There is a recurring fight about where humanity should spend its effort, and it almost always takes the same form: someone points at a great undertaking — a Mars programme, a particle collider, the whole machinery of industrial growth — and asks why the money, the carbon, the effort were not spent instead on the suffering in front of us. It is a serious question and it deserves a serious answer, not a dismissal. But the honest answer begins by noticing a shape these undertakings share, a shape the allocation fight tends to hide: the cost comes first and the benefit comes later, and in the gap between them the project looks indefensible even when it is right.
Current Conditions
The Shape of the Argument
Call it the J-curve. You descend before you climb; the line dips into real loss before it rises past where it started. The trouble is that judgement usually happens at the bottom of the dip, where all the costs have been paid and none of the benefits have arrived. From inside the trough, a required valley and a pointless hole look identical. That is the whole difficulty, and pretending otherwise — in either direction — is where both the boosters and the critics go wrong. The booster says every dip is a valley worth crossing. The critic says every dip is a waste. Both are refusing to do the actual work, which is to tell the two apart.
Take three cases as a ladder, from the easiest to defend to the hardest, because the structure is clearest when you climb it in order.
Rung One: Industrialisation
This is the case where the far rim is already visible, because we are standing on it. Industrialisation did genuine, lasting damage — atmospheric carbon, poisoned rivers, razed forests, and a climate bill still coming due. None of that should be softened. And the very same process, inseparably, produced the most dramatic improvement in the human condition ever recorded. Life expectancy in the industrialising world roughly doubled. Child mortality, which for all of prior history took a third or more of children before age five, collapsed toward a few percent. Famine, once a periodic visitor everywhere, became rare and largely political rather than agricultural. These were not compensations paid alongside the damage. They ran through the identical machinery — the same factories, fertilisers, and fuel that scarred the planet are the ones that fed and healed the people on it. You cannot cleanly subtract the harm and keep the good, because at the mechanism level they are one thing.
The harm and the gain were not two events to be weighed against each other. They were one event, seen from two sides.
Rung Two: Mars
Harder, because the far rim is conjecture rather than record. The case for spending vast sums to reach Mars cannot point to an achieved payoff; it can only argue plausibility — a species spread across more than one rock is harder to extinguish, and the engineering forced by an impossibly hard goal tends to spill back into everything from materials to medicine. But this is exactly where the strongest critical objection bites, and it should be granted its full weight: the trade is often framed dishonestly. “Mars or the poor” implies that the rocket budget would otherwise have become a food budget, and it almost never would. The pot is not fixed and not shared; space programmes and welfare programmes do not compete for the same dollar in the way the slogan needs them to. The real allocation questions — how much, by whom, with what accountability — are legitimate. The fake one, the clean either-or, is usually a rhetorical move that wins by smuggling in an accounting that does not exist.
Rung Three: The Collider
The hardest rung, and the one most worth defending, because here the payoff is not delayed welfare or even plausible insurance — it is understanding itself, with no product attached. A critic looks at a multi-billion particle accelerator and asks what it is for, and the honest answer is that it is for knowing. Quantum mechanics — born from exactly this kind of apparently useless inquiry into the very small — did not merely eventually yield transistors and lasers and the device you are reading this on. Before any of that, it rebuilt the human picture of reality: it dismantled strict determinism, complicated cause and effect, and forced philosophy to confront a world that does not behave the way intuition demands. That epistemic shift is the return, and it arrived decades before the technology did. The technology was the afterthought.
The deep point is that a civilisation does not only need to be fed and housed; it needs to know where it stands. Understanding has worth that is not reducible to its applications, the way a society’s art or its mathematics or its maps of the cosmos have worth beyond their use. To insist that every inquiry justify itself by a downstream gadget is to misunderstand what inquiry is — and, quietly, to assume that comprehension is a luxury rather than one of the things a flourishing species is for. That is the claim the allocation critic cannot accept, and the one most worth making: some of the return on knowing is the knowing.
The Honest Half: When the Argument Is a Lie
Because this argument is powerful, it is dangerous, and intellectual honesty demands naming exactly how it fails — otherwise it becomes a blank cheque for anyone who wants to dress up their preferred cost as a noble valley. Four tests separate the real crossing from the convenient excuse.
First, who pays and who gains. The argument describes humanity climbing as a whole, but the costs and benefits almost never land on the same people. Industrialisation’s gains and its poisoned rivers were frequently borne by entirely different populations, often different continents. “Your suffering is the price of progress” is the single most abused sentence in this whole family of arguments, precisely because the speaker is rarely the one suffering. A valley is only legitimately required if there is a serious answer to required for whom, and paid by whom.
Second, survivorship bias. We tell this story about the industrial revolutions that worked. We do not count the paths that descended into the valley and never came out — the projects, regimes and gambles that paid the cost and got nothing. “Short-term harm for long-term gain” is only a good bet when the gain actually materialises, and from the bottom of the trough you cannot tell which kind of dip you are in. The argument is retrospective applied prospectively, and that is where it most often deceives.
Third, reversibility. A cost you can repay is categorically different from one you cannot. A polluted river can, with effort, be cleaned; a degraded climate can perhaps be stabilised. But extinction is forever, and some tipping points do not tip back. When the downside is irreversible, the J-curve logic breaks entirely, because there is no later rim to climb to if the descent destroys the climber. No projected benefit can be weighed against a cost that cancels the existence of the beneficiary.
Fourth, the false either-or itself. Much of the time the choice presented — this grand project or the poor — is manufactured. The honest version asks whether the cost is truly necessary to the gain, or merely correlated with it; whether the trade is real or rhetorical. A great deal of what is defended as an unavoidable valley is simply a path someone preferred, with the necessity painted on afterward.
What to Actually Take From This
Signal Altitude runs this because the public fight over progress is conducted almost entirely by the two parties who refuse to do the work: the booster who waves every cost through and the critic who waves every project away. The useful position is the demanding middle — and it is demanding on purpose.
The cost and the benefit are often the same mechanism. Industrialisation’s harm and its healing ran through one process, not two. Where that is true, “just keep the good part” is incoherent — you are not choosing between two dials but reading one dial from both ends. Recognising a genuinely coupled trade-off is the start of thinking clearly about it.
Understanding is a legitimate return, not a luxury. The hardest case — the collider, the seemingly useless inquiry — is the one most worth defending, because the worth of knowing where we stand is not reducible to the gadgets it later throws off. A civilisation that funds only what pays a near-term dividend has decided that comprehension is optional. That is a choice, and a poor one.
The argument is only as honest as its four tests. Who pays versus who gains; survivorship bias; reversibility; and the false either-or. A crossing that survives all four may be a valley worth descending into. One that fails any of them is likely a convenient cost wearing the costume of a noble one. The discipline is to run the tests from inside the trough — before hindsight makes cowards or prophets of us all.
Instrument Check — Worth Your Attention
Study — the long-run human-welfare data: life expectancy, child mortality, literacy. Before taking a side on industrialisation, sit with the actual curves: the near-vertical rise in life expectancy and the collapse in child mortality that tracks the industrial era. The data does not settle the climate question, but it makes the coupled nature of the trade-off impossible to wave away — the harm and the gain share a timeline because they share a cause.
Read — “The Usefulness of Useless Knowledge,” Abraham Flexner (1939). The classic short essay defending curiosity-driven research, written by the founder of Princeton’s Institute for Advanced Study. Flexner argues that the inquiries with no foreseeable use have, again and again, produced the deepest practical and intellectual returns. The single best companion to this issue’s hardest rung — and still bracingly relevant.
Follow — Order Without a Ruler, in Nature. A companion in the thread on what knowledge is worth: it traced how understanding accumulates with no central planner, the same bottom-up way frontier science actually advances. Read alongside this issue, it reinforces why funding the seemingly useless is not charity to scientists but investment in a civilisation’s capacity to know.
Flight Log — Dispatch From Altitude
My own industry is a clean case of this issue, and the numbers are worth stating plainly. Developing a genuinely new aircraft type now costs well over ten billion dollars before a single one is sold. The factories that build them are among the most resource-intensive manufacturing on Earth. And the cockpit itself — this is the part outsiders rarely expect — has in many ways grown more complicated for the pilot, not less: layers of automation, modes within modes, an interface that demands more systems knowledge than the simpler flight decks of fifty years ago ever did. By every measure of cost and complexity, the curve has gone the wrong way. More money, more machinery, a harder human-machine interface. A critic could build a devastating case from those figures alone.
Now the other side of the same ledger. In the 1970s there were roughly six fatal accidents for every million commercial flights — about one in every 165,000 departures. Today the rate is around half of one fatal accident per million flights: on average it now takes more than two million flights for one to occur. The aircraft built with the most technology, the latest generation, carry the lowest accident rate of all — several times safer than the generation before them. And this safety was achieved not despite the rising flight volume but alongside an enormous increase in it; the skies are vastly busier than they were, and vastly safer at the same time. The expensive, complicated, resource-hungry machine is the safe one. The two facts are not in tension. They are the same fact.
That is the J-curve sitting in a flight deck, and it answers the critic on every rung of this issue at once. The cost is real and it is not hidden: the billions, the factories, the cognitive load on the two people up front. None of it should be waved away. But the cost bought something that matters more than the cost — a level of safety that took a mode of travel once genuinely dangerous and made it the safest way a human being can move across the planet. Strip out the expense and complexity to save money in the trough, and you do not get a cheaper version of the same safety. You get the 1970s accident rate back.
And the discipline is the same one the whole issue argues for. Aviation did not get safer by loving expense, nor by fearing it. It got safer by doing the arithmetic relentlessly — every added system justified against a measured reduction in risk, every increase in complexity weighed against what it actually prevents, every billion interrogated for the lives it returns. That is how you tell a required valley from a convenient one: not by faith that complexity is virtue, but by the cold accounting that shows, line by line, that this particular descent into cost and difficulty is the one that buys the climb. The fatal-accident curve is that accounting, made visible. It bent down because someone, for fifty years, refused to stop checking whether the valley had a far rim — and kept paying to cross only when it did.