do some amazing things...

Pot pouri
Well we've made it through the shortest day and now summer is just around the corner...

Pushing back via tech
The UK and Australian governments are leading the Western world in rolling out online tracking and censorship...

Look out, I am armed and dangerous!
I really can't believe that the clowns at the South Waikato District Council are still at it -- but they are...

No science here, we are the media!
For quite a few years now, I've been suggesting that the power currently feeding the aluminium smelter run by Rio Tinto in the South Island could be better used...