Signal vs Noise

The loudest, most dramatic news is usually noise — real signal is the quiet trend that keeps showing up.

The Future of Technology · Lesson 2 · 7 min read

Two things land in your feed the same week. One is a jaw-dropping video — a robot does a perfect backflip — with the headline “Robots are taking over!” The other is a dull industry footnote: the cost of a certain technology dropped again, the way it has nearly every year for a decade. Your attention lunges at the backflip. But which one actually tells you where the future is going? You have limited attention and the internet is engineered to spend it for you. Hold that question — by the end you should know which one to bet on, and why.

Noise is loud; signal is persistent

Noise is the dramatic single event: the viral demo, the breathless headline, the chart that spikes and reverts, the one impressive result cherry-picked from many flops. It is short-term, surprising, and emotionally loud.

Signal is the durable underlying trend — usually quiet, often boring, visible only when you step back across months and years. Here is the catch that fools almost everyone: the systems that deliver you information select for noise, because surprise earns clicks and a steady trend is “not news.” So the loudest thing in your feed is, on average, the least durable. Importance and volume are barely related.

Three filters for finding the signal

You do not need a crystal ball — you need three questions:

1. Is there a driver? A real, persistent force — a falling cost, a law of physics, a strong incentive — that keeps pushing in one direction beats any one-off event.
2. Does it persist? Look across many data points over time. A trend that survives years is signal; a spike that reverts next quarter is noise.
3. Direction over level. A thing improving steadily matters more for the future than a thing that was impressive once.

Run anything through those three and the hype usually sorts itself out.

Apply the filters to the two feed items:
• Robot backflip — one demo, no persistence shown, likely the best take out of dozens. Driver? Not visible. Persistence? None shown. → noise.
• Steady cost decline — a clear driver (scale + competition pushing prices down), persistent across many years, pointing in one direction. → signal.
The backflip is forgotten in a month. The cost curve quietly decides which products even become possible. Bet your attention on the second one.

Beware the hype cycle — and your own attention

New technologies tend to travel a predictable arc: a burst of inflated expectations, then disappointment when reality lags, then a quiet, genuine climb to a useful plateau. Notice the timing: loudness peaks early, often before the thing is actually useful. So “everyone is talking about it” measures attention, not importance.

Two cheap tools keep you honest. Base rates: most flashy new things fizzle, so start a little skeptical and let persistent evidence move you. And the hype cycle itself reminds you that the quiet plateau — not the screaming peak — is where real adoption lives. This is what grounded optimism looks like: patient, betting on durable drivers, unbothered by whatever is loudest this week.

Suppose a new gadget gets a viral launch: huge demo, sold-out pre-orders, “this changes everything.” How do you read it?
1. Driver: is anything forcing it forward — a falling cost, a real need — or is it riding novelty? If only novelty, expect a fade.
2. Persistence: come back in a year. Are people still using it, or did the spike revert? One launch is a single data point.
3. Direction over level: a clunky product that improves every quarter is more telling than a slick one that stands still.
A skeptic who ran these filters during many past “revolutionary” launches would have correctly ignored most of them — and correctly held onto the few with a real driver underneath. Same method, both ways.