What makes another couple big steps like that inevitable in a short time frame?

Before the recent floodgates cracked open, AI research made only slow incremental progress for decades. Why couldn't we already be back near that rate of progress?