I've never found E-Prime very useful because I've always been rather empiricist in philosophy outlook and aware that I should always be able to reduce my statements down to something referring to my observations, and I suspect most LWers would not find E-Prime useful or interesting for much the same reason. But I could see it being useful for normal people.

I personally think of it as a tool, not unlike "lint" for C programmers. It shows things in your code (speech) that may contain errors.

To put it another way, if you know how to spot what isn't E-Prime in a sentence, you can dissect the sentence to expose flawed reasoning... which actually turns out to be a pretty useful tool in e.g. psychotherapy.

Whether or not RET (rational-emotive therapy) and CBT (cognitive behavioral therapy) directly derive from General Semantics and E-Prime (or their logical successor, the linguistic meta-model), modern psychotherapy is all about map-territory separation and map repair.