The spreadsheets are still stuck. For two years, we’ve been told that silicon is about to liberate us from the drudgery of the nine-to-five. The pitch was simple: flip a switch, pay a subscription fee, and watch the output soar. It’s a nice story. But if you look at the actual data—the cold, hard GDP and productivity numbers—the needle hasn't moved. In fact, it's barely twitched.

We’re currently living through the "Great Prompting." Every C-suite executive from Seattle to Seoul is obsessed with the idea that LLMs will shave 40% off the work week. They’re buying enterprise licenses for Microsoft 365 Copilot at $30 a month per head, hoping for a miracle. That’s a massive tax on a workforce that is still trying to figure out if the bot is actually helping or just making more mess to clean up.

It’s the "babysitting" problem. You ask a chatbot to draft a memo. It spits out three paragraphs of high-fructose prose in four seconds. Impressive. But then you spend twenty minutes checking the facts, stripping out the weirdly formal jargon, and making sure it didn't accidentally hallucinate a new company policy. You haven't saved time. You’ve just shifted your labor from "creator" to "unpaid editor for a machine that doesn't understand context."

The friction is everywhere. Take the coding world. Sure, GitHub Copilot can autocomplete a function faster than a human can type "Hello World." But senior engineers are now reporting a "code rot" problem. The bots are churning out massive volumes of boilerplate that’s technically functional but architecturally sound as a sandcastle. We’re producing more code than ever, but the quality is diving. We’re building technical debt at the speed of light.

Then there’s the hardware bill. NVIDIA is printing money because everyone is terrified of being left behind. We’re pouring billions into data centers that consume enough electricity to power a small nation, all so a middle manager can get a slightly better summary of a meeting he didn’t attend anyway. The return on investment isn't just invisible; it’s currently a hole in the ground.

Tech evangelists love to talk about the "J-curve." They say productivity always dips before it spikes because it takes time to reorganize how we work. They point to the steam engine or the electric motor. It took decades for factories to stop lining up machines in a row and start placing them based on the flow of work. Maybe. But those were physical tools that did physical things. AI is a cognitive layer that mostly generates more noise in an already noisy digital world.

We’ve reached a point of diminishing returns on information. We have more "content" than we can consume, more emails than we can read, and more Slack messages than any sane human can process. Adding a button that generates even more words doesn’t solve the bottleneck. The bottleneck is us. Our brains don't have an API update.

There’s a specific kind of exhaustion that comes with this era. It’s the feeling of running a race where the finish line keeps moving because everyone else is also using a motorized scooter. If everyone uses AI to write their emails, and everyone uses AI to summarize those emails, we’re just two bots talking to each other while we sit in the middle, paying for the electricity.

Silicon Valley keeps promising we’re on the verge of a post-scarcity utopia. They say the "agents" are coming to handle our calendars and buy our groceries. But right now, the most tangible result of the AI boom is a massive increase in the cost of enterprise software and a suspicious amount of extra "vibes" in the quarterly reports.

Companies are caught in a prisoner’s dilemma. They can’t afford not to buy the tools, but the tools aren't actually making them more profitable yet. They’re just making them more automated in their mediocrity. We’re spending hundreds of billions of dollars to see if we can replace a $20-an-hour copywriter with a $40,000 server that occasionally insists that 1+1 is 3.

Maybe the boom is coming. Maybe in five years, we’ll look back and wonder how we ever lived without these digital ghosts. But for now, the revolution looks a lot like a very expensive spell-checker.

Are we actually getting more done, or are we just getting better at pretending to be busy?