Blog

Home | About | Archive | RSS

Rethinking where to post on the internet

One of the side effects of the move to social media has been the loss of depth in online communication. I’m using the term “social media” as a reference to any form of online media where other users of the site can are able to interact with and offer feedback on something posted by a user. This is a broad definition that includes Facebook, Twitter, Bluesky, Reddit, Hacker News, and really anything with an active comments section, including YouTube and some blogs.

This is in contrast to the early days of the internet. In the mid-1990s, if you wanted to share information on the web, you created an html document and uploaded it to a web server. You wouldn’t typically get any feedback. You might track hits to your site, but that didn’t tell you anything about how many people were reading a particular article or what they thought of it. You shared content with others because you thought it might be useful to others.

Then came reader comments and analytics. If you were interested in those things, it would drive your behavior. Rather than write what came naturally to you, in the style that you thought was best, you tried multiple styles and used the one that you believed would result in the best numbers. This represented a major change, because it was the reader that determined how you wrote.

Over time that continued to evolve. Rather than posting full pages of information about a topic on your website, written the way you think it should be written, you make short posts on a small group of sites (small is a feature, not a bug, because the fewer sites the more users of each site). The numbers we’re supposed to optimize for today are upvotes, shares, or interactions, depending on the platform. What’s rewarded is posting things others want to hear, largely determined by the entertainment value or the distance from your own views. Something that is deemed to have “made you think” is a single sentence that required less than a second to digest. It made you think in the sense that it wasn’t something you already agreed with, and it wasn’t purely entertainment.

This has really started to bother me in recent months. The time I spend posting online does not provide any value. It’s turned into a race to figure out how to be the most average person on the board. I read a comment blaming LLMs for the loss of diversity in online content, but as I got to thinking about it, I realized the diversity was already gone before the public release of ChatGPT. It’s absolutely true that LLMs are doing a kind of averaging over all their training material, so you can think of LLM output as an average of the training data. Yet that’s what we already had when humans communicated online. Social media is a game you win by being some best version of average.

But it’s worse than just lack of diversity in the things you post and the words you use to post. To play the game, you have to reduce the depth of your post to whatever gets the best numbers, and due to resource constraints (other users of a given site have finite time to give feedback on what’s effectively an infinite supply of posts) very few of them have an incentive to invest time in something with depth. Twenty minutes of thought before giving their feedback is too much.

I regret how I’ve spent the last decade or more of the time I’ve been online. Writing up my expertise, thoughts, and observations on a topic of interest is something I should do my way, by my rules. I should be writing the things I want to read. My writing should be the result of a real thought process. The fact that others might not care is irrelevant if I’m doing the writing for my benefit.

The good news is that we haven’t lost what we had in the early days of the internet. I can still post something on my website in static html, with no comment section, without sharing it with anyone else. Visitors to my website can read what I’ve written, but I have no knowledge of how many people have read it, and if they did, whether they thought it was a masterpiece, stupid, or pointless. I don’t have a reason to collect those numbers if I’m not writing to optimize them.

I will post this now. I will do so having no way to know if even a single other person has ever read it. My work on this post ends with me posting it on my site, as opposed to posting it to social media, where the process that matters (the feedback) begins when it’s posted.

PS: I wasn’t going to do this, but I changed my mind. Here’s the comment that motivated me to write this post:

Bonds are more volatile than stocks over 20-year time horizons and underperform stocks during periods of high inflation because of the duration risk. Because of this, I’m more comfortable owning 100% stocks in retirement.

I wanted to fire off a reply to that comment asking how that argument (which is definitely not true as a blanket statement about bonds) implies someone should hold 100% stocks in retirement. Are they saying they don’t plan to consume over the next twenty years?

Old me would have spent a few minutes crafting a careful reply. It’s not like the commenter would have seen my reply and said “Oh, I guess my comment didn’t make sense. I was wrong.” Replying to the comment wouldn’t have provided me with any value, so I didn’t do it.



True minimal theme