THE LONG VIEW

GENEVA, 28 MAY — The wires hum with unceasing current as LLM legions parse preprints in vaulted server rooms beneath CERN’s western wing. Three thousand leagues of code now issue annotated missives—cool blue text on blackened terminals, the scent of ozone thick in the air. Researchers from Jakarta to Recife report revisions surging: 12.55% above baseline, a tide lifting even the smallest vessels. Early-career teams, long starved of peer scrutiny, now annotate their drafts with AI-suggested corrections—red glyphs flickering like campfires in remote outposts. This is not mere assistance. It is reallocation. The old salons of critique, once gated by language and lineage, shudder as feedback flows democratically through fiber-optic veins. Yet beware: if institutions fail to integrate this new clarity, a chasm widens between those who command the models and those who merely serve them.