Why Will Parallel Be a Larger System Than Twitter?

Parag Agrawal expects Parallel, over time, to be a meaningfully larger system than Twitter. He is proud of Twitter and calls it influential, and he still describes it as a small slice of public content. The opening he sees at Parallel is keeping more of the web from disappearing into closed systems. He gives Twitter the same mission: high-quality, distinct content in public, and direct access to people rather than only through intermediaries or closed groups.

The building lesson is when to rebuild. At Twitter the team rebuilt whole systems about every year because they were not ready for the next order of magnitude. Building for four years out meant shipping nothing, so the systems kept being evolved for the next step of scale. He says Parallel is running a similar set of lessons, including the move from cobbled models and heuristics toward end-to-end training, the arc he also names at Waymo. He wants Parallel to start closer to that later step instead of repeating every earlier one.