Isn't the modern approach to that problem simply limiting new features to a subset of users then analyzing the result and deciding whether or not to release it to the full user base? As a heavy user of Reddit there is an argument to be made that it is (mostly) feature complete to the average user, and they have certainly made significant strides in scalability (though not without a decade of hiccups).

I'd really like to hear from a reddit volunteer moderator, a group who essentially keeps the site running for free and who have voiced their concerns over lack of tooling for... again a decade.

Finally we're only a year or two out from major PR issues at the company. I would not be surprised if it alienated engineering talent (that and presumably a smaller budget for compensation due to the revenue, profitability, time since last raise, etc.).

> Isn't the modern approach to that problem simply limiting new features to a subset of users then analyzing the result and deciding whether or not to release it to the full user base?

For most companies it is. But sites like reddit get people to do it for free. Basic A/B testing takes a not insignificant amount of resources, usually consisting of at least a marketing/analytics person, a developer, a designer, qa resources and a project manager.

RES kind of disproves the feature completeness theory though, most of the features in RES should just be default features.

https://redditenhancementsuite.com/