Healthy Competition

I wrote this up awhile ago as a shortform post on the EA forum [EA(p) · GW(p)]. My motivation to post it now was because of an interesting conversation contrasting https://roamresearch.com with lesswrong.com.

My conversation partner was excited about Roam as a multi-purpose tool for intellectual progress and collaboration, and listed some features they were considering that were moving in directions that were similar to directions the LessWrong team was considering. "Roam is gonna do all the things! Collaboration! Blogposts! It could replace google docs as a collaborative thinking tool"

[note: this was a friend who doesn't work at Roam, I'm not sure how concrete those plans are]

And I had a few flinch reactions of "Aww, LessWrong was gonna do all the things!", followed immediately by "okay every writing platform that tries to do all the things right-off-the-bat fails and probably both of the teams should be focused a bit more". Both of these were followed by a more interesting observation:

One of LessWrong's primary focuses is being an attention allocation platform. Whether it follows the simple "show latest posts, ordered by date" or a hackernews algorithm, or things like curated and recommendations, it's fundamentally aiming to be a place where centralized conversation of some sort happens.

There's a thing in EA/rationalsphere space where people notice that a thing isn't being coordinated on, and their first impulse is to build a coordination platform to fix it. And this is usually a mistake because it's real costly to get everyone to switch to a new platform, and if you screw it up you not only waste everyone's time but make them less likely to switch to the next platform that might actually be good enough.

I think it's somewhat dangerous (albeit necessary) that LessWrong is natural-attention-monopoly shaped, which makes it hard to directly compete with. I think this gives us something of an obligation to do a good job, and to enable things like GreaterWrong, and to be careful taking on too many different domains that we won't have the capacity to be good at.

Single Player

But there's something sort of nice and valuable about having other writing platforms whose primary focus is on the "singleplayer" aspects of writing/thinking/intellectual progress. (My underrstanding is that this is the general advice to startup founders trying to corner a market that requires network effects – start with something that doesn't require network effects)

Right now Roam is young, I'm not sure how serious their plans are for adding collaboration and blogpost-type features (this was mentioned to me second-hand and might have just been a "interesting idea to consider" thing rather than a "concrete plan.") But after some reflection I found it actually kind of reassuring that there were ways to build up competing platforms interacting with the same ecosystem, not via initially starting as similar products, but by starting from pretty different vantage points and then gradually adding various supporting social features.

Self Review. 

I still think this is true, and important. Honestly, I'd like to bid for it being required-reading among org-founders in the rationalsphere (alongside Habryka's Integrity [LW(p) · GW(p)] post)

I think healthy competition is particularly important for a (moderately small) constellation of orgs and proto-orgs to have in mind if they are trying to scale up and impact the world at large, while maintaining integrity. (i.e. the rationality/x-risk/EA ecosystem). 

I think this is one of the key answers to "what safeguards do we have against evolving into a moral maze?"

Meanwhile the piece is pretty short, which makes me feel better about saying "hey guys, please actually read this."

This post does not make a comprehensive case for its claims (in part because it's aiming to be short). I would definitely appreciate someone who has differing intuitions, or who thinks I'm missing something major, doing a substantive response. 

This topic seems very important!

Another potential consideration: In some cases, not having any competition can expose an org to a ~%50 probability of having net-negative impact, simply due to the possibility that a counterfactual org (founded by someone else) would have done a better job.

Note that many of these things can be accomplished with intra-organizational competition - you don't necessarily need a separate group entity for each set of ideas and behaviors, just an acknowledgement that there are different values and beliefs in play.

The question of how best to cooperate with partially-aligned agents, in a world with MANY agents that are less aligned than that, is important and under-modeled.

Nominating this post as much for the main body as well as Ray's top-level comment. I guess maybe this post is somewhat downstream of me, so it's not super surprising I like it, but I do think many many parts of the world could really benefit from more healthy competitions, and I've set many plans into motion that try to create more competition in ways that I think improves things quite a bit.

This seems basically right to me.

Research orgs benefit from having a number of smart people bouncing ideas around.

Probably most of (more of?) this benefit can also be unlocked by an ecosystem of multiple orgs in friendly competition who regularly talk to each other in ways that feel psychologically safe.

If you want to challenge a monopoly with a new org, there's likewise a particular burden to do a good job.

(This seems to depend on whether the job/project in question benefits from concentration.)