[Question] How effective are tulpas?

Edit: After further consideration, I’ve concluded that the risk:reward ratio for tulpamancy isn’t worth it and won’t be pursuing the topic further. I may revisit this conclusion if I encounter new information, but otherwise I’m content to pursue improvements in a more “standard” fashion. Thank you to everyone who posted in the comments.

If you don’t know what a tulpa is, here’s a quick description taken from r/​tulpas:

A tulpa is a mental companion created by focused thought and recurrent interaction, similar to an imaginary friend. However, unlike them, tulpas possess their own will, thoughts and emotions, allowing them to act independently.

I’m not particularly concerned whether tulpas are “real” in the sense of being another person. Free will isn’t real, but it’s still useful to behave as if it is.

No, what I’m interested in is how effective they are. A second rationalist in my head sounds pretty great. Together we would be unstoppable. Metaphorically. My ambitions are much less grand than that makes them sound.

But I have some concerns.

Since a tulpa doesn’t get its own hardware, it seems likely that hosting one would degrade my original performance. Everyone says this doesn’t happen, but I think it’d be very difficult to detect this, especially for someone who isn’t already trained in rationality. Especially if the degradation occurred over a period of months (which is how long it usually takes to instantiate a tulpa).

A lot of what I’ve read online is contradictory. Some people say tulpas can learn other skills and be better at them. Others say they’ve never lost an argument with their tulpa. Tulpas can be evil. Tulpas are slavish pawns. Tulpas can take over your body, tulpas never take over bodies. Tulpas can do homework. Tulpas can’t do math.

Then there’s the obvious falsehoods. Tulpas are demons/​spirits/​angels (pick your flavor of religion). They’re telepathic, telekinetic, and have flawless memories. They can see things behind you. There’s not as much of this as I expected; most of the claims are at least plausible. Some guides are cloaked in mystic imagery (runes, circles, symbols), but even they usually admit that the occult stuff isn’t really necessary.

It does seem like there are clear failure modes. Don’t make a Quirrel tulpa. Don’t abuse a tulpa. Make sure to spend enough time tending to the tulpa during creation, etc etc. And everyone seems to agree that tulpas are highly variable, so a lot of the contradictions could be excused. On the other hand, if tulpas were really so useful, wouldn’t the idea have spread beyond niche internet forums?

Perhaps, but perhaps not. The stigma against “hearing voices”, plus people’s general irrationality, plus the difficulty… those seem like powerful forces inhibiting mainstream adoption. Or maybe the entire thing is comfortable nonsense and the only reason I find it remotely plausible is because I want to believe in it.

Given that the guides mostly say that it takes months and months of hard work to create a tulpa… well, I’d rather not waste all that work and get nothing. And it only gets worse from there. An out-of-control dark rationalist tulpa that fights me for mental and physical control sound absolutely terrifying.

Most people seem to agree that the chance of that happening is basically zero unless you deliberately try to do it. And the potential gains seem at least as potent. Being able to specialize in skills seems absurdly overpowered, especially if we each get natural talents for our skills (which some people claim is what happens). A minor drop in cognitive resources would probably be worth it for that.

So, if you have a tulpa, please chime in. What’s it like? How do you know that the tulpa isn’t less efficient than you would be on your own? Was it worth it? Does it make your life better?

And if you don’t have a tulpa, feel free to comment as well. If I get a hundred LWers saying that I’ve been suckered by highly-evolved memes, that’s pretty strong evidence that I’ve made a mistake.