The second to last week of October 2026 has been rough for the people tasked with keeping the internet ticking over. Two exploits were announced this week that allowed attackers to take control of servers, with millions of attempted attacks & websites/ services going offline as IT teams scrambled to protect their systems.

It would be easy to see this week as a one-off event, but it’s symptomatic of massive changes sweeping the internet & interconnected technology in general (which is most technology, these days).

IT Security as Home Security

So we don’t get lost in technical jargon, I’m going to use the analogy of home security.

As home security is familiar & tangible it is easy to understand that while a locked door makes your house more secure, not all locks are equal & there are plenty of other ways to get into a house. A 100-year-old lock will be far easier to break than a brand new lock, but for most people - a lock is a lock.

So each homeowner chooses their level of security, but in most cases we lock our doors with whatever lock the house came with & that’s pretty much that. A security minded homeowner might upgrade their locks, put security latches on their windows & even install a camera system or alarm.

The machines that serve you websites, online shopping & infinite scrolling videos also need security. Broadly, the level of security will be a combination of the level of security the machine was built with & whatever maintenance & upgrades have happened since then.

In both cases as the perceived threat grows, so does the applied level of security - for those people paying attention, at least.

So, What Changed?

Two inter-related technologies have combined to cause the current security headache - Artificial Intelligence & Mass Automation.

One thing that AI is very good at is cybersecurity. It’s cheap, tireless & will find ways to break into systems that no one has ever thought of. This is a double-edged sword because the “good guys” need to find & fix issues before the “bad guys” find & exploit them - & both sides have access to AI.

Using our home security analogy, imagine a criminal gang tasking an AI bot to work on better house breaking & it comes up with a way to break into houses through a popular model of skylight using a step ladder, coat hanger & bottle of lemon juice.

That doesn’t sound too scary - the criminal gang would need to know which houses had that model of skylights & then set about breaking into them one-by-one. Which leads us to…

Mass automation turbocharges these novel exploits. Back online, millions of criminal-controlled machines - from servers to home computers to smart TVs - are currently connected to the internet. These machines can be used for probing for security weaknesses & carrying out the attacks themselves.

When the “bad guys” get there first it looks something like this:

1. AI bot finds novel way to break in.
2. Attackers already have information telling them where this will work.
3. The novel break-in can be run simultaneously across thousands to millions of locations.

Everything, everywhere, all at once.

What Do We Do?

Allow me to briefly talk about cheese.

In the late 1980s James Reason introduced the world to the concept of the “Swiss cheese model”. This envisions a system like a block of sliced Swiss cheese - complete with its famous holes. Arrange the slices in one way & you can look right through the block of cheese, rearrange the slices & very little can get through.

The concept is that a single slice won’t stop everything, but multiple slices properly arranged will. We apply the same thinking to layers of security.

For a house this could be multiple locks, a door chain & an alarm, for computers this is an increasingly bamboozling stack of technologies tasked with detecting, deflecting & mitigating attacks. Bypass one layer & there are plenty more layers to stop you.

The challenge for technology is that a lot of what is currently connected to the internet is poorly protected with minimal elderly layers of security which attackers already know how to breach.

UpShift’s Approach to Security

First up I should say that if you’re running a small website then I’d strongly recommend you move it to a managed platform like Squarespace, Framer or Webflow. These platforms have centrally managed security - teams of engineers working around the clock to identify & mitigate risks. If you are using WordPress I’ve written about why you probably shouldn’t.

Likewise, many e-commerce websites can be run on platforms like Shopify, which also take a centrally managed approach.

Medium-to-large websites or web apps often have requirements that can’t be met by centrally managed website platforms. For our clients we take a managed approach where client sites & applications are managed, monitored & maintained.

We’ve built our servers to meet an internationally recognised security framework - CIS Controls IG2 - & install layers of cheese security to detect, deflect & mitigate attacks.

The reality is no one can guarantee complete security. AI & mass automation have significantly increased the risks to IT systems & if you’re not sure about the security of your systems, now is the best time to review them.

Talk with whoever manages your websites/ IT & make sure - at the bare minimum - they are monitoring & patching your systems.

We're Here to Help

UpShift can help with replatforming/ securing PHP websites & applications built on frameworks like Drupal, Joomla, Silverstripe & Laravel. We can also move e-commerce websites to Shopify. If that sounds like you, please get in touch - we’re happy to review your situation.