I remember sitting in a TV newsroom in 1992 telling a colleague the internet was going to eat us alive. I was right, just not for a while. This midweek piece is a bit different. I’m thinking out loud. No easy answers, but a few patterns worth watching.

I’m not your investment adviser, but something odd is happening in public and private markets that will touch all of us eventually.

The sensible path for most people has been clear for years. Do not try to pick single stocks, you are not competing with full-time analysts. Diversify, buy the whole market, go with ETFs. Right now, that strategy looks genius because an S&P 500 tracker has been on fire.

Here is the twist. As of writing, the top ten companies in the S&P 500 make up roughly 40 percent of the index. That concentration is a risk in itself. And what do those top ten have in common? They are deeply tied to AI, one way or another. The market is making a very large bet that AI will reinvent how companies operate and spin off huge profits to the winners.

But what if it does not, or at least not for years?

There is a scratch-my-back dynamic in play. Companies with real revenue, like Nvidia, Oracle and Microsoft, are backing companies with little revenue, but with forecasts that look heroic. Nvidia invests in OpenAI. OpenAI commits to buying Nvidia chips. Oracle announces eye-watering future growth that drives the share price higher, then you read the fine print and see it is based on OpenAI booking cloud capacity with revenue it doesn't have. It is musical chairs.

What happens when the music stops and there are not enough seats?

We have lived through the internet age long enough to know that even the biggest organisations still fail at basic plumbing. Finance does not talk to HR. HR still cannot calculate public holidays properly, let alone across countries. The cure has often been monster all-in-one systems that promise everything, cost a fortune and never quite work.

That is the real-world challenge for AI. Not clever demos, but integration. Getting ten different agents to safely read the right policy, update the right database, and pull the right file from OneDrive every time. Hard to do, boring to fix, essential to scale.

Gen X readers (us Slackers) will remember when the web was going to be the be-all and end-all, then it wasn’t. Dotcoms popped. Then, quietly, the web over time ate the world anyway. I suspect AI rhymes with that.

My hunch: boards push AI spend on to CEOs to signal progress and chase savings. Many firms buy tools without a plan, do not see the returns they wanted, then rein it in. Valuations plummet back to earth. Some players go bust. Meanwhile, a slower, more durable revolution takes hold where the integration work actually gets done.

So will AI take your job? Yes, no, and maybe. It will take tasks first, jobs later, and it will create new ones. The opportunity for midlife professionals is to be the adult in the room who can turn buzzwords into working processes.

AI is neither a miracle nor a mirage. The market has sprinted ahead of the real economy. The work now is patient, practical integration. If the hype cools, that is not failure. It is the start of a more mature phase.

Also, this is a hot take. I want to hear yours.