The Room Wherethe Future Was Already Talking

In 1977 Ken Olsen stood in front of the World Future Society in Boston and said he could see no reason for any individual to have a computer at home. He was the president of Digital Equipment Corporation, one of the most successful computer companies then alive, and he was addressing a convention of people who had assembled specifically to think about what was coming. That same year, in garages an hour from nothing important, the Apple II shipped. Olsen was later at pains to explain that he had meant something narrower, a central machine running the household rather than a desk computer, and the correction is fair. It is also beside the point. The signal was in the room. So were the people paid to hear it.

01 · The argumentWhat the book actually claims

Amy Webb’s claim is deceptively administrative: forecasting is a procedure. Not a gift, not a temperament, not the private property of people who read a great deal of science fiction, but a repeatable sequence that a team can run on a Tuesday and audit on a Friday. Futurists, she argues, are not seeing further than everyone else. They are looking somewhere else, earlier, with a method that stops them from mistaking noise for news and stops them from dismissing news as noise.

The sequence has six steps, four for finding a trend and two for acting on it. Go to the fringe, where the unusual suspects are running experiments nobody has commercialised. Test what you find against six pattern markers she compresses into CIPHER. Interrogate the pattern to establish whether it is a trend at all rather than something merely trendy. Calculate its timing. Build scenarios across the probable, the plausible and the possible. Then pressure-test the strategy you have chosen against a six-part checklist. The step that gives the book its title sits underneath all six: signals are already talking, continuously, and the discipline is a listening discipline before it is a predictive one.

02 · Author’s vantageWho is making it, and from where

Amy Webb is a quantitative futurist, a professor of strategic foresight at NYU Stern, and the founder of what was then the Future Today Institute and now trades as the Future Today Strategy Group. Her training is in journalism and economics rather than in strategy consulting, and it shows in the book’s best passages: she reports the fringe rather than theorising about it, and she has physically gone to the places she describes. The book won the Thinkers50 Radar Award and an Axiom Business Book Award, and it sits upstream of her later work on artificial intelligence.

03 · Key insightsFive things the book gets right that most planning processes get wrong

The fringe is a place, not a mood

The instruction to watch the periphery is the oldest cliché in strategy and the least operationalised. Webb’s contribution is to make the fringe addressable. It is the population of researchers, hobbyists, hackers, artists and cranks working adjacent to your problem without any commercial obligation to solve it your way, and it can be enumerated: name the domain, name the people running experiments in it, read what they publish, attend what they attend. She builds a fringe sketch as a physical map, with your organisation at the centre and nodes placed by distance from your current business. The discipline is that the map must contain names. A fringe you cannot list is an anxiety, not an input.

CIPHER converts a hunch into a testable pattern

The book’s most portable device is the acronym it uses to interrogate what the fringe is doing. Six markers: contradictions, inflections, practices, hacks, extremes and rarities. Each is a specific thing to look for rather than a category to feel. A contradiction is behaviour running against the stated trend. An inflection is a bend in an established curve. A hack is a workaround people build because the official product will not do what they need. What the acronym does is force the analyst to say which marker fired, which converts an intuition into a claim that a colleague can dispute. That is a small thing and it is most of the value.

Trend and trendy differ by structure, not by excitement

Webb spends more effort separating a trend from a fad than most foresight books spend on prediction itself, and the distinction is structural. A trend, in her usage, is a new manifestation of a persistent human need, driven by identifiable forces, that converges from more than one direction and is sustained by more than novelty. A fad meets none of those conditions and often looks more exciting because novelty is what it is made of. The test is not whether something is growing. It is whether the growth is anchored to a want that predates the technology by a century. That framing survives contact with almost every cycle since publication.

Timing is a calculation, not an instinct

The step readers skip is the one that costs them. Webb argues the arrival time of a trend is a function of internal technological development set against external events: regulation, capital, culture, the price of a component. The consequence is that being right and being early are different failures with different remedies, and a firm that cannot tell them apart will alternately over-invest and under-invest for years while believing itself consistent. Most corporate foresight is not wrong about what. It is wrong about when, and it treats that as a small error.

Scenarios are built in three registers, and the third one is not decoration

Webb asks for three kinds of story about any trend. The probable extends what already exists. The plausible respects physics, economics and politics but allows the arrangement to change. The possible violates none of the laws of nature yet requires something we do not currently know how to build. Most corporate scenario work produces four versions of the probable and calls the set a range. Her insistence on the third register is not indulgence; it is a stress test for the assumptions the probable case is silently carrying. And each scenario must be written as a narrative with an if-then spine, because a scenario nobody can retell is a scenario nobody will act on.

04 · Where the argument echoesThree fields that formalised this before management did

The book’s structural claims rhyme well outside foresight. Each parallel sharpens something Webb leaves implicit.

Signal-detection theory supplies the discipline the fringe sketch lacks. Radar operators and radiologists work with a formal apparatus for exactly Webb’s problem: separating signal from noise when both are present and the cost of the two errors differs. What the theory adds is the concept of a criterion that can be moved deliberately, and the recognition that lowering it to catch more true signals necessarily catches more false ones. Webb tells the reader to look harder at the periphery. Signal-detection theory says the yield of looking harder is fixed unless you also decide, in advance, what ratio of false alarms you are willing to fund.

Epidemiology already runs the fringe as infrastructure. Sentinel surveillance places monitoring sites deliberately at the periphery of a population rather than at its centre, tracks syndromic proxies that move before confirmed cases, and defines the action threshold before the outbreak so that crossing it triggers a response automatically. The parallel exposes what Webb’s method leaves to judgement. She asks leaders to interpret what they find at the moment they find it, which is precisely the moment epidemiology concluded judgement cannot be trusted. Pre-commitment is the missing component, and it is cheap.

Materials science explains why the fringe is a production site rather than a viewing gallery. A supercooled liquid can sit below its freezing point indefinitely because a phase change requires nucleation, and nucleation almost never begins in the pristine interior. It begins at an impurity, a scratch, a boundary. Read this way, the periphery is not merely where change becomes visible first. It is the only place where the new state can form at all, because formation requires the irregularity that the well-run centre has spent years engineering out of itself.

05 · The enterprise translationWhat the method would have caught, and what it would not

A framework published in 2016 has had an unusually severe decade to survive. Generative AI ran the full fringe-to-mainstream sequence inside five years, GLP-1 medicines reorganised two consumer categories, and a long list of incumbents with functioning research departments were caught by both. The useful question is not whether Webb’s method was available. It was, in paperback, for a modest price. The question is what it would have produced in the hands of a firm that actually ran it.

Take the clearest case. The transformer architecture was published openly in June 2017 by researchers at Google. It was, in CIPHER terms, an inflection: a bend in a curve that had been improving steadily for years. By 2019 GPT-2 was a rarity, an artefact whose fluency should not have existed at that parameter count. By 2021 shadow usage of language models inside software teams was a hack, the marker Webb rates most highly because it shows people routing around a product that will not do what they need. Every one of those markers fired in public, in papers, in developer forums, in job postings. The signal was not scarce. Chegg, whose entire business sat in the path, acknowledged the threat in a regulatory filing within ten weeks of ChatGPT’s launch and shipped a countermeasure inside six months, then lost roughly ninety-nine per cent of its market value anyway. Its failure was not in step one. It was in steps five and six.

The GLP-1 case tests the timing calculation rather than the detection. Exenatide was approved in 2005; the appetite-suppression effect was documented in the clinical literature long before it became a consumer phenomenon; the arrival in 2023 was triggered by external events Webb’s equation names explicitly, namely reimbursement, supply and a shift in cultural permission. A packaged-food company running her ETA calculation in 2019 would have found a trend with a strong internal-development score and a weak external-event score, which is exactly the profile that says watch closely, commit small. Nestlé launched a portion-aligned brand in May 2024. WeightWatchers, which occupied the arena the drugs took over, bought a telehealth provider to prescribe them and filed for Chapter 11 in May 2025 to shed $1.15 billion of debt. Same signal, same decade, opposite readings of the clock.

The counter-case matters as much. Run the method in 2016 and it would also have flagged consumer virtual reality, blockchain settlement for trade finance and the autonomous vehicle timeline, all three of which cleared the fringe test, showed several CIPHER markers and generated confident arrival dates that were wrong by five to fifteen years. Webb’s framework does not distinguish these from the transformer case in advance, and no framework in the literature does. What it does offer is the sixth step, the one that asks whether the strategy survives if the arrival is late, and organisations that ran it emerged with option positions rather than committed factories.

06 · DisciplinesFour things to change in the operating rhythm

07 · RelevanceWhy a 2016 book about listening reads differently in 2026

A curious thing has happened to this book. Its method has held and its examples have aged into evidence, but the constraint it was written against has inverted. Webb was writing for organisations that could not see the fringe: the periphery was genuinely hard to reach, the papers were hard to find, the practitioners were scattered across conferences nobody in the executive team attended. Ten years on, the fringe arrives daily, pre-summarised, in more feeds than any leadership team can read. Scarcity has moved from detection to discrimination.

“We’ve made a devil’s pact, swapping convenience and efficiency for an ever-increasing tyranny of information and choice.” Amy Webb, The Signals Are Talking

That shift changes which steps carry the weight. Step one has been substantially automated, and the tools that automate it are the same ones that generate the noise. Step two matters more than ever, because a marker-based test is one of the few filters that scales without collapsing into taste. Step four is where the modern damage happens: the lag from fringe to mainstream has compressed to the point where the arrival lands inside a single planning cycle rather than across three, which means the calculation has to be run continuously rather than at the offsite. And step six, the pressure test, is now the difference between a firm that holds a portfolio of small positions and one that makes a single large bet on a date it has not examined. A reader who takes this book as an argument for better scanning has read the easy half.

08 · Where the argument strainsFour lenses that sharpen the case rather than dismiss it

The book invites engagement from the disciplines it borrows from. Four lenses bring its genuine claims into relief.

Through the lens of forecasting research, the book has a scoring problem it never confronts. Philip Tetlock’s work established that the hard part of prediction is discrimination rather than sensitivity: raising alarms is free, and skill lives in the ratio of true alarms to false ones. Webb’s cases are selected on the outcome. Every fringe signal she describes turned out to matter, which tells the reader nothing about the far larger population of fringe signals that correctly went nowhere. A team that runs her method without keeping score will generate more warnings and no more accuracy, and will feel more prescient while doing it.

Through the lens of information theory, the acronyms do less work than their structure suggests. CIPHER is a compression scheme, and every compression scheme discards. The six markers are heuristics for spotting departures from an expected distribution, which is genuinely useful, but the book presents them as a taxonomy rather than as a lossy filter with a known failure mode: signals that arrive as slow, smooth, distributed change register on none of the six. Climate adaptation, demographic ageing and the erosion of institutional trust are all consequential and all invisible to a marker set tuned for anomalies.

Through the lens of organisational theory, the method assumes an organisation that will act on what it finds. Webb gives four of six steps to detection and two to response, and the two are the shorter ones. Chris Argyris’s distinction between espoused theory and theory-in-use is the relevant apparatus here: firms espouse curiosity about the periphery while their promotion, budgeting and reporting systems reward the defence of current revenue. The book’s remedies are almost entirely informational, and the failure it diagnoses is almost entirely political. Read alongside McGrath, the gap closes; read alone, it stays open.

Through the lens of political economy, the author’s position is worth holding in view. Publishers Weekly was blunt about it in 2016, calling the framework acronym-heavy and arguing the overall effect left futurology looking like a dark art navigable only by hiring experts such as the author. The criticism is unfair to the method, which is genuinely teachable, and fair about the packaging, which is proprietary in a way that a genuinely open method need not be. Three rules, six steps, six markers and a six-part test is more scaffolding than the underlying insight requires, and the surplus scaffolding is what a consultancy sells.

Refracted through these lenses, the friction above is not an objection to Webb’s perspective. It is scaffolding the reader can build around the book’s argument, a place to anchor their own engagement and test the framework against disciplines it touches. What she built survives the engagement, and the reader gains a firmer grip on why it does. The fringe sketch and the marker test are durable instruments that a competent team can run without her; what the lenses expose is that the book is strong on how to look and thin on how to keep score, and that keeping score is what separates a foresight practice from a subscription.

09 · Adjacent readingWhere to go next

10 · ClosingWhat the room already knew

Olsen was not a fool, and the correction he spent years issuing was legitimate. He meant a central machine running the household, not a desk computer, and on that narrower claim he was closer to right than the anecdote allows. What survives the correction is the setting. He was standing in front of a room convened specifically to think about the future, in the year the Apple II shipped, describing the edge of his own market as a place with no reason to exist. Webb’s method is a procedure for the person sitting in that audience: go where the unusual suspects are, name the marker that fired, work out when rather than whether, and write down what you would do if the arrival came early. None of it requires unusual foresight. It requires the willingness to write down a claim that a colleague can dispute, and to be the one holding it when the arrival is late.

If someone in your organisation heard the signal this morning, what would it cost them to say so out loud?

11 · What to carry away

12 · ReferencesSources for the business cases and figures used above

Every business figure used in the sections above links to a primary filing, company announcement or the original research paper. Interior text of the book itself was not independently accessible during synthesis.