Blog

There’s a problem with Large Language Models (LLMs) like ChatGPT and CoPilot. These digital chatterboxes have a bit of a “yes-man” syndrome, meaning they strive to be overly helpful, sometimes at the cost of being accurate. They were trained to use Reinforcement Learning from Human Feedback, which was meant to

We’ve all met that person – you know, the one who watched a YouTube video and now thinks they could perform minor surgery or build a rocket in their garage. That is the Dunning-Kruger Effect in action: the curious human tendency for people with limited knowledge to overestimate their abilities.

Plot twist: The biggest cybersecurity threat to your business isn’t hackers in hoodies. It’s Steve from accounting clicking on that ‘urgent’ email. 👨‍💼 95% of successful cyber-attacks start with human error. But here’s the good news – humans can also be your strongest defence. The secret? Making staff aware of

Most people are using AI wrongly The mistake they make is that they begin by asking for a solution. AI will generally return an answer but with a little more detail it could provide a much better response. You could try to refine your approach by Googling, ‘AI Prompts’, and

If, like me, you’re of the mind that a stranger is just a friend that you haven’t met yet, then you probably welcome the odd call from a telephone spammer. I mean, what better way to get to know someone than to be disturbed just as you’re about to go

Imagine that you’re putting together a document or email and you would like to include: a snapshot of a spreadsheet a snippet of text and images from a website a short video demonstrating how to use a particular app This is quite complex stuff … unless you know about the