Guth's Original Inflation

Guth's original proposal used a first-order phase transition with tunneling. The bubbles of true vacuum nucleated within an expanding false vacuum background. The problem: bubbles expanded relativistically, never collided, and so did not produce the homogeneous reheating needed to start standard Big Bang cosmology. This was the "graceful exit problem," fixed by slow-roll models (new inflation) shortly after.

The reason Guth's framework opened the door to eternal inflation, even though Old Inflation itself died from the bubble percolation problem, is that the false vacuum geometry contains the seeds of self-reproduction. A region trapped in the false vacuum expands exponentially with Hubble rate H set by the false vacuum energy density. Bubble nucleation removes patches of false vacuum at a finite rate per unit four-volume, but as long as the nucleation rate per Hubble volume per Hubble time stays below a critical value, the false vacuum domain grows faster than bubbles can eat it. New false vacuum is created exponentially while bubbles only remove it at a sub-exponential rate. The unconsumed false vacuum keeps inflating forever, even as our own bubble formed and stopped inflating internally. Vilenkin (1983) made this geometric argument explicit, and it became the structural foundation that Linde's chaotic inflation later generalized to scalar-field potentials with metastable plateaus.