Trending Articles
Trending Topics
Trending
This article is part of our Opinions section, where we invite industry professionals to share their views on the most pressing technology questions of our time.
In recent months, many large enterprises and mid-sized companies have been focused on modernising their software delivery. It’s been a journey full of cloud-native wins, cultural shifts, and lately, nonstop AI hype. You can’t open LinkedIn, Slack, or even a coffee machine without someone promising that an AI agent will do your job better than you ever could.
From analyst reports to vendor decks, we’re told AI will handle entire workflows, replace whole teams, and transform every part of the business. If you’re not already “agentizing” your operations, are you even trying?
The latest are AI agents, which, if press releases are to be believed, are almost as good as a true AI. They’re apparently able to string together all kinds of use cases, handling tasks autonomously without minimal input from us poor, slow-witted humans. Without AI agents, you’re giving up your business, pretty much, surrendering it to competitors who are making money hand over fist using these applications.
Then you actually use an AI agent, and, well, it all kinda falls apart.
I mean, don’t get me wrong, AI can do some pretty cool stuff. It’s already helping teams shave time off big tasks or outsource the boring stuff, freeing them up to focus on something more interesting. That said, I’ve never encountered one of the magic agents that can actually take over the work of a human. They’re just not there yet.
The way the industry is going, though, separating AI hype from reality may soon become a career path of its own — agentic or otherwise. Even with hard evidence from people on the ground, spokespeople from the lowliest marketer to CEOs won’t shut up about how wonderful AI is and all the things it can do.
Like tech-crazed carnival barkers, they’ll espouse the benefits of their AI agents, despite it being clear that they can’t do half the stuff they claim. They’ll jazz up their claims with meaningless buzzwords like seamless (a great favourite), claiming that AI agents will jumpstart my processes or fuel expansion or whatever other word they’ve decided to appropriate that day.
They’re not just pulling the wool over consumers’ eyes, either: the massive PR push has also made its way to investors who will be told wondrous tales of everything AI agents (or whatever other AI-driven tool they’re peddling) can do, with ROI and adoption predictions that make the eyes water.
These lies probably won’t sucker more savvy investors, but there are always going to be a few greedy enough to believe all this hot air, dropping millions into useless technology. That wouldn’t be too bad – far be it from me to tell you how to spend your money – but they’ll further amplify the cacophony of false promises being spread over the web, further clogging up your LinkedIn feed with white noise.
Take Builder.ai, for example. They pitched “AI-powered software development” and raised hundreds of millions from the likes of Microsoft. On paper, they automated app building through magical AI. In reality, it was dozens of underpaid offshore developers working behind the scenes — AI mostly in name.
Once the truth emerged, the company collapsed almost overnight. Investors lost big. So did startups and SMEs who bet their growth on what turned out to be human sweat wrapped in a sci-fi wrapper.
The saddest part? Until the fall, Builder.ai was everywhere — praised in media, hyped on stage, and hailed as “the future of software.” If you listened to the headlines, you’d think we could fire our engineers and replace them with a chatbot and a backlog.
This kind of hype reminds me mainly of the crypto craze a few years ago, where we were all told that digital tokens would replace fiat currency. Not sure if I’m the only one, but last time I bought my groceries using money issued by my central bank, not bits and bytes; that’s still making the rounds in dodgy exchanges, making some rich while impoverishing many others.
On a more serious note, this rampant speculation may end up fueling rivalries between the United States and China, as each believes that the other is developing some superior capability that they need to counter. One is rumoured to have AI weapons, so the other needs to develop countermeasures. Is there any evidence of these weapons? Does it matter?
It’s Superpower Chicken 2.0, though instead of nukes, it’s an AI that’s little more than an empty computer pumped up by sales departments. At a time when cooler heads should prevail, the only voices we’re hearing are those screaming for more and more of…nothing.
After reading hundreds of articles and testing dozens of AI tools, I’m convinced that within the next three to five years, AI will not replace software developers or the skilled analysts required to work alongside it. The tools are improving rapidly, but the narrative of full automation replacing engineering teams (or any other teams) is just wishful thinking wrapped in marketing spin.
What we need more than anything is an AI agent that can monitor and filter all the noise surrounding AI and can help us get rid of the marketing spin and let us focus on the actual good AI can do – and there’s plenty of that. If we’re going to have a future powered by AI, we need to get rid of the sales pitch, first.
