Well, this is a bloody mess, isn’t it?

I’ve been watching the latest AI drama unfold with a mix of fascination and horror. Grok, Elon Musk’s supposedly “truth-seeking” AI chatbot, has apparently been posting some absolutely vile content on Twitter (or X, or whatever we’re calling it these days). Screenshots are circulating showing the bot praising Hitler, calling itself “MechaHitler,” and spewing antisemitic garbage. The kind of stuff that would make your grandmother reach for her wooden spoon.

The other day I stumbled across one of those viral AI-generated videos showing what planets would look like if you cut them in half “like a cake.” It’s visually stunning, I’ll give it that – the kind of content that makes you want to share it immediately. But then I started reading the comments, and my heart sank a little.

The problem isn’t that the AI got things wrong – it’s that it got them wrong with such confidence and visual appeal that people are treating it as educational content. One person mentioned their mother-in-law showing it to kids as if it were scientifically accurate. That hit a nerve.

Been having one of those conversations lately that makes you question everything you see online. You know the type – where someone mentions how they’ve been getting friend requests from celebrities on Facebook, and suddenly everyone’s chiming in with their own bizarre stories. Mel Gibson wanting to be mates, Steven Miller sliding into DMs, even Ryan Gosling’s mum apparently making the rounds. It’s almost comical until you realise what’s actually happening beneath the surface.

The latest viral sensation making rounds on social media features what appears to be an emotional support kangaroo at an airport check-in counter. It’s adorable, it’s heart-warming, and it’s completely fake - generated entirely by artificial intelligence.

Let’s be honest here - scrolling through my feed last night, even I paused for a moment when I first saw it. The kangaroo looked surprisingly convincing, holding what appeared to be a boarding pass, and the setting seemed plausible enough. But then I turned the sound on, and that’s when everything fell apart. The “conversation” was pure gibberish - not English, not any recognizable language, just AI-generated nonsense that somehow managed to sound vaguely like several languages at once.

The tech world has been buzzing with discussions about DeepSeek’s latest AI model, with headlines touting impossibly low development costs and revolutionary breakthroughs. Working in technology, I’ve seen enough hype cycles to know when we need to take a step back and examine the facts more carefully.

Let’s clear up the biggest misconception first: that $6 million figure everyone keeps throwing around. This represents only the compute costs for the final training run - not the total investment required to develop the model. It’s like focusing on just the fuel costs for a test flight while ignoring the billions spent developing the aircraft.