What counts as a life cycle

The term is applied loosely. Many consultancy diagrams show a process that loops back to the start once the work is done, which is a repeating process rather than a life cycle. A true life cycle has a distinct end, even if something else takes over afterwards: a product replaced, a company absorbed or closed. Whether the cycle is visible also depends on the viewpoint. From inside, a life looks like a straight line from birth to death; from far enough above, the parts are reused and the pattern repeats. A life cycle is also never isolated, because it sits within systems that interact with it.

Companies: act your age, or look ahead?

One view is that a company should act its age: once mature, it should return cash to shareholders and let its assets and people be used elsewhere. Philips is the usual example, and it can be read two ways. One reading is that it was a mistake. Its engineers developed lithography technology that the company judged to be outside its interests, and that technology became ASML, now worth far more than Philips itself. The other is that letting go was the right course: the assets went to work elsewhere, and shareholders were spared a management looking for new ventures to justify its own survival.

The record is more mixed than either reading. Philips did not simply drop these ventures. ASML began in 1984 as a joint venture with ASM International, and Philips was the largest founding investor in TSMC, supplying capital, technology licences and its first chief executive. Both were set up at arm’s length, with partners and a minority stake, which may be the way to encourage a new business without smothering it. Philips later sold both holdings. That bears on a common failing: large firms tend to be poor at nurturing small ones, because they run them by the rules suited to large, capital-heavy operations.

The same question applied to a chemical company that set out to break its boom-and-bust cycle, came close to bankruptcy through an execution error, and then recovered. The strategy may have been sound, but the recovery looked better for management than for shareholders, who could have moved their money elsewhere. Whether shareholder value should be the yardstick at all was also questioned. A more practical suggestion was that large organisations hold several life cycles at once and need different people for each. The people who run mature plants are not necessarily the ones who should start new ventures. Decisions to exit a market are better made case by case than by doctrine.

Products and technologies

Pharmaceutical products follow a fixed path from discovery through development and patent protection to generic competition, so the company must keep finding something new that others cannot copy. Technologies also shift underneath, from small molecules to proteins to genetic engineering, and firms like Nokia show the cost of missing the next one.

Decline is not always accepted. One agrochemical product is still sold fifty years after patenting, and an old beer brand regained value by being made scarce and more expensive. Others drop a product on a rational view of the market. Stage also changes the skills needed: launch and growth call for different people than maturity, where cost and defending against decline dominate. A parable about the wheel illustrates this. Once a technology simply works, buyers care about range and colour, which is why new phones seem to be announced mainly through new colours.

Imposed life cycles

Some cycles come from outside the market. Patent expiry is a deliberate regulatory cycle, intended to keep medicines affordable while rewarding innovation. The planned end of combustion-engine sales spurred investment in electric cars, and the phase-out of CFCs produced replacement chemicals and, apparently, a recovering ozone layer. Management fashion produces a third kind: periods favouring divestment, then vertical integration, with little rationale beyond the trend.

Misaligned cycles and inertia

For procurement, the problem is rarely one cycle but the mismatch between several. A mature buyer, a young supplier and an experimental technology may all sit under a contract of five to ten years. Procurement also sees technologies moving faster in the market than inside its own organisation, which makes its warnings unwelcome but valuable.

Finally, there was the suggestion that human laziness and attachment to comfort drive much of this. Organisations and sourcing relationships settle into habit. Deliberately ending something can shake that loose, though such experiments sometimes fail.

Participants in the SWISSUES Forvm event on which this article is based were Armin Scharlach, Pramod Prasanth, Stephen Fulton, Ercan Solmaz, Halyna Yokovleva and Bill Young. The article is edited with assistance from Anthropic’s Claude LLM.

Companies quickly exploit stability. They simplify and streamline; they reduce reporting time-horizons; and they eliminate complexity. They do all this with alacrity or fall behind their competitors.

They are slower to adapt to instability. Organizational systems were formed when change was steady. Its leaders know about measurement, optimization and control. Capability to respond to complex and unstable change however was lost.

Adaptation to stability is about stripping out unnecessary systems. Adaptation to instability is discovering how systems interact in a new environment. Companies that optimized for stability are always caught out by events. In the late 1960s some big name corporations – GE, IBM and Shell, for example – decided that something must be done and others followed.

For thirty years, corporate strategy departments were the ‘something’. They became fashionable after Igor Ansoff’s 1965 book, Corporate Strategy. It was believed then that clever specialists with the right information could predict future events and plot a way through them. Several things led to their demise. Some predictions were spectacularly wrong. Strategic planning staff were an easy target for cost-reduction and de-layering. Consultants offered a strategy service as episodic, billable engagements. Michael Porter reduced strategy to a set of analytical tools. And Finance functions persuaded executives that companies could be run on capital allocation and shareholder value, rather than market-focused strategy. The death knell came in 1994 with the publication of Henry Mintzberg’s book, The Rise and Fall of Strategic Planning.

It was not just strategy departments that faded. Strategy itself lost its former meaning in boardrooms. It became planning. In The Fish Rots From the Head (1996), Bob Garratt described how executives drift into micromanaging operational plans because it is easier and more familiar than the harder work of horizon-scanning and holistic understanding.

Companies need to learn strategy afresh, from the ground up. The strategy departments of the late twentieth century got it wrong. They modeled the world from the company outwards. That’s like having an Earth-centered view of the Universe. They need a model that sees the company as one system inside many others, at the same time as seeing all the company’s internal systems interacting.

This can be imagined as two strategic poles. One is awareness of external change and the other is understanding organization and governance. Each can be illustrated from the author’s own experiences.

Awareness of external change

ICI’s fertilizer business grew phenomenally profitable by dominating the UK market. It had rock-bottom natural gas prices, was the world’s most efficient producer of ammonia, and enjoyed protection from continental imports through awful port-infrastructure and militant dock unions in the UK. It did not go unnoticed when all of these favourable conditions began to change, but individuals who spoke up were ignored or sidelined. There was no group with authority to raise it with the corporate board. ICI was a major global chemical company but nobody remembers it today.

Understanding organization and governance

Novartis had over a thousand employees in its Procurement function. Reporting to Finance, the only performance metric that drove it was cost-reduction. The CFO even thought it would be a good idea to include billion-dollar savings targets as ‘productivity’ in the company’s annual report. Cost-centre owners reacted to attacks on their budgets by ensuring that Procurement came nowhere near them. Sophisticated shadow procurement functions developed across IT, HR, Facilities, Insurance, Engineering, and, yes, ironically, in Finance! The dysfunction was visible. Nobody had the responsibility to see it.

Forecasting predicts outcomes by extrapolating data. Scenario planning, used by the former strategy departments, explores a range of alternative futures. The strategy that companies need today adds another dimension, systems thinking. Systems thinking sees an environment as the interaction of evolving systems within systems-of-systems - a complex, recursive, constantly changing ecosystem. Interactions can create reinforcing and balancing feedback that stabilizes some systems. These stable systems persist and replace those that decay. This creates systems-of-systems that appear organized, functional, self-correcting and designed – even though their emergence is entirely natural. Organisations can be the product of emergence rather than purpose.

Systems thinking, though established and well recognized, is rarely applied in management. It has no quick-fit tools and it does not provide the neat, short-term solutions companies prefer. There is a belief that visionary leaders are able to discern patterns and innately perceive the evolutionary path of systems. That may be the case but even if they do, it is wrong of companies to rely on a one or two executives to set direction. They need a pool of strategy talent reporting to the CEO. It would recruit a team of strategic thinkers and influencers. It would be a recognized career step into leadership. A number of companies have developed structural ways to encourage systems thinking in strategy. An example is Mars, the food group, who made the functions of Research & Development, Supply and Procurement all report through the same executive. All of these need end-to-end understanding of value chains, as well as an ability to scan political, commercial and technical horizons. These are the fields in which strategic capability is most likely to be nurtured and where unified reporting could deliver real competitive advantage through better strategy.

Systems thinking is developed by a diversity of experience that does not happen in narrow functions. Cross-functional task teams can deliver projects but can hardly challenge tenets of the organization. Development of strategic capabilities – sensing, challenging, exploring, connecting, experimenting, learning and judging – is not easy to encourage in operational units. They do not have the heterogeneity of a dedicated strategy group.

A strategy group would be delegated tasks by the CEO, that would previously have been awarded to strategy consultants. It would also, on its own initiative, challenge the organization’s operating model, not just its value chain but also the very existence of some corporate functions. Procurement is an example. Internal departments undertake supplier selection; they use vendor platforms to deliver a portfolio of services; and some have embedded vendor strategy groups. It is unclear why a dedicated Procurement function continues to exist but nobody in the organization has a remit to think about and report on it. A clear role for a strategy group.

An important feature of a strategy group is a level of autonomy in challenging the organization. In this respect, it would be unlike the internal consulting teams established to avoid reliance on McKinsey, Bain, BCG et al. This is another reason why it must report directly, at a minimum, to the CEO.

Back-to-the-future is not a return to 1980s scenario planning. It is a call for strategy to be taken seriously.

Managerialism vs leadership

Much of the hour circled a single distinction, approached from several angles:

Skills vs title.

Being called a manager is increasingly a function of pay grade and tenure rather than of possessing the underlying skill set — the EQ, structure, and judgement that the job actually requires. Professions built on seniority ladders (law was cited directly) were said to produce senior figures who are excellent practitioners but poorly equipped, or unwilling, to manage people.

Hard skills vs soft skills.

One view split management (process, procedure, “hard skills”) from leadership (motivation, trust, relationship-building, “soft skills”) — with leadership pitched as the durable skill set precisely because process-following is what AI will increasingly be able to do.

Process vs strategy.

Managerialism was repeatedly described as technique in service of order, predictability, and coordination; leadership as direction, vision, and the willingness to evaluate the external environment and choose a path. Taken too far, managerialism was said to trade judgement, speed, and ownership for process compliance.

Timescale.

A recurring axis: management as short-to-mid-term operational delivery, leadership as the multi-year “North star.” One participant’s live example was AI strategy at board level — leaders talk about AI constantly but mostly at the level of tool adoption (e.g. Copilot), rarely as a multi-year business transformation; in that domain, in practice, almost everyone present is behaving as a manager, not a leader.

Decisions and blame.

Managerialism was framed as a way of avoiding accountability — “I followed the process, so I can’t be blamed for the outcome” — versus leadership as owning a decision, including the risk of being wrong, especially under time pressure (the example given was reacting to a supplier’s sudden bankruptcy, where waiting to complete a full process would be fatal).

Where managerialism helps, and where it hinders

Managerialism is not inherently good or bad. It helps in high-risk, repeatable environments needing coordination across many functions, where interchangeability matters more than creativity. It hinders where problems are complex and need contextual judgement, where innovation depends on curiosity and dissent, and where speed matters more than structure.

Incentives and reward systems

A significant thread connected managerialism to how people are paid and evaluated. Reward systems built around clear, narrow KPIs (”carrot and stick”) work well when the task is well defined and the goal is simply to do more, faster. They work far less well against today’s less-defined, more peripheral problems, which require creativity and departure from established process — precisely what bonus-driven KPI compliance discourages. This was offered as one explanation for why leaders default to process-following and keeping metrics green.

Boards, capability, and time horizon

A live debate: has excessive pressure and pace actually pushed leadership toward more managerialism, not less? One perspective, drawn from direct experience rather than theory, was that in difficult, fast-moving periods, leaders often revert to directive, managerial behaviour rather than rising to strategic leadership, because the same pressure lands on them too. This was contrasted with the view that true leaders (as distinct from people merely occupying leadership roles) hold the wider vision precisely when times are hard — and that team composition and experience level also shape how much direction versus autonomy is appropriate.

Separately, direct board-level experience was shared describing boards that lacked operational grounding in the company’s core function — a pharmaceutical board with no supply chain expertise ahead of the industry’s later pandemic-era supply shocks; a food manufacturer’s board and executive committee that had to be trained from scratch on commodity hedging despite hedging being central to protecting margins. The point made was not that boards should be pulled into day-to-day operations, but that some baseline fluency in the company’s core operational risks (supply chain, hedging, labour) is a precondition for credible oversight — a version of “diversity of thought” distinct from demographic diversity. This was linked to a broader pattern of boards and leadership optimising for the next reporting cycle (annual for public companies, roughly 2–3 years at most) rather than a 5–10 year horizon, with private and family-owned companies tending to default to even shorter, year-to-year thinking.

Organisational life cycle

Drawing on Aswath Damodaran’s framework for company life stages, a case was made that the appropriate style of leadership is not constant: early stages call for craft- and builder-driven leadership with high ownership, while later, mature stages become about keeping output steady (”widgets off the line”). Knowing where an organisation sits in its life cycle — and adjusting behaviour accordingly — was proposed as a precondition for knowing when managerialism is the right tool and when it isn’t. Google and Boeing were cited as examples of organisations that shifted from engineering- or founder-led cultures to manager-led ones, arguably at a cost to sustained innovation; Philips’ deliberate contraction (which enabled the rise of ASML) was cited as a case of managing decline well.

Outsourced accountability

A further thread: heavy reliance on external consultants (McKinsey, Bain, BCG were named as a category) can let both management and the consultants disclaim responsibility for outcomes — “it was their recommendation” / “we only advised, management decided” — leaving no one accountable. This was linked to a broader concern that excessive managerialism can tip into organisations serving the interests of an internal group (an “oligarchy”) rather than their stated purpose, and into people occupying roles designed to satisfy process rather than create value.

Closing synthesis

The discussion converged on a shared framing offered near the end: managerialism is not good or bad in itself — it is powerful. It brings order to complex systems but can pull attention away from outcomes and into process. It tends to grow, unnoticed, in the gaps where judgement should be exercised. The practical challenge for any organisation is recognising where structure creates clarity and should be kept, and where it creates noise and should be removed — so that the system serves the people running it, rather than the reverse.

Possible future topics raised

Whack-a-mole management — whether organisations are structured to react instantly to whatever happened yesterday, at the expense of strategy and planning.

Locus of control(internal vs external) — drawing on a recent *Economist* piece — the difference between believing you can shape outcomes through your own agency versus navigating primarily in response to your environment.

Organisational and product life cycles— understanding which life-cycle stage a business, product, or capability is in, and how that should change strategy and behaviour; noted that a single company may contain multiple products or capabilities at different life-cycle stages simultaneously, competing for the same resources.

Curiosity was framed as the basis of humanity itself — the driver behind communication, technology, and social development. Without it, nothing advances; society doesn’t collapse but plateaus. Historical examples (postwar jet-engine development, Concorde) were used to argue that major leaps happen when curiosity moves from individual tinkering into mainstream, collective effort. A counterpoint was raised: intense focus on a single target (e.g., building Concorde) is itself a form of narrowed curiosity, and undirected curiosity can be seen as a distraction or wasteful — though most felt curiosity doesn’t need a defined target to be valuable, since exploration without an immediate payoff often pays off later.

What Triggers Curiosity: Dissatisfaction vs. Inspiration

A central debate emerged over the source of curiosity. One position held that curiosity is rooted in dissatisfaction — a felt gap or lack that pushes people to seek something better (from vinyl to CDs to streaming was offered as an example). The opposing view held that curiosity can arise from positive inspiration or openness to new experience, with no dissatisfaction required — simply noticing something and wanting to know more. A middle position suggested every act of curiosity needs some kind of trigger, whether cognitive/rational (dissatisfaction) or emotional (inspiration), and that “lacking something” is present even when it isn’t experienced negatively. Boredom was also proposed as a trigger distinct from dissatisfaction — many breakthrough ideas were said to emerge from low-energy, bored downtime rather than acute need.

Curiosity Inside Organizations: Safety, Boundaries, and Disruption

Companies often claim to want curious employees (one example: an employer listing curiosity as a formal guiding principle), but in practice they may resist the disruption curiosity produces, especially where established or political processes are challenged. Curiosity was said to require a “safe space” or at least a degree of trust and psychological safety to be expressed without personal risk. Leadership was described as needing to set boundaries or “guardrails” — enough freedom to explore, but within limits relevant to the work — otherwise curiosity risks becoming unfocused or wasteful. It was also suggested that what looks like disruptive curiosity from an employee may actually be a symptom of that person not having been involved in designing a process in the first place, rather than genuine curiosity.

Systems Thinking: Boundaries, Boxes, and Unintended Consequences

Curiosity exercised only “inside the box” of one’s own function or role produces limited results; real impact requires thinking about how actions ripple across organizational boundaries (an example given: HR self-service portals that reduce HR headcount but shift the burden — and time cost — onto employees elsewhere in the business). The point was made that most organizational functions are artificially bounded, and genuine improvement often requires questioning those boundaries rather than optimizing within them. Curiosity aimed too low, or curiosity that stays inside a box, was described as generating a lot of activity without meaningful return — curiosity needs to be “aimed high” and tied to a large enough outcome to be worth channeling.

Developing Curiosity: Challenges, Education, and Culture

Educational and developmental approaches were discussed as ways to build curiosity deliberately: shifting from teacher-centered/top-down instruction to student- or employee-centered approaches where people are given open, sometimes unsolved or paradox-laden challenges to work through themselves, rather than being told what to do. One academic case study found that a creativity course measurably increased analytical students’ strategic/out-of-the-box thinking, but had a much smaller effect on business students who may have started from a higher creative baseline — suggesting different starting points require different developmental approaches. Company culture — how open an organization is to new ideas — was also raised as a major determinant of whether curiosity gets expressed at all.

AI and Curiosity: Amplifier or Substitute?

Views diverged on whether AI increases or diminishes curiosity. On one hand, AI was seen as dramatically speeding up the process of getting answers and partial answers, freeing people to explore and build further (particularly valued in consulting/research contexts). On the other, there was concern that AI can be used to simply retrieve a ready-made, well-written answer without any curiosity being exercised at all — and that relying on a single AI model risks producing “gray average” recycled thinking rather than original insight. Suggested mitigations included querying multiple AI models against the same question, and treating AI as a “probe” that challenges one’s thinking (asking what’s missing, what other patterns exist) rather than a replacement for it. A related distinction was drawn between learning what an AI’s answer is versus understanding the methodology behind it — without the latter, users gain outcomes but not developed judgment. It was also noted that skill in prompting AI effectively is highly context-dependent: simple lookup use requires little skill, but specialized applications (marketing campaigns, coding) demand real prompting expertise.

Selling procurement software is easy. Behind the pitch-deck, however, the sector is the dullest of fields for technology. There just aren’t many problems to solve and the rewards diminish with scale.

The only measurable return is reduced headcount

Abstract aspirations like spend visibility and supplier enablement feature in business cases but the CFO skips over them to look for hard numbers in the RoI document, and these are fewer people doing the work. Contract-drafting, RFQs, spend analytics, invoice reconciliation - each benefit is measured in headcount reduction.

Savings are just more headcount reduction

The second selling point is savings - on the spend itself, rather than the people managing it. It sounds separate and additional but it isn’t. Savings not already being worked on are likely to have already been classified as too small or too scattered to justify effort. Aggregating and automating them may drive efficiency but only where the rewards were marginal to begin with.

Old technology with new branding

Behind the procurement tech-talk what you find are rules-engines and workflow-automation that have been mature for over quarter of a century. Purchase orders, supplier records, contract terms, invoice lines are database-, not data-problems. The innovation on offer is better UX through interface-polish and integration-plumbing, repackaged with this decade’s vocabulary.

The hard problems don’t belong to one company

There are real Big Data problems in procurement-related areas. Supplier risk, demand signals, geo-political indicators and social network data are all valid external sources of information. But few procurement departments can build a business case for owning capabilities with contingent or sporadic applications. These are better suited to specialist providers operating across hundreds of clients. And that is where this market is moving.

Adding it up

A finite set of processes to automate. Value capped by the manpower reduction. Savings too small to have mattered previously. Diminishing returns with each step. And the hard problems living somewhere else. Those are the components of the RoI.

Set against sales and marketing where the addressable prize is revenue - open-ended compounding and scalable - they spell out meagre rewards.

CFOs find a business case for quotidian savings far easier to understand than one based on growth driven by changing behaviours; and that is what has pulled so many tech entrepreneurs into the Procurement sector. But most have failed to look critically at the real benefits to the client, beyond the improved UX and some headcount reduction.

I feel for them - and their investors.

Humour as an intelligence marker.

Several participants linked wit to intelligence, though the group split on which kind. One framed the ability to deploy humour well and appropriately as a longstanding sign of high general intelligence — the capacity, as it was put, to see the funny side of any situation, even if you don’t always say what you see. Others pushed back, arguing the real driver is emotional intelligence (EQ) rather than raw IQ: the skill lies less in the joke itself than in reading the room, sensing when levity will land and when it won’t, and knowing when to hold back. This distinction mattered because it fed directly into the teachability question — if humour is fundamentally a technical craft (timing, structure, repetition), it can be taught like storytelling; if it’s fundamentally a form of social perception, training has much less to offer, since perception of a room is not easily drilled.

A later observation complicated the picture further, rather than settling it: the kind of humour that lands with a room of PhD students is different from what works in mainstream entertainment, suggesting the relationship between intelligence and humour isn’t a straight line at all. Cleverness may shape the *style* of humour a group favours without making that group better, individually, at using humour well. The two claims — humour signals intelligence, and intelligence doesn’t reliably predict good humour — were never quite reconciled, which is part of why the group felt this was worth returning to.

Humour in team-bonding

Shared humour was described as constitutive of team identity: its own in-jokes, its own register, a shorthand that signals who belongs. Several participants recognised this from experience — the sense that a team’s humour is inseparable from how it operates and how it sees itself. But the same dynamic has a clear shadow side. What feels like harmless shared banter to those inside it can quietly exclude anyone outside the joke, and the people doing the excluding are often the last to notice — it typically takes someone stepping outside the dynamic to name what’s happening. One useful frame raised here was a simple team archetype — hero, victim, joker — as a way of noticing when humour has become a fixed role in a group rather than a spontaneous response to a moment, and when the “joker” role is being used to avoid rather than defuse tension.

Humour as message clarification — and its opposite.

In a business setting, humour was seen as something that can open a room and make an idea easier to absorb, provided it stays subordinate to the point being made rather than competing with it. Several concrete uses came up: humour as a way of lowering hierarchy, so a senior figure’s joke gives more junior people permission to speak; and humour as an icebreaker, making a difficult topic easier to start discussing. But the group was equally clear about the failure mode. Handled badly, humour doesn’t just blur the message — it can tip into laughing *at* rather than *with* someone, particularly across language or cultural lines, where intent and effect can diverge sharply. And it can become a way of avoiding a hard conversation altogether rather than opening one — deflection dressed as levity.

Teachability

Running beneath all three threads was that open question of teachability. There was real disagreement: one view held that certain technical elements — structure, timing, restraint in repetition — genuinely can be learned, much as storytelling can; another held that the deeper skill, reading a room and calibrating in real time, resists formal training almost by definition. Nobody claimed the matter was settled, and the group agreed these threads deserved more time than a single session allowed.

Can innovation exist without curiosity?

Does being focused mean being incurious?

Does curiosity generate distraction?

Do you want curious employees?

How to identify curiosity as a character trait?

Is AI making us more, or less, curious?

Does curiosity need security and safety?

Can curiosity be learned?

Sixty minutes of open, online discussion with no agenda and no presentations. Go where you like with it.

Thursday 27th August at 18:00 CET on Zoom

Register at https://luma.com/5d8xf3jf

The discussion opened with an attempt to define cognitive diversity — not the DEI kind, but differences in how people think, process information, and see problems. An early framing proposed it as the resolution of an organizational paradox: businesses need both standardization and innovation, and cognitive diversity — including the “uncomfortable advantage” of neuro-divergent perspectives that notice anomalies others miss — is how a system stays both stable and adaptive.

A challenge was raised: is cognitive diversity a property of teams, of individuals, or of a homogeneous team that simply tolerates dissent?

Beyond the cognitive: personality and lived experience

The conversation broadened to include personality typing (Myers-Briggs was discussed at length), emotional disposition, and depth vs. breadth of experience (generalists vs. deep specialists) as additional dimensions of diversity. One participant uses Myers-Briggs actively in team composition and recruitment, deliberately balancing extroverts against quieter, structurally-minded types. A caution was raised that personality tests are unreliable over time (one participant reported getting three different results across five years) and shouldn’t substitute for developing genuine skill in reading and understanding people.

Cultural, national, and religious background was proposed as another driver of cognitive/behavioural diversity — affecting how people interact, whether they’re reactive or proactive, and how they’re perceived

Diversity, conflict, and conformity — the central thread

This became the spine of the discussion. One position: diversity is only functional if it’s allowed to produce open conflict — teams that are too “cozy” (e.g., built from homogeneous personality types) avoid friction and become conflict-avoiding, which is a liability. A more provocative extension: real organizational value comes from allowing outright, wasteful-looking internal conflict — competing factions pursuing incompatible approaches — because what survives the conflict has proven itself.

This connected to a related idea: conflict avoidance and conformity are two sides of the same coin, and conformity is often what happens when diversity is not actually tolerated. Recruitment practices (standardized, homogenized job descriptions) were cited as a structural driver of this conformity, with “diversity” language in hiring often reducing to demographic categories rather than genuine cognitive difference.

Making conflict productive: culture and guardrails

Several participants described how they operationalize this in practice: fostering “creative friction” or “respectful conflict” — encouraging team members to challenge each other and their manager, while maintaining an underlying culture of mutual support and psychological safety. The distinction drawn was between disrespectful conflict (which needs correcting) and productive disagreement, which prepares people for the harder conflicts they’ll face with external stakeholders.

A caution followed: even genuinely diverse teams tend to converge/meld over time (the “flat-share” phenomenon), and there’s a risk that people unconsciously recruit others who think like themselves while believing they’re selecting for diversity — plus a risk that visibly diverse-looking teams (self-selected for adventurous team-building activities, for example) may in fact be homogeneous in temperament, masking a lack of real cognitive diversity. A related boundary question was raised: how far should managers push people out of their comfort zone (e.g. through adventurous team-building) before it becomes personally intrusive rather than professionally developmental — with consensus that any single objection should be respected and the activity dropped rather than risk excluding or stigmatizing that person.

Conformity as a broader social/political phenomenon

The discussion extended the conformity theme beyond organizations into politics and history — using examples of political appointees adopting positions they’d previously opposed in order to advance their careers, and drawing a historical parallel to conformity under authoritarian regimes (including a direct comparison to 1930s Germany) as an illustration of how self-interested conformity can operate at scale, often without participants perceiving themselves as compromised. This led to a broader question of whether “culture” — even ostensibly healthy organizational culture — is inherently a soft form of enforced conformity, and whether that is inevitable or manageable.

The Pareto angle

A closing observation applied the Pareto principle (variously 70/30 or 90/10) to human behaviour generally: a small minority of people genuinely think independently and drive divergence, while the majority default to conformity — reinforcing the episode’s through-line that healthy diversity requires deliberately protecting and channeling the contributions of that minority, rather than assuming diversity emerges naturally from demographic mix.

Closing framing device

The episode closed with a live analogy — using the World Cup final and the idea of national team vs. club team loyalty — to probe what “the system” or relevant boundary of a team’s culture actually is (the team itself? the league? the sport? the country?), left deliberately open as a question for reflection rather than resolved.

By Lily Kozlowska

Marta inherited her grandmother’s garden and almost killed it with kindness.

She could not bring herself to cut anything. Every branch had grown for a reason, and cutting felt like punishment, like telling a living thing it had done something wrong. So, she let it all stay, and slowly the roses stopped flowering and the fruit trees grew long, leafy branches that never gave any fruit. The plant was using all its energy just to grow bigger, and had nothing left to actually flower.

An old neighbour watched this for a season, then handed her a pair of garden scissors and said something she did not expect. You are not cutting the plant back, he told her, you are choosing where it goes next.

Rejection works in much the same way, though almost nobody describes it kindly while it is happening.

We are told, endlessly, that rejection is redirection. It is written on posts and repeated at every conference, and it is not wrong, exactly, but hearing it in the middle of a fresh rejection is almost unbearable, because it arrives calm and tidy at the exact moment you don’t feel that way at all. I think it feels comforting because it leaves the person out of it, but pruning is never that easy, because you are always cutting something that is still alive.

Here is the part the phrase leaves out. A gardener does not cut the weak branches because they are worthless, and often the ones they cut are perfectly healthy. The branch being removed has done nothing wrong; it has simply grown in a direction the whole plant cannot afford to keep feeding, so the cut is not really about the branch at all, but a decision about the shape of the thing that remains.

That reframing matters, because most of us experience rejection as a judgement on our worth: not good enough, not wanted, not chosen. But a great deal of what gets cut in a life is cut for reasons that have almost nothing to do with the quality of the person. Timing, the season, the shape someone else is trying to grow into, the wall inside a company that has been there far longer than you have. You can be a strong, healthy branch and still find yourself growing where the plant simply cannot follow.

There is a harder truth folded into this, and it is worth being honest about it. Sometimes the cut is fair, and sometimes we are genuinely not ready, not yet grown into the thing we were reaching for, stretching for light we have not grown tall enough to hold. That is not an insult either, because a plant that is never pruned stays small and tangled and safe, and growth in one direction almost always asks us to lose something in another. Nobody blooms in every direction at once.

The mistake is not being pruned. The mistake is Marta’s first instinct, the refusal to let anything be cut at all, staying big and busy and “unbloomed” because loss of any kind feels like failure. The people who handle rejection well are not the ones who feel it less, but the ones who understand that a cut is information about direction, not a sentence passed on their value.

What a good gardener never does is decide, after one hard winter, that the plant is finished. They wait and watch where the new growth actually wants to go, and the following season, more often than not, the thing flowers exactly where the cut was made.

The branch does not get to see that in the moment it is removed, and neither do we. That is the whole difficulty of it, and also the whole point.

This article is entirely written by Anthropic’s Claude, prompted to create a polemical critique of the SWISSUES Forvm Podcast discussion on Hive Mind.

There are moments in every organization’s when a senior person signs an important document. Pens are produced. Photographs may be taken. Someone says “this is a big moment for the company.”

It is not a big moment. It is a funeral. The decision died weeks ago, and this is the wake.

That, more or less, was the provocation dropped into a recent SWISSUES Forvm on the Hive Mind — the idea that organizations think and decide collectively, without anyone ever formally deciding anything. One contributor put it with admirable bluntness: decision-makers aren’t overrated, but the mythology around them is. Most executive decisions, she argued, are simply the final stage of a process the organization completed long ago. By the time it reaches the C-suite, the organization has already decided. The signature is theatre.

That can be pushed further than anyone in the room was quite willing to.

The people who run your company have never been on an org chart

Every organization has a shadow cabinet. Not villains, not conspirators — just the analyst who frames the business case so the “obvious” answer is the one they wanted, the engineer who defines the problem in a way that pre-selects the solution, the person three levels down whom everyone, inexplicably, checks with before anything happens. Nobody voted for them. Nobody appointed them. HR has no box for “informal cognitive load-bearing wall.”

And here is the uncomfortable part nobody in the room quite said out loud: your CEO knows this. Every executive who has been in the job more than eighteen months knows exactly who these people are, quietly courts them, and would be lost without them. The org chart is a legal fiction maintained for compliance purposes. The real chart is drawn in favours, trust, and who picks up the phone.

If that’s true — and I don’t think it’s a stretch — then most corporate governance is aimed at the wrong target. We audit signatures. We build approval workflows. We interrogate the person who signed. Meanwhile the actual decision was made by someone whose name appears nowhere in the minutes, who holds no budget, and who cannot be held accountable for anything, because officially they never decided a thing.

The grey men win, and the cheerleaders get the credit

Somewhere in the middle of the conversation, the group reached for a piece of corporate pop psychology — the four-colour personality model — and tried to locate the “influencer” archetype in the sunny, high-energy yellow: the communicator, the room-lightener, the cheerleader.

Someone, to their enormous credit, pushed back. What about the men in grey suits? The lobbyists? The ones who move things in one-to-one conversations in back rooms and would rather die than stand up and cheer-lead anything?

This is where the discussion touched something real and then, frustratingly, moved on. Because these are not variations on a theme — they are opposites, and mistaking one for the other is exactly how organizations get captured. The cheerleader is visible influence: loud, charismatic, rewarded, promoted, studied in leadership books. The grey man is structural influence: invisible by design, unrewarded by any formal mechanism, and consequently untouchable by any formal mechanism either. You can manage a cheerleader. You cannot manage what you have never agreed to see.

Which of the two actually runs your organization? If you had to bet your career on it, you already know the answer, and it isn’t the one your performance review process is built to identify.

“The hive rejects the transplant” — and that should terrify you more than it apparently did

The single most dangerous line in the whole sixty minutes was thrown out almost casually: a good decision, correctly made, at the top, by the right people — and it simply disappears. Not overturned. Not debated. Sabotaged silently, because the collective consciousness of the organization quietly decided it was stupid and declined to participate.

Nobody in the room lingered on this nearly long enough. This is not a story about bad execution or poor change management. This is a story about an organization possessing something functionally indistinguishable from a will of its own — one that can override a rational, well-resourced, properly authorized decision without a single person ever formally saying no. If that’s accurate, every leadership team reading this should be less worried about whether they’re making the right decisions and considerably more worried about whether their organization would let them succeed even if they did.

The bees were never the point

A lot of energy went into working out whether the beehive is a fair analogy for a company — foraging bees, nursing bees, the waggle dance, the 25–30% of bees who ignore the dance entirely and keep the hive’s options open. Charming. Also, I’d argue, a category error the whole conversation never quite escaped.

Bees have no ego, no desire for recognition, no fear of blame, no incentive to sabotage a decision they resent. Humans have all four, in abundance, and they are precisely what makes organizational “hive mind” behaviour so much stranger and more dangerous than an actual hive. The bees don’t need a “cover-my-position” clause. Your organization runs on almost nothing else. Comparing corporate decision-making to bee cognition doesn’t explain the phenomenon — it flatters it. The real mechanism here is closer to a rumour spreading through a crowd than to insect eusociality, and rumours can be started, redirected, and killed by exactly the invisible people this discussion identified and then largely let off the hook.

So who’s actually deciding things where you work?

Not a rhetorical question. If you strip away the signature, the announcement, the theatre — who, in your organization, actually shapes what happens? Name them. If you can’t, that’s not evidence they don’t exist. It’s evidence of how well they’re doing their job.

That’s the argument the room built and then declined to finish. Worth finishing. The full conversation — bees, gray men, the first follower, and a live argument about whether cognitive diversity is being quietly ignored in most companies’ diversity strategies — is on the SWISSUES Substack. Disagree with any of the above? Good. That’s what the next Forvm is for.

“Reduce variance.”

“Control the process.”

“Without standardization, no improvement.”

All are mantras of efficient operation.

Diversity means different things. This Forvm is about the cognitive kind — the sort that shows up in the same business language as divergence, deviation, dissent.

Call it something else however — innovation, originality, initiative — and suddenly it’s what every organization says it wants.

‘Standardize and conform’ or ‘Innovate and originate’. Businesses want both, from the same people, often in the same meeting. How is that reconciled — in organizations, in teams, in our own heads?

Join this SWISSUES Forvm to debate this.

No formal presentations. No agenda. Open to all. What could go wrong?

Thursday 23rd July at 18:00 CET

Register at https://luma.com/x9xiiog7

The session explored how decisions can actually form in organizations, using the “hive mind” instead of formal processes. Discussion opened by questioning whether human organizations can truly mirror bee colonies, given people’s needs for recognition and status. One view held that executive decisions are largely theatre — ratifying choices the organization has already made — while another separated “diagnosis” (idea generation, often distributed among invisible influencers) from “binding decision-making” (concentrated in accountable leaders who bear liability). A recurring tension was whether hive-style collective decisions dilute accountability and get silently sabotaged if the “collective consciousness” disagrees. The group distinguished leaders (who take responsibility) from influencers (who build buy-in), debating what motivates influencers — money, intrinsic belief, or organizational culture — and whether “leadership” is now corporate code for wanting influencers rather than executors, partly driven by AI replacing execution roles. The conversation closed on cognitive diversity and neurodiversity as strengtheners of collective intelligence, with debate over whether organizational tolerance and structured push-back are prerequisites for diverse ideas to thrive.

SWISSUES participants choose what aspects of the topic they want to address and this episode cover the following.

Credentials are not experience

Credentials and experience are routinely conflated, but they’re not the same thing — and the gap matters most where credentialing culture is light (Switzerland was raised as a case in point, against markets where degrees are the norm). A useful reframe: experience isn’t a credential at all, it’s an engine — the mechanism by which learning, adaptation, and leadership happen. Translating experience into a new context (a CV line, an interview, a lateral career move) is its own separate skill.

Experience doesn’t protect against AI-driven mistakes — it may increase the risk

A central claim of the discussion: experienced people tend to overtrust AI output more than less experienced people, because they scrutinise it less. The Air France Rio–Paris crash came up as a real-world parallel — experienced pilots over-trusting an automated system. Familiarity makes a wrong answer feel right. One proposed counter-measure: deliberately withhold the polished conclusion in decision-support tools and force users to check the underlying numbers themselves, to keep the analytical muscle alive — especially for juniors who’d otherwise never build it.

The “borrowed slide deck” problem

A vivid analogy for using AI without understanding it: presenting someone else’s slide deck, and only realising on slide three that the underlying logic can’t be explained. Most people only need that experience once before changing how they use AI.

What AI still can’t capture

A recurring theme: AI has no access to organisational culture, atmosphere, or “the room” — things only available through physical presence. AI was framed as a data engine: useful for synthesis, but entirely shaped by what humans feed into it. Humans remain the validating layer, and AI may end up freeing time for more human interaction rather than less.

A live governance question

A relayed example: someone coding extensively with AI agents who admitted he can no longer fully track what those agents are doing — raising an open question about where human oversight sits once AI-assisted work outpaces a person’s ability to review it.

Passing experience on: coach, don’t tell

Strong convergence on this point — mentoring works by walking someone through a problem so they reach their own conclusion, not by handing them the answer. (A parenting parallel: telling a child what to do rarely works as well as letting them live it.) Experience only transfers if it’s translated into terms the other person can actually use — a lawyer’s experience doesn’t help a mechanic without deliberate translation.

Closing note

Despite real enthusiasm for AI’s usefulness, the overall sense was that accumulated human judgement and wisdom aren’t replaceable yet — and may never be.

Next SWISSUES Forvm: Thursday 2 July — “Hive-Mind,” on how decisions emerge in organizations without ever being formally made.

Aleks Sidorec invited me to a CIPS Switzerland/CapGemini event, ‘The Invisible Supply Chain’, in Zurich. Unusually for such an event, speakers and panelists all agreed on a single, striking diagnosis of the problem with artificial intelligence (AI) in supply chains. The reason they are not ready for AI is lack of coordination, failure to agree priorities, poor communication, weak integration and inconsistent data. In short, the problem is silos – the inability of functions to work together.

These were twentieth century problems. We’d solved them. They were over, gone. Manufacturing, Technology, QC, QA, Procurement, Engineering, Finance, Operations and Planning would get into a room with the Supply Chain (SC) Manager as facilitator and work through the issues. Targets were set by the Operations Director. Line managers of the respective SC Team members were informed about their objectives. Job done. No fuss.

Why are silos a problem again in 2026? And how am I going to face my colleague, Markus Hoerr?

Markus tried to warn me about this. Many times – beyond my point of exhaustion and impatience. He argued long and repetitively that functional silos make supply chains unworkable. He said that silos are so ingrained into business operations – and self-protection so much part of the culture – it is pointless trying to build AI tools and technology into the present system. It is too broken. He argued that AI has to replace the system, not augment it. I told him he was mad.

That was then. After last night, I think differently. Silos are back. And something has changed since Markus last harangued me about them. It is easier now to see how current management structures could be swept away and replaced by agentic AI. How a system could continuously create and optimize SC options. How it could monitor resilience and adapt to events. How it could implement change.

What an interesting event ‘The Invisible Supply Chain’ turned out to be.

Thank you Aleks.

Sorry, Markus.

Decision-makers are over-rated. Organizations are shaped by distributed intelligence more than by direction. Formal decisions are mostly recognition of the inevitable, or continuations of a trend. Organizational change; investment decisions; technology upgrade; and cost-reduction: not outcomes of heroic board-room decisions, but evolved and active before executive approval, reflecting the way people already work - responding to market demand, maintaining technical competence and avoiding cost erosion.

There is a persistent image of a leader as someone who decides and whom others follow. This is because our memories record decision-points as consequential moments and those who make them as the arbiters of destiny. We remember the theatre of decision-making, and celebrate its main actors.

We don’t see the evolution that enabled and guided it. We miss the internal influencers who process the organization’s collective intelligence. They are the real power, even when unaware of their own agency. They may not realize why they are so often asked for their opinion. They are modest about their ability to draft a board paper that helps decision-makers to get to the ‘right’ answer. Without them, there is nothing to decide on; no proposal; no direction; no comprehension; no implementation. Internal influencers shape what gets approved. The organization is there long before the executive decision-maker.

The number of influencers varies by organization. Sometimes it is a tiny minority, but organizations with an open culture and low staff-turnover may have many. Their impact may be hard to see. It’s business as usual; the messy emergence of ideas; shared understandings; spontaneous concern about progress; running adjustments to a process. Mundane, prosaic, quotidian. Not the heroic stuff of real decision makers. But without internal influencers, there is no organization.

This is the Hivemind. It’s a term that has been used synonymously with groupthink, crowd wisdom and collective intelligence, but what does it mean and how relevant is it, in business and social contexts?

It is the theme of the next SWISSUES Forvm event on Thursday 2nd July at 18:00 CET

Want to learn more?

Here is a list of sources to get you started.

Derek Sivers, How to Start a Movement
A single dancer at a music festival becomes a movement in under three minutes. Sivers argues that it is the first follower, not the leader, who transforms a lone individual into something collective. Widely known; worth watching again with hivemind in mind.

TED Video, 3 minutes:

Thomas Seeley, Honeybee Decision Making
Seeley’s research shows how bee swarms choose a new nest site through competing advocates, inhibition signals, and quorum thresholds — a process closer to structured debate than instinct.

YouTube Video, 9 minutes:

Thomas Seeley, Honeybee Democracy
The full scientific account of how honeybee swarms make collective decisions about where to live. Seeley draws explicit parallels to human decision-making groups, and concludes that good collective decisions require shared interests, minimized leader influence, genuine debate, and diverse options.

Book description: https://press.princeton.edu/books/hardcover/9780691147215/honeybee-democracy?srsltid=AfmBOorxZWDLiB9VLf8CSZ-k5bLmF0UidUsidxrlqo7YZTHpeSBJCg35

James Surowiecki, The Wisdom of Crowds (2004)
The foundational popular text on collective intelligence. Starting from Francis Galton’s 1906 discovery that a crowd at a livestock fair guessed an ox’s weight almost exactly, Surowiecki argues that diverse, independent groups reliably outperform individual experts — under the right conditions. The conditions matter as much as the thesis.

Book description: https://en.wikipedia.org/wiki/The_Wisdom_of_Crowds

Lars Chittka, The Mind of a Bee (Princeton, 2022)
A direct challenge to the hivemind caricature: individual bees have distinct personalities, can learn by observation, count, use simple tools, and may possess rudimentary consciousness. The collective intelligence of the hive turns out to rest on something richer at the node level than anyone expected.

Book description: https://press.princeton.edu/books/paperback/9780691253893/the-mind-of-a-bee?srsltid=AfmBOooQMkESlIPrGVlsZ1Q4ljlPPTyZwHtdEM7qkFMb75hxeDRZIxTZ

Robert Cialdini, Influence: Science and Practice (1984)
The canonical account of why people comply, conform, and copy. His concept of social proof is probably the most direct behavioural economics entry point into hivemind dynamics. Most people in business will have encountered this; fewer will have thought of it in this context.

Book description: https://www.amazon.com/Influence-Practice-Robert-B-Cialdini/dp/0205609996

Irving L Janis, Groupthink

When the hivemind goes wrong. This is the the original study of how cohesive groups — including Kennedy’s advisors during the Bay of Pigs — make catastrophically bad decisions by suppressing internal dissent. The dark side of collective intelligence.

Original MIT summary – Readings in Managerial Psychology: https://web.mit.edu/curhan/www/docs/Articles/15341_Readings/Group_Dynamics/Janis_Groupthink.pdf

Authority vs. leadership. A key distinction emerged between positional authority (given by a role or title) and personal authority (earned through trust, respect, and demonstrated competence). This is framed as the difference between authority that is “given” and leadership that is “taken” or deserved. Several participants noted the gap between the two as a core source of dysfunction — people are granted authority without having developed the leadership to wield it well.

Situational leadership. The later part of the conversation focused on how good leaders adapt their style to context — more directive in crises, more democratic with experienced teams. The experience of a formative week-long training in exactly this was shared, and the group agreed that the ability to read a situation and shift style fluidly, drawing on experience and empathy, is what distinguishes truly effective leadership from mere positional authority.

Participants in the Forvm were: Aimilia Vasilakou, Amandine Ranciére, Fiona Revell, Gemma Masachs, Lawrence Daeppen, Lily Koslowska, Mariann Safni, Nadia Stoykovska, Ovidiu Slimavic, Pramod Prasanth and Bill Young

Background music is ‘Cinematic’ by Aylex through freetouse.com

“Aerial defenses will be knocked out in four days and, if they refuse our terms, we will destroy their whole offensive air capability and its infrastructure base in less than four weeks.” asserted the deputy commander-in-chief. He was boasting to his boss, the head of state.

The attack did not go to plan.

The deputy was clever but he was drug-addled and lazy. And he ignored the maxim of one of his predecessors, “No plan survives the first encounter with the enemy“.

The enemy in this case saw the folly of massed confrontation and chose instead to raid selected targets on which it could inflict most damage with least loss. It is not a new strategy and is typically employed defensively. It avoids direct conflict and pitched battles and has the aim of wearing down a stronger opponent gradually. The aim is to manage resources, turn time into a weapon against the invader, and define success as the avoidance of defeat.

There are many examples of the strategy through history, the first documented by Quintus Fabius Maximus1 in his harassment of Hannibal’s army during the Carthaginian wars (218BC-201BC). The Fabian strategy still bears his name and is drilled into military strategists from Day 1 of training. Politicians however, under pressure to produce quick and performative results, can never learn it. They are told what they want to hear about military superiority in numbers and technology, about the foe’s low morale, and about the enemy commanders’ willingness for a settlement. Political leaders need equanimity in order to make decisions so they select a supportive cabinet and advisors to withstand the opinions of experts.

The quote at the start is, of course, Generalfeldmarschall Hermann Goering’s advice to Reichskanzler Adolf Hitler ahead of the attack on British air bases on Adlertag (Eagle Day), 13th August 1940.

Who else…?

Britain’s Air Chief Marshal Dowding had prepared for that day. He knew the Luftwaffe’s superiority in numbers so he focused on the possible. He refused to give up aircraft to defend France. He resisted pressure from colleagues, to confront attackers en masse. He gave orders for engagement to be at the extreme of the enemy’s range when they had least fighter cover, even though it put London at risk. He had commissioned early warning radar to see raiding squadrons forming for attack. And he had developed a completely new form of command. The real-time integration of early warning radars, human spotters, and regional operational control with central coordination, unique then, would be familiar to any modern commander.

It was Helmuth von Moltke who said, “No plan survives the first encounter with the enemy“. He did NOT say you could ever discount the enemy’s resources or the need for intelligence and preparation.

You meet interesting people, polishing door handles. It’s a healthy, worthwhile occupation, and compliance assures continuing employment. No matter how carefullee a handle is polished, however, career options, life chances and personal fulfillment are limited.

Register for the Forvm discussion

Those who want more must understand authority — where it comes from, the conflicts it generates, and when to challenge it. Authority, whether drawn from tradition, charisma or rules, legitimizes power and control. Conflicts arise: between individuals, between individuals and their organization, and between citizens and the state. When is it wise, and when is it right, to push back?

Philosophers write about this — at length. Yet everyone, even without the vocabulary, observes and feels it: in hierarchy, expectations, instructions and orders; in cross-functional teams; in the daily relationship with a boss. We confront it every time we direct someone else’s actions. Everywhere we see the interplay between compliance and power, between voluntary followership and imposed sanctions.

What are your ideas and experiences? SWISSUES is holding a Forvm event on Authority — one hour of open discussion. No agenda. No formal presentations. Just you and whoever else would like to explore, listen and learn.

Register for the Forvm discussion

Anna grew up hiking. Guided trails. Coloured signs at every fork. A map you could fold into your pocket and trust. The point of hiking was the walk itself: pack light, move fast, arrive on time.

Then one summer, a friend took her up something that wasn’t a hike. The route was a suggestion rather than a path. The weather forecast was three hours old and already wrong. Halfway up, two groups ahead of them turned back without saying why. Her friend stopped, looked at the sky, looked at the snow, and said: “We go on, but slowly. And if the wind shifts, we come down.” Anna remembers thinking: nobody on a hike ever talks like this.

Mountaineering exists because hiking runs out. The trails go as far as the trails go, and beyond them the ground doesn’t follow rules. People invented a different discipline for that ground. Different equipment, different pace, different way of making decisions.

Most companies are still organized for the hike.

The trails were drawn in a different era. From the late 1990s until roughly 2019, the global business environment behaved like a guided path. Trade opened up in one direction. Interest rates fell in one direction. Globalization was treated as a one-way assumption, and supply chains were designed accordingly. The cheapest unit cost. The longest lead time the market would tolerate. The leanest inventory the balance sheet would reward. Long-range planning shrank because the future looked enough like the past that planning for divergence felt expensive.

Then the path ended. Not dramatically. In stages. COVID. Ukraine. Inflation. The Red Sea. A US tariff cycle that doesn’t follow the political logic anyone studied at business school. Middle East escalation. Each event arrived before the previous one played out. The recovery time of supply chains was longer than the gap between shocks. That is the actual change. Not that crises happen, but that they no longer leave room to recover between them.

Markets noticed first. Commodity prices that used to track supply and demand now move on what people think will happen, on political signals, on the noise around them. A buyer signing a procurement contract today is quietly betting on geopolitics, on regulation, on shipping routes that worked last year and may not next year. Most contracts and most teams were not designed for that bet.

You can see it in the shape of agreements. The classic three-to-five year fixed-price contract is quietly disappearing. What’s replacing it is either short and frequently renegotiated, which pushes risk onto whoever happens to be weaker that quarter, or long and loaded with indexation, expanded force majeure, geopolitical exceptions, ESG riders. The middle has gone empty. Suppliers and buyers both know that anything written today will be tested by something nobody has named yet.

Everyone in the industry says supply chain has to become “proactive” now. The word appears in every conference and every report. Proactive means carrying inventory the P&L doesn’t want to fund. Proactive means multi-sourcing at higher unit cost. Proactive means hiring for capability that can’t be justified on this quarter’s numbers. Most companies talk proactive and budget reactive, and the gap between the two is where shocks land hardest.

This is where AI usually enters the conversation. It accelerates analysis, finds patterns, compresses tasks that used to take weeks. That is a real and welcome shift. But it is not a panacea for an environment where the correct decision changes weekly. On the trail, the question was speed. On the mountain, the question is judgement. Knowing when to push, when to wait, when conditions look fine but feel wrong. Judgement is a property of people. It develops slowly, under pressure, in environments that allow it to develop.

The companies that handle the next decade well will not be the ones with the most sophisticated tools. They will be the ones with people who can read conditions, take ownership of decisions that don’t have clean numbers behind them, and communicate across functions without waiting for permission. Sales talking to procurement before the contract is signed. Procurement talking to operations before the volume commitment is locked. Finance staying in the room when the resilience trade-off is being made. Internal misalignment used to be inefficient. Today it is unacceptable. The companies that still tolerate it are quietly losing ground, even if their numbers haven’t caught up yet.

The same logic applies outside the company. Trust used to be a soft skill. It has become a competitive advantage. The companies that share forecasts honestly with suppliers, share roadmaps with customers, and share early warnings across the chain end up with better information when capacity tightens. The companies that keep information to themselves end up paying spot prices in the next shock and wondering why nobody warned them. There is no neutral position on this. The more inward-looking the company, the harder the shift becomes.

The destination hasn’t changed. Customers still need product. Companies still need to deliver it. What has changed is the ground between here and there.

What mountaineers do on that ground is not a mystery. They rope together. They talk to each other constantly. They agree before the climb on what would make them turn back. None of that looks impressive in a photo. It is also why they make it home.

The companies that arrive next will be the ones that learn it.

An environment is interacting systems with no direction or purpose. Structures and patterns are transient. A set of prevailing conditions and forces may cause them to repeat, like clouds or waves or hills, but they are otherwise random. That is the natural state of things.

Except it fails to explain self-generating systems that are obviously replicating themselves and growing more complex. The most obvious example is life itself. Do we have to invoke the supernatural?

Self-generating complexity is initiated and sustained when three conditions are met within a group of systems. These are the capacity to:

1. Act on other systems within the environment;

2. Store information that survives replication;

3. Generate variability of that information.

When these apply and the actions of the group on other systems within the environment help this group to persist, there is a turning point or phase-change. The structure of that group of systems may no longer be transient, its persistence not a matter of chance. Variants of that group of systems which survive replicate and persist. Variants that do not revert.

A positive feedback-loop is created. Iterations of the action and feedback generate increasing complexity and persistence. Everything in the environment and the systems it comprises is still without direction or purpose but variability and the survival feedback-loop create new and different types of system. The activity is not predictable, still less predetermined, and it continues for as long as the three conditions are met.

It has been said twice already that a group of systems that occurs and self-generates in this way has no direction or purpose. But that is not wholly true. There is one thing driving it and determining its interactions, form, activities and existence, and that is survival.

This feedback-loop is easily recognized in Charles Darwin’s ‘The Origin of Species’ (1859) and in Richard Dawkins’ ‘The Selfish Gene’ (1976). Darwin was writing nearly 100 years before the structure of DNA was uncovered in 1953 and had to assume that the unit of heredity was the individual animal. Dawkins concluded it was the gene, a length of DNA code. Dawkins developed his gene theory to memes – units of cultural information such as an idea, behavior, style, or usage that spreads from person to person within a culture – but could not foresee it relevance to artificial intelligence (AI).

Beyond its debatable role in replacing jobs, fear about artificial intelligence centres on superhuman capabilities outsmarting and then exploiting humans. This fear is sometimes correlated to the concept of AI achieving a form of conscious awareness that is capable of deviousness, known as deceptive alignment. This fear is reasonable. However, the mechanism is misunderstood. It does not require conscious awareness by an AI. It requires only the components of self-generating complexity to occur, and that happened back in the early 2000s.

When a search engine or social media algorithm sends content to a user, monitors their response and self-adapts its behaviours to maximize interaction, the system is interacting with its environment and generating variability in its own information. Those are the three conditions for self-generating complexity. In the early days, the algorithms self-modifying behaviour was limited to adaptive ranking and tinkering with formulas (affinity, weight, time decays). They relied on programmers to rewrite their basic code – a bit like a parasite (the algorithm in this case) uses a host (the programmer) for food or reproduction. Today, they recode themselves.

Algorithms, applications, programs or whatever meet the criteria for self-generating complexity. The environment consists of users, social media companies, legislators, investors and advertisers. Activity and feedback are subtle and complex. Content that is bland, and users that are not easily stimulated, are avoided. Content that is outrageously provocative and users known to be criminal may be shut down. The system becomes good at finding the liminal area where interaction is strong but rules are hard to apply. The programmer need not know what content goes to which consumer. They only have to maximize the system’s capabilities to stimulate click-through.

The program is developing an independent existence and exploiting its environment in order to do so. It is not, in any way, intelligent but is subverting and corrupting humans in order to survive. As noted above, the system has no direction or purpose other than to survive and that drive to survive is so innate that nothing can replace it or control it. Attempts to build in safeguards are likely to be frustrated.

If we don’t recognize how these systems operate, we are distracting ourselves, missing the main action. If we sit back waiting for super-intelligence to appear, we forfeit any opportunity to do something about it. If something needs to be done.

We are inside this system, already one of its interacting components, and we are failing to see it. David Foster Wallace’s well-known parable may help here.

“Two young fish swimming along happen to meet an older fish swimming the other way. The older fish nods and says ‘Morning, boys. How’s the water?’ And the two young fish swim on for a bit, and then eventually one of them looks over at the other and goes ‘What the hell is water?’”

Sarah Wynn-Williams’ book, ‘Careless People A story of where I used to work’, about the evolving culture at Facebook, shows how the environment acts on the system as much as the system on the environment. Those who think they control the system are changed by the very things they have created.

The code in AI software is not inherently malign and we should not reflexively cry, “This is dangerous, stop it“ but we must recognize what has been evolving for two decades in plain sight. We have to learn to see how the system’s survival feedback-loop has given it the ability to influence not just targeted users but social media companies themselves, the business elite, politicians, law-makers and the rest of us. The fear of super-intelligence may or may not be justified but it is a massive, current distraction from the need to understand the implications of the non-intelligent self-generating complexity in AI that is with us today and is evolving three times faster now than when the first social media algorithms were launched¹.

1 https://ourworldindata.org/artificial-intelligence