In March 1975, a few dozen hobbyists crowded into a garage in Menlo Park, California, to talk about building their own computers. To most observers, it looked like a niche pastime for people who enjoyed soldering. The machines were kits. The software barely existed. Steve Wozniak was in the room.

Within a decade, personal computers had reshaped offices, schools and homes. The future was visible early, just not where serious people were looking. Futurists call these early hints weak signals, and the practice of systematically noticing them horizon scanning. It's less about prediction and more about attention: learning where to look, what to write down, and when a scattering of oddities starts to form a shape.

What Makes a Signal Weak

Also in 1975, a young Kodak engineer named Steven Sasson built the first digital camera. It was roughly the size of a toaster, captured black-and-white images at 0.01 megapixels, and took 23 seconds to record a single photo onto a cassette tape. Next to Kodak's thriving film business, it looked like a curiosity.

That is the defining trait of a weak signal: it rarely arrives looking like the future. It looks worse than what already exists. It's slower, cruder, more expensive, or used only by a handful of enthusiasts. It often comes from outside the industry it will eventually disrupt, presented in unpolished language by people without the usual credentials.

Established players tend to judge new developments against their current customers and current metrics. By those measures, an early signal almost always fails. Yet its awkwardness and smallness are exactly what let it grow quietly, unchallenged, until its trajectory becomes impossible to ignore. Theodore Levitt warned that companies often define themselves too narrowly. Kodak saw film. The signal was about images. Weak signals tend to live just outside those narrow definitions.

Takeaway

Ask not how good something is today, but how fast it is improving and whom it serves that incumbents ignore. Unimpressive is not the same as unimportant.

Where Signals Hide

Weak signals cluster at the edges, in places that feel too small, too technical or too odd to matter. Niche communities are one of the richest sources. Hobbyist forums, maker spaces and open-source projects are where people experiment without needing a business case. What they tinker with today often becomes a product category later.

Patent filings and research papers reveal where organisations are quietly placing bets, sometimes years before a product appears. Unusual job postings work the same way: when companies start hiring for a role that has no established title, something new is being built. Even small regulatory changes, such as a pilot licence or a new testing zone, can signal that a technology is moving from lab to street.

Perhaps the most overlooked source is the clumsy workaround. Early Twitter users typed 'RT' by hand to share each other's posts, and invented the hashtag to group conversations. Both later became official features. When people go to awkward lengths to solve a problem, they reveal unmet demand. Spreadsheets doing the job of software, group chats running small businesses, duct-taped tools: each is a rough sketch of something that doesn't fully exist yet.

Takeaway

Look wherever people are spending effort without being told to. Workarounds, side projects and odd hires are demand made visible before the market catches up.

From Noise to Pattern

A single weak signal means very little. A strange paper might lead nowhere. Forum buzz might fade by next month. Horizon scanning becomes powerful only when signals are collected over time and compared. The goal is not to judge each one on its own, but to notice when several begin pointing the same way.

A simple habit works well. Keep a running signal log: a note or spreadsheet where you record anything that surprises you, along with the date and a sentence on why it caught your eye. Once a month, review it and group entries loosely by theme. A patent filing, a new job title and a hobbyist project that seem unrelated may turn out to describe the same emerging capability from three angles.

This is the logic behind scenario planning, which Pierre Wack and his colleagues developed at Shell in the 1970s. They didn't claim to know the future. They watched for converging forces and asked what would follow if those forces continued. Clusters let you do the same: not predict, but prepare. The discipline also protects you from noise. You don't have to chase every novelty, only notice when the pile grows.

Takeaway

One oddity is an anecdote; several pointing the same way is a trajectory. Your job is to keep the notes long enough for the pattern to appear.

Tomorrow's technologies rarely announce themselves. They appear as toys, hacks, odd job listings and papers few people read. The trick is not brilliance but posture: staying curious about the fringes rather than dismissing them.

A practical framework fits on an index card. Notice what surprises you. Record it with a date and a reason. Revisit regularly and look for clusters. Then ask one strategic question: if this keeps growing, what changes? You won't catch every shift, but you'll stop being surprised by the ones that matter.