I recently found myself staring at a kitchen tap and realised that the way we treat knowing things is transforming into the same way we treat water.

In the pre-Victorian era, if you wanted water, you dug a well or carried a bucket to a communal pump. It was a discrete, laborious task. Today, you turn a handle, and the miracle of plumbing delivers it to you for a fraction of a penny. You don’t think about the reservoir; you just think about the tea you’re making.

Due to AI, we are currently transitioning from the Bucket Age of knowledge to the Utility Age.

For the last thirty years, the internet has been a giant library where we had to go and fetch information. We Googled. We sifted, we browsed. But with the rise of Large Language Models, knowledge is being pressurized and piped directly into our existing workflows.

The value of a product isn't just its utility, but also its convenience. If you make something 10% easier to use, people don't just use it 10% more; they use it in entirely new ways. When knowledge becomes a utility, that is to say, on tap, always on, billed by the token, then the sociological shift won't be that we know more, but rather that we stop valuing the act of knowing. We used to admire the learned man - the person with a deep well of internalised facts. In the Utility Age, that looks a bit like someone who keeps fifty gallons of water in their bathtub just in case the mains fail. Impressive, certainly, but a bit eccentric.

The danger, of course, is the Roman lead pipes problem. When water is a utility, we stop worrying about how it’s purified - until the pipes rust or the supply is poisoned. When knowledge is a utility, we stop checking the sources. We trust the reservoir of the AI's confidence.

We are moving toward a world where intelligence is no longer a personal attribute, but a monthly bill. The question is: when that bill comes due, will we still remember how to dig a well?