You Probably Think You Know AI

The meteoric rise of artificial intelligence in recent years has everyone—from tech moguls to your tech-illiterate neighbor—chattering about its potential. It's being heralded as the savior of human efficiency, a tool that could eliminate drudgery and usher in a new era of prosperity. But equally, it's cast as a villain that threatens the livelihood of artists, and in some circles, it's even seen as the harbinger of human extinction. Amidst all this noise, it’s easy to think you know AI. But do you really?

Welcome to the Circus

In the AI space, buzzwords flow like free champagne at a networking event: "machine learning," "neural networks," "natural language processing," "LSTMs (internal joke between my friends, don't worry about it)." Each term is weaponized to get the attention of media outlets, VC investors, and, of course, corporate decision-makers looking to sprinkle some AI magic dust onto their companies.

But for those of us who’ve spent time inside this so-called AI revolution, the reality is often more farcical than fantastic. Many of the so-called "AI breakthroughs" are either rebranded ideas from the '70s or are simply not applicable outside of narrow academic settings. As someone who’s seen the belly of the beast firsthand, let me share a secret: most "AI initiatives" are just expensive make-believe.

The Glory Days of the Grift

Back in 2022, I entered the AI space wide-eyed and hopeful, like a kid entering Willy Wonka’s Chocolate Factory. I was captivated by the allure of working on breakthrough technologies that could redefine entire industries. But as I dug deeper, the sheen of enthusiasm wore off, revealing the rusted, crumbling machinery underneath. It became painfully clear that the field was often less about creating real innovation and more about inflating headcounts, securing promotions, and getting "thought leader" status on LinkedIn.

The disparity between what AI was supposed to be and what it actually is became glaringly obvious. Leaders who hadn’t spent more than thirty minutes reading a Wikipedia article were proposing multimillion-dollar AI initiatives with zero practical value. As a result, the number of companies boasting about their AI initiatives far outstripped the number of companies doing anything meaningful with it.

Despite that, I haven’t fully "packed my bags" and left. I still work in AI, but with a more cautious approach. It’s less about blindly following the hype and more about identifying genuine use cases—solving problems that actually need solving.

When AI Becomes the Driver, Not the Passenger

And then ChatGPT happened, and everything changed. Overnight, every developer with an internet connection and a few ChatGPT prompts suddenly became an “AI expert.” Codebases started to fill up with GPT-generated snippets—often implemented from start to finish with little human oversight. It became a bit like trusting a toddler to build your house using LEGO blocks. The problem is that while ChatGPT’s code might get things working most of the time, it doesn’t guarantee the use of the latest or best practices. Developers now scramble to patch up bugs and security issues introduced by these seemingly helpful AI-generated solutions.

The real irony? While ChatGPT can generate code at breakneck speed, it often regurgitates out-of-date methods and practices. Developers who rely on it too heavily without understanding the nuances of the technology are inadvertently creating code that's not only inefficient but also dangerous in production environments. It’s a house of cards waiting to collapse, and no amount of breathless blog posts or social media hype can prop it up for long.

That’s why I stick to building solid systems. Sure, I’ll use AI tools as assistants, but I’ll never let them drive the car. In the end, the code and architecture should reflect my expertise, not a chatbot’s latest educated guess.

Data Science: The First to Fall

When the data science hype began to wane, the job market adjusted. The shiny promises of machine learning models solving all business problems faded, and people who once lived in data science glory found themselves staring at evaporating job opportunities. The grifters, who had skillfully ridden the AI wave, simply moved on to the next fad, leaving behind the true practitioners who suddenly found themselves looking at a much more competitive job market.

The reason? Data science was—and still is—hard. There’s a steep learning curve, a requirement for genuine expertise, and a demand for domain-specific knowledge that can’t be faked or Googled. In essence, the people who genuinely cared and were good at their jobs were left to clean up the mess.

Then Came GPT-3, GPT-4, and the Death of Nuance

Just when you thought things couldn’t get more ridiculous, along came ChatGPT and its more evolved sibling, GPT-4. This opened Pandora’s box. We now live in a world where a chatbot is treated as the crown jewel of technological innovation. Forget sustainable, meaningful progress. Now everyone’s on a mad dash to slap GPT-based conversational support onto every conceivable app and website.

The fallout? We've re-entered the era of absurd promises. Those who couldn’t understand AI's limitations before are now under the delusion that LLMs (large language models) will soon be running every facet of business and life.

AI: Still Just Another Tool, but with a Better Marketing Team

AI is powerful. There's no denying it. But it's still just a tool—a tool that’s particularly effective at certain tasks like pattern recognition, data generation, and language understanding, but not much else. It’s no omnipotent entity that will automate our world and bring about utopia. It’s as flawed and fallible as its creators. Yet the hype machine barrels forward, powered by the hopes of execs and influencers who see it as the golden ticket to instant relevance and profitability.

So, the next time you hear someone proclaiming that AI will save the world or, conversely, doom us all—just remember that they’re probably just following the winds of the latest trend.

Conclusion: Chill, Dude. It’s Just an Overhyped Tool.

If you really think you know AI, step back for a moment. Consider what’s actually happening. AI isn’t some messiah, and it’s not the end of the world either. It’s a tool with certain capabilities and a ton of marketing hype behind it. Yes, it has implications—both good and bad. But if history is any guide, the people who will suffer most are those who bought into the hype without understanding its limitations.

So, whether you’re a true believer or a skeptic, remember this: It’s just another chapter in the ongoing book of technology hype cycles. In the end, we’ll probably find ourselves muttering in the cold night air:

“Just use Postgres, you nerd.”