Aardvark Daily

Please visit the sponsor!

The real reason behind the AI cautions

AI companies are calling for a pause in development and the implementation of stronger controls on their industry.

Why would they do that?

Surely they should be in favour of less control and a faster pace of innovation so as to become leaders in their field, right?

Most recently we've seen warnings that unfettered AI development poses an existential threat to the whole of mankind.

So what's *really* going on here?

There are several possible scenarios that have caused the industry itself to call for the brakes to be applied. Let's evaluate them.

The first one is the one they want us to believe. This is that AI is progressing so rapidly that we risk losing control, especially now that recursive self improvement (RSI) is a thing. RSI, they tell us, has the potential to create improvements at an exponential rate due to the positive feedback loop that is created. There is a risk that we might simply lose control once a threshold is crossed and have no way to claw back that control.

Really?

While it's true that RSI could rapidly facilitate the capabilities of AI, we need to remember that ultimately we can always pull the plug on the hardware every AI system relies on for its existence. We are not going to wake up one morning and find that some super-AI has become sentient and that the assets of the military are now being operated by humanoid cyborgs intent on our destruction.

At worst, we might get the situation where internet-connected infrastructure is disabled by a rogue AI but there'd be no way for that AI to stop us simply shutting down the power and effectively turning it off.

Okay, so if that's just a smokescreen, why would AI companies want governments to put the brakes on things?

One possibility is that they realise they're reaching a point of diminishing returns. Things aren't progressing as quickly as they'd hoped and with all the money being invested in AI datacentres and the like there's the very real risk of the bubble bursting soon. By having governments slam on the brakes, they can keep hyping the potential of AI while blaming governments for not seeing that realised yet. "Just keep funding us and we'll get there despite government intervention" would be the claims they'll make to investors.

This strategy would likely allow the investors to IPO their companies and slip out the back door with wads of cash before the reality of the situation crushes stock valuations and financial shockwaves echo across the world.

However, I think I know the real reason:

Right now anyone can go online and download an AI agent or LLM and run it on their own computer -- for free. No charge. Zero cost.

As regular readers will know, I've already done this and my home network has its own AI LLM which has been quietly humming away on a lowly Raspberry Pi 5 for over a month now. It just works.

Okay, it's no ChatGPT or Claude but it provides 90 percent of all the AI capacity I need on an average day. It'll summarize long documents, help me research a specific topic, check my logic and even cut code snippets when I'm writing software.

I have friends around the world who have built far more capable systems using regular PCs and GPUs. They have generative AI systems that produce images and even video -- without having to spend a single red cent on AI tokens.

It's my belief that the AI big-boys have suddenly woken up to the fact that the ability for companies and individuals to build their own perfectly adequate local AI systems is going to decimate their business model. They can see that nobody is going to potentially huge sums to them for cloud-based AI if a one-off cost and a free software download will deliver sufficient capabilities in-house.

What better way for big-tech AI to strangle this looming threat to their profitability than to convince governments that AI is dangerous? Tell the wooden-headed politicians of the world that AI poses an existential threat to all of mankind and demonstrate its abilities by showcasing examples where it's hacked other computer systems and infrastructure.

You know that the idiots in power will go into "knee-jerk" mode and roll out huge restrictions on the premise that they're saving mankind from a hideous threat "won't someone think of the children?"

I'm sure those same big-tech AI companies will say "you don't have to worry about us, we have highly skilled people ensuring that *our* AI systems are safe -- but you need to make sure this dangerous tech isn't in private hands".

The inevitable response will be that AI will become regulated and restricted. No more free downloads and the posession of an unauthorised local AI system will be a crime akin to owning an arsenal of machine guns.

If I had to place a bet, this would be where I'd put my money. Big tech never thought for a moment that regular people and businesses would be able to run a useful AI model on desktop hardware but now that it's happening they're scrambling to protect their revenues from such things.

Watch this space.

We all know that the western world is entering an era of neofascism so governments will leap at the chance to regulate or ban something else, especially if they can use a trumped-up excuse such as saving us from SkyNet.

Carpe Diem folks!

Please visit the sponsor!

Here is a PERMANENT link to this column

Rank This Aardvark Page

Change Font