arXiv let physicists publish first, review later

On 14 August 1991, the automated repository hep-th@xxx.lanl.gov at Los Alamos National Laboratory received its first email submission. Paul Ginsparg had built it to let theoretical physicists circulate preprints instantly, a machine-age replacement for mailing paper copies to a circle of colleagues.

The system, now arXiv.org and hosted at Cornell, spread from one physics subfield across physics, mathematics, computer science and beyond, becoming the primary channel through which those fields communicate results, with journals persisting alongside for certification and archiving.

Move the bottleneck. Instead of waiting on editors and referees to distribute a finding, researchers deposit the paper themselves and the community reads it immediately: peer scrutiny arrives as commentary and citation after visibility, not as a gate before it. An automated server with light moderation replaced both the postal preprint circle and the journal's distribution function at near-zero marginal cost.

Moved the bottleneck: researchers self-deposit on an automated server, making distribution instant and free, so peer scrutiny becomes post-publication commentary.

Authors want speed, so they deposit immediately.

Free access draws readers and citations to the open copy.

Automation keeps marginal cost near zero.

Field-specific moderation preserves enough trust.

When an intermediary's main value is distribution, not judgment, a self-deposit commons can displace it; incumbents should identify their true gate.

From one email submission on 14 Aug 1991 to the default distribution channel of physics and math, now arXiv.org at Cornell. The template for every open repository since, SSRN, bioRxiv, medRxiv, and the wedge that forced journals to accept preprint-first publishing.