Similarly to what some others have written, my attitude toward LessWrong is that it would best thrive with this model:

1. Embrace the Eternal September.

If LessWrong is successful at encouraging epistemic and especially instrumental rationality, people who have benefited from the material here will find less value in staying and greater opportunities elsewhere. LessWrong doesn't need to be a place to stay any more than does a schoolhouse. Its purpose could be to teach Internet users rationality skills they don't learn in ordinary life or public school, and to help them transition into whatever comes next after they have done so.

Since culture is always changing, to best aid new waves of people, the Sequences will need to be scrapped and crafted anew on occasion.

2. Aim lower.

Eliezer had motives in writing the Sequences in the way he did, and he also had a very narrow background. It has often been noticed that the demographics here are absurdly skewed toward high IQ people. My presumption is that our demographics is a consequence of how things like the Sequences are written. For example, Eliezer's supposedly "excruciatingly gentle" introduction to Bayesianism is in fact inaccessible for most people; at least it was difficult for me as a high-but-not-very-high IQ person with (not-recent) years of statistics training, and I pointed friends toward it who simply gave up, unable to make progress with it. A new Sequences could do well to have multiple entry points for people of different backgrounds (i.e. abandon the programmer jargon) and ordinary IQs.

3. Extend higher.

If we want to keep longtime participants from moving on, then we have to give them additional value here. I can't give advice here; I feel I've already learned more theoretical rationality here than I can effectively ingrain into habit.