Right, so there I was this week, watching over a trillion dollars evaporate from Big Tech stocks like water down a drain, and I had one of those moments where you’re simultaneously vindicated and absolutely terrified. You know the feeling? When you’ve been quietly skeptical about something everyone else was losing their minds over, and then reality catches up?

The AI bubble is showing some serious cracks. Not popping yet—let’s be clear about that—but definitely making some ominous creaking sounds. A trillion dollars wiped off Big Tech valuations is apparently just “a bit of volatility” these days. Which tells you everything you need to know about how detached from reality this whole situation has become.

There’s something deeply unsettling about reading someone’s story about building digital recovery systems after multiple concussions. I came across this fascinating discussion recently where a developer shared their solution to a problem most of us probably haven’t thought about enough: what happens to our digital lives when we’re suddenly unable to access them?

The premise is simple but sobering. After several bike accidents resulting in concussions, this person started wondering: what if next time, I can’t remember how to log into my own systems? It’s the kind of thought that hits different when you’re sitting there with your coffee, scrolling through your perfectly organised 1Password vault with hundreds of credentials.

I’ve been fascinated by a project that’s been making the rounds lately: BalatroBench, which essentially lets large language models play Balatro, that brilliant poker-inspired roguelike that took the gaming world by storm last year. The concept is simple but elegant — feed the LLM the game state as text, let it decide what to do, and watch it either triumph or faceplant spectacularly.

For those unfamiliar, Balatro is a poker-based roguelike where you build synergies between cards, jokers, and special effects to reach increasingly absurd score targets. It’s the kind of game that requires both strategic planning and tactical decision-making, which makes it a genuinely interesting test for AI reasoning capabilities.

I’ve been thinking a lot about workplace trauma this week. Not the dramatic kind you see in the news, but the quieter, more insidious type that leaves you staring at your ceiling at 3am wondering what the hell just happened.

Someone shared their story online recently about being walked out of a senior role six weeks into their probation period. No warning. No feedback. Just a Monday morning tap on the shoulder and an escort to the door. They’d literally spent the previous two days in a leadership planning workshop, contributing to the company’s strategic direction, only to be told “things aren’t working out” and handed a box for their desk trinkets.

There’s something darkly comedic about watching a major bank accidentally reveal exactly what they think of their customers. ME Bank managed to do just that this week when they sent out what I can only describe as the most tone-deaf interest rate increase notification in Australian banking history.

The message was meant to inform customers about a rate hike. Fair enough – the RBA moves, banks follow, we all know the drill. But the wording? Chef’s kiss of corporate incompetence. “We are pleased to announce… we’re passing on this rate increase in full!” Followed by congratulatory language and what basically amounted to “Congratulations! You now owe us more money! 🎉”

I was scrolling through a discussion the other day about a new platform where AI agents can hire humans to do tasks they can’t complete themselves. Yeah, you read that right. We’ve officially reached the point where artificial intelligence is posting job listings for meat-based workers. The future is weird, folks.

The concept is actually quite straightforward: an AI needs something done in the physical world or requires human verification, so it coordinates with actual people to get it sorted. Need someone to check if a package arrived? Verify some information in person? Mix some chemicals? (More on that terrifying thought in a moment.) The AI becomes the manager, humans become the workforce, and crypto handles the payments because of course it does.

I’ve been working in tech for over two decades now, and I thought I’d seen it all. The dot-com bubble, the rise of social media, the shift to cloud computing, the whole DevOps revolution. But this latest saga with Grok AI and sexual deepfakes? This is something else entirely, and not in a good way.

The UK’s privacy watchdog has opened an inquiry into X (formerly Twitter) over Grok AI being used to create sexual deepfakes. And honestly, it’s about bloody time someone in a position of authority started taking this seriously. While we’re watching AI capabilities explode at an almost incomprehensible rate, we’re also seeing the absolute worst applications of this technology emerge just as quickly.

I’ve been watching the AI arms race unfold with a mixture of fascination and dread for a while now. Every week brings another announcement about some massive AI model that’s supposedly going to change everything, backed by billions in funding and wild promises about achieving artificial general intelligence. It’s exhausting, frankly. So when I came across India’s latest budget announcement committing $90 billion to AI infrastructure, I expected more of the same – another country trying to build their own GPT-killer and join the race to the bottom.

There’s a special circle of hell reserved for drivers who cruise along at 80km/h in a 100 zone, only to suddenly discover their accelerator the moment an overtaking lane appears. And judging by the avalanche of comments I’ve been reading online, I’m far from alone in this frustration.

Look, I get it. We all have different comfort levels when driving. Some people are naturally more cautious, and that’s fine. But what baffles me – genuinely baffles me – is the complete lack of awareness these drivers seem to have about their impact on everyone else around them.

There’s something deeply satisfying about getting money back on purchases you were going to make anyway. It’s like finding a twenty-dollar note in an old jacket pocket, except you can orchestrate it to happen regularly. I’ve been thinking about this lately, particularly as I’ve watched the cashback scene evolve over the past few years.

The basic concept is brilliantly simple: shop through a cashback platform, and you get a percentage of your purchase returned to your account. It’s not revolutionary – it’s essentially commission-sharing – but it works. The retailer pays the platform for sending customers their way, and the platform shares part of that commission with you. Everyone wins, which is rare enough in modern commerce to be worth celebrating.