The biggest problem as I see it is loss of members and lack of talented new ones. I'm willing to bet if you plotted the histogram of user activity, you'd see virtually all of the posts and comments coming from a very small number of members. The 'Top Contributors' section in the side panel probably contains most of them, and has been relatively stable in composition over the past two years. If the first step in instrumental rationality is to identify reality in an objective way, then we have to realize this site has become an echo chamber for a small few, with their own vocabulary and system of thought which is incompatible with the outside world. The barriers to entry (reading the sequences, reading the 'seminal' comment threads, etc.) are too high for most people. HPMOR offers a pathway for new members to start reading the rationality materials, but it doesn't equip them to meaningfully contribute.

Another thing is that there is only so much 'low-hanging fruit' lying around. In terms of general rationality, we've covered most everything. There are only so many threads you can have about biases and logical inconsistencies. The topic has become quite stale.

I like AI because at least it offers the possibility for new material to arise every once in a while, leading to useful discussions. Other people might like other topics. I have a huge list of topics whose implications are quite relevant to the art of rationality and so would be quite compatible with the goals of this site:

• Thermodynamics
• Neuroscience
• Neural Networks
• Social organization & Forms of government
• Human sexual dynamics

The problem is, the types of people likely to be knowledgeable about these things probably have no idea this site exists. And if they do it is unreasonable to expect them to learn the required information to be 'on the same page' as this site's core users. And this is very bad, because it means that this site's users will attempt to foray into these topics themselves without any help from actual experts.

I don't know what the solution is. Maybe it's already too late to do anything.