By 2017, online dating had become the single most common way American couples met, surpassing introductions through friends, family, work, school, and church combined. That is a structural change in the mating market with a roughly two-decade history, and its consequences for commitment are still propagating through the population. Michael Rosenfeld's longitudinal "How Couples Meet and Stay Together" survey captured the inflection point, and the line on his graph kept rising. The romantic ecosystem stopped running on local social networks and started running on national, then global, search engines.
A commitment market behaves differently from a search engine. Commitment is built on closing options, accepting friction, and tolerating ambiguity in a particular person. Apps invert all three. They open options, remove friction at the front end, and let users resolve ambiguity by swiping to the next profile. The medium is optimized for first contact, not the messy middle. The result is not that nobody pairs up — pairing happens constantly — but that the cost of staying single has collapsed while the cost of choosing has stayed roughly the same. When the outside option is cheap and infinite, the inside option has to be unusually compelling to hold.
Eli Finkel and his collaborators reviewed the science of online dating in 2012 and concluded that the algorithms' matching claims were largely unverifiable, and that the volume of available options tended to produce a "shopping" mindset that worked against the commitment frame. A decade later, those findings look prescient. Users describe partners as "options," compare them across slots, and feel the friction of choice precisely because the choice is real and reversible. The technology did not invent fickleness; it lowered the activation energy.
The long shadow falls in three areas. First, the duration of the search phase has lengthened. People meet more partners before settling, and the median age at first marriage has crept upward in every country where apps have penetrated. Second, the early-relationship attrition rate has risen. Couples that would have once moved from "dating" to "committed" through inertia and social pressure now stall because either party can re-enter the app market at no cost. Third, the meaning of commitment has shifted from a default state ("we are together unless something breaks it") to an active state ("we are together as long as we keep choosing it"). The active form is in some ways more honest, but it demands more cognitive and emotional labor, and it never lets either party rest.
The 2nd Law — Think — is the law most strained by this environment. Thinking well about a partner requires sustained attention to one person, the discipline to ignore the counterfactual, and the willingness to update beliefs about them based on accumulated evidence rather than fresh comparisons. Apps train the opposite habit. They reward rapid evaluation, surface-level pattern matching, and constant comparison. A user who runs the app's logic on their existing partner — "could I do better?" — has already left the relational frame and entered the market frame. The two frames do not coexist for long.
At a collective scale, the effect compounds. Each individual decision to keep options open is rational; the aggregate is a society in which fewer people pair, those who pair pair later, and those who paired earlier feel the gravitational pull of the market they left. Rosenfeld's data show that couples who met online are not less stable once married — the selection effects favor people who have done more searching and arrived more deliberately — but the population around them has changed. Friends are single longer. Singleness is no longer a temporary status. The infrastructure of coupled life — shared mortgages, shared schedules, the social weight of "we" — has thinned, and the thinning feeds back into the next cohort's decisions.
The shadow is long because the technology is sticky. There is no return to the parish dance, the office romance is increasingly suspect, and the bar scene was already declining before apps arrived. The apps are not an addition to the older venues; they are the replacement. A generation now exists for whom "meeting someone in real life" is the exception that has to be explained. That generation will set the next baseline of what commitment means, and the data suggests the baseline will continue to drift toward looser, longer, and more conditional pairings — with all the gains in autonomy and all the losses in continuity that this implies.