optimise this

Machine learning is not the danger. Optimisation without sufficiently broad objectives is. Every optimiser faithfully pursues whatever objective it is given while systematically ignoring everything excluded from the objective function. Markets optimise for return. Platforms optimise for engagement. Political campaigns optimise for victory. Bureaucracies optimise for self-preservation. Artificial intelligence merely accelerates a civilisational reflex that long predates it: sacrificing tomorrow for a marginal improvement today. We celebrate efficiency while quietly optimising away resilience, trust, wisdom, culture, meaning, and long-term persistence because none of them fit comfortably inside a quarterly report.

The current fascination with artificial intelligence risks missing the point entirely. AI is not the story; it is simply the latest instance. Steam engines, industrialisation, financial markets, mass media, social media, recommendation algorithms, machine learning, and whatever follows are successive expressions of the same underlying dynamic. The technology changes. The objective function scarcely does. Again and again, extraordinary advances in human capability are reorganised around extraction, concentration, acceleration, and advantage. We speak as though AI has become our master when, in reality, it is faithfully amplifying objectives civilisation selected long before the first neural network was trained.

Artificial intelligence will not destroy civilisation. Civilisation will destroy itself through the objectives it chooses to optimise, with artificial intelligence serving as the most powerful amplifier yet invented. That is the unpleasant little knot at the centre of this entire spectacle. AI is not our existential threat so much as our existential mirror. If we continue optimising for extraction, concentration, and short-term advantage, AI will not decide our future. It will simply help us arrive there much faster.

History has rarely been undone by a shortage of intelligence. More often, it has been undone by institutions that rewarded immediate advantage while systematically discounting consequence. There is an important distinction between ignorance and wilful blindness. Ignorance can learn. Wilful blindness often pays remarkably well. Civilisations seldom fail because nobody noticed the warning signs. They fail because acknowledging them threatened too many incentives, too many assumptions, and too many concentrations of power. By the time reality submits its invoice, those responsible have usually convinced themselves they were creating progress.

Civilisation cannot exist without optimisation. The question has never been whether we optimise, but what we optimise for. Money, markets, and technology are indispensable instruments, but disastrous masters. Intelligence without wisdom merely accelerates error. Continue mistaking accumulation for success, extraction for value, and efficiency for progress, and artificial intelligence will not end civilisation. It will simply become the clearest possible expression of objectives that were already consuming the future long before anyone thought to call them intelligent.