Stepping into the AI era feels strangely familiar, another shift, another cycle, another wave of excitement that’s as unsettling as it is intriguing. We’ve just lived through Blockchain, Cloud, and everything in between, and now AI arrives with the confidence of a newcomer at a party.

AI/Agentic AI is loud, charming and little too impressive due to all the marketing charms but has found a way. Beneath the noise, something real is happening. Whether we deny it or try to frame it, this technology is slipping into daily life. One day we will wake up and realize we can’t imagine the world without it, as it slowly appears on smartphones and every device in-between. Before thinking this is yet another bubble, consider the following unfolding in the background:

1. Emerging protocols like MCP and A2A, are quietly rewriting how models and agents speak to one another.
2. NPUs are landing on our phones and laptops, bringing privacy, on-device reasoning, and strange new efficiencies.
3. TPUs are training new models faster than platforms can release them and tech giants are rushing racing to beat the odds., whilst Nvidia is collecting its winnings.
4. And of course, users are suffering from what I call choice apnea. So many benchmarks, so many trade-offs when it comes to choosing a model, that most of us simply test a few and settle on what works that week.

And then there the marketing charms, the pixie dust as it may. Marketing has grown immensely alongside the tech the past few years. Digital products are not simply enough on their own. Linus Torvalds wasn’t wrong when he said AI is “90% marketing and 10% reality.” Companies slap “AI” on old services the way people added “.com” in the early 2000s. The question isn’t whether AI creates value, it clearly does, but where that value ultimately lands and is tangible is another topic of discussion. One needs to wonder whether Big Tech can stomach the shock if the hype cools? For now, the giants are paying the bill. And yes, they can.

Look, I’m naturally more conservative, low-risk, slow-adopting, measured. So, part of me wants to step back and breathe. For years, machine learning meant curated datasets, tedious cleaning, and models built for narrow, specific use cases. That hasn’t vanished. But Big Tech has now widened the aperture. We’ve moved from small, single-purpose systems to large general-purpose models, LLMs that echo the shift from local servers to the cloud and just like the cloud didn’t erase local servers entirely, LLMs won’t erase traditional ML.

So where does that leave us?

AI will continue to elevate capabilities. It may be that we’ve been standing on dormant potential for years, waiting for the right spark. I’m not pretending to predict the future. There is hype. There are real limitations, but we need to remember new tech always brings new problems, and those problems always create the next wave of opportunities. I prefer to stay in that space, the space of solutions, of building, of steady innovation.

A new era is certainly here. A new era on top of the information tech era. Let’s anchor ourselves in the human condition.

If history has taught us anything, it’s this: new tech we will always make new challenges and we will always find new ways to solve them