This piece combines relatively uncontroversial points with some justification (“we’re not near the compute or data efficiency limit”) with controversial claims justified only by Steven’s intuition (“the frontier will be reached suddenly by a small group few people are tracking”). I’d be more interested in a piece which examined the consequences of the former kind of claims only, or more strongly justified the latter kinds of claims.
This piece combines relatively uncontroversial points with some justification (“we’re not near the compute or data efficiency limit”) with controversial claims justified only by Steven’s intuition (“the frontier will be reached suddenly by a small group few people are tracking”). I’d be more interested in a piece which examined the consequences of the former kind of claims only, or more strongly justified the latter kinds of claims.