It's 2024, and forums and other alternatives for mass communication still don't work well. Everyone is complaining about the inefficiency of WeChat groups and putting up with Telegram's tens of thousands of unreads (or marveling at its smooth one-click reads); wanting to escape from fragmented social networks, but also lamenting the decline of blogs and personal websites and their fragmentation; bashing Q&A platforms for being spammy, and content-sharing platforms for being un-nutritious. In short, it's all getting worse. There's often a nostalgia argument that harkens back to the vibe among the elite cliques of seed users in the early days of the product.

What's not working well?

I've tried to summarize some aspects based on my own experience.

Difficulty in context understanding

Difficulty in tracking context and historical consensus in long discussions. When a user clicks on a thread with hundreds or even thousands of replies, he/she needs to read all the replies in order to understand the focus of the discussion, or whether there is a consensus on a certain assertion, or whether there is any stage of conclusion. Therefore some summary replies are often welcomed because he has done the painful process for the general public, and no one accuses them of dramatizing or complains that the summary posting takes the fun out of it, because it's not like a well-written piece of literature that really isn't fun to read.

Of course, there are some products in which special displays are given to some valuable replies, such as marking them as best answers and popular replies. The special typography takes it out of context and often makes understanding more difficult if it's not a very good straight-to-the-point targeted response.

Information cocooning

More and more products are using personalized recommendations and offering the ability to follow and block. The aim is to reduce users being inundated with topics they hate. Examples include Twitter / Tiktok and the like. Users reach a topic's content primarily through popular recommendations and feeds; unfortunately, they also apply algorithmic sorting in their reply system. Users miss out on some of the topic content, and they miss out on some of the replies in the topics they are interested in. This causes users to misjudge opinions, leading to misunderstandings and unnecessary conflicts.

Low-value responses

A portion of participants, having encountered the difficulties mentioned above, participate in the discussion without acquiring sufficient information, leading to off-topic, repetitive, misunderstanding, and even pointless arguments, thus further deteriorating the whole thread and making it more difficult for future participants.

Difficult to retrieve and utilize

There is a term for what is called a menstrual problem. It refers to a recurring topic that comes up every month. The viewer of the topic encounters the same dilemma as the participants of the discussion, and is unable to conveniently profit from previous discussions and has to start another thread.

Information overload

An overload of low-quality, low-value content makes it more difficult to find good, valuable content.

Dependence on community management

Given all the above experiences, many products have introduced administrators. For example, Baidu Tieba has an elder system, and Reddit has a democratic voting system. There are also review and reporting mechanisms and so on. Administrators actually create more uncertainty.

Will it work better?

AI seems to be very good at summarizing and inventorying, and the fact that there are teams working on related auxiliary products suggests that they have a similar view to me. In contrast, the likes of Github Issues and Twitter are still good, they've made a lot of detailed improvements, and they're more popular as a result. From that perspective, there are viable directions we can take beyond introducing AI, and that's redesigning.