The death of independent content creators?

I recently tweeted about this article.

I’ve had a few conversations, both online and offline about this recently, so I thought I would write something to bring together a few of my thoughts.

I think it’s also worth mentioning this post I wrote a couple of years ago, as some of it is relevant to this discussion.

Why do people create content?

There are several reasons to create content on the internet, and everyone will have slightly different motivation, but here are some obvious reasons to do it.

  • Money. Some people and companies create content as a source of income. That income comes from advertising. The only way you get paid is if people visit your website or YouTube channel etc. Anything that reduces the number of people visiting your content directly affects how much money you earn.
  • Reputation. Some people use their content to build their reputation, or personal brand if you prefer that term. If nobody visits your content, then what is it really doing for your reputation?
  • Personal Reasons. I often say I write for myself. It’s how I learn stuff I need for my job. I recently put out three posts about ORDS. In all three cases I was writing something that I needed, and which I could point colleagues to. I will still do this if nobody else visits the website, but I have to admit the activation energy required to start a new post is extremely high when I think nobody will care about it. I will definitely write less often, and about fewer subjects, if I am just shouting into the wind. πŸ™‚

Where does traffic come from?

This will vary depending on the content, but the vast majority (98%) of my traffic comes directly from Google searches. That has been pretty consistent for many years.

I don’t know how many people now search on ChatGTP directly, or via Bing, but it would seem zero change to the old workflow of searching on Google will result in an AI answer, so it hardly matters where people are going to do their searches.

What is the problem?

Having said the number of visitors to your content is a big deal, anything that reduces the people visiting your content is a problem.

In approximately 2012 Google added Knowledge Panels to the Google search output. When you searched for something, the search results might include a Knowledge Panel at the top of the page. This was typically a quote from one of the top search results that might answer your question. I would often see quotes from my own posts in them. The Knowledge Panels might work well for a request for a specific command, but I often found them a bit crappy, and ended up clicking on the link to the site anyway.

In May 2024 Google started to include an AI Overview in the search results for some questions. Rather than being a quote from a website, this was a “generated” answer. The quality of these answers varies, but they are substantially better than the old Knowledge Panels.

How has this affected my website traffic?

Even though Knowledge Panels were crappy, I did notice a drop in the visits to some of my articles. If you are just searching for a bit of syntax and the Knowledge Panel happens to show it, there is no point visiting the link. The impact was noticeable, but not catastrophic.

The timing of the release of the AI Overview, and of course other AI search options, had a more dramatic affect on my website traffic. It’s about half what it used to be. If I were doing this for cash, it would be a really big problem, because a drop in traffic like that, along with the general reduction in advertising revenues across the industry is a big deal.

Are you sure it’s because of AI?

No. It could be a coincidence. A number of things are probably at play here. Things that spring to mind are…

  • The AI search thing generally, as well as the Google AI Overview already mentioned.
  • A reduction in the number of people searching for Oracle content generally. Most people I encounter have to do a variety of tasks, not just Oracle, so the amount of times they are searching for answers about Oracle tech is reduced.
  • People usually search for things they are working on. The lack of the regular on-prem release of Oracle 23ai means the total number of people searching for 23ai content will be lower. Content about newer versions of the database are always less popular for a while, but the delay to the release has meant 23ai has dropped off the radar for many of us.
  • I’m publishing less content. Once Oracle 23c/23ai Free was released I published over 60 posts about it. I’ve got another 15+ waiting for the on-prem release. The lack of the on-prem release has left me feeling rather deflated, and I’m struggling to be bothered to write much. With less new content hitting the site, there is a tendency to get lost in the crowd. I’m sure some of the drop off in views is because of this. I don’t know how much though.
  • Improved documentation. It’s still far from perfect, but the Oracle documentation has improved. I’m sure some people will click a link to the docs in preference to some random website.

Even though it is likely to be a combination of these things, the timing of the inclusion of AI Overviews in Google search seems like a smoking gun to me.

So what?

It’s obviously disheartening for existing creators to see their viewership dwindle, but my ego is not really the central issue here. I was chatting over DMs to someone about this and I said this.

“The bigger issue is what happens to the next generation of content creators? With a million AI generated sites, and nobody actually hitting your site, what are the chances of them bothering?”

If there is nobody stepping in to fill our shoes, what happens when we all retire, as many of us already have?

AI feeding on itself

If there are no new people stepping up to produce new content, or even worse the new folks just generate their content, what happens? Everything I’ve read suggests that AI feeding on itself causes a downward spiral in terms of quality. You only have to look at what the YouTube algorithm promotes to know that crappy AI generated content is saturating the platform now. I’ve read numerous stories of the percentage of bot traffic on social media sites going through the roof. Where will it end up? See Dead Internet Theory.

Conclusion

I really hope I’m wrong, but I don’t think the future is positive for independent content creators. I’m sure there will always be some, but they will be increasingly lost in the noise of AI generated slop.

I’m an old man shouting at clouds, and it won’t be my problem for much longer. πŸ™‚

Cheers

Tim…

PS. While I was posting this to social media I noticed Jon Dixon had written Why AI Doesn’t Give Credit to Bloggers? It’s worth a look.

PPS. Someone sent me a link to this video on Twitter. Very scary.

Getting inspiration from daily use (or not)

It’s been an odd few months on the content creation front.

When Oracle Database 23c/23ai Free was released I was no longer under NDA for the shipped functionality, and I pushed out a glut of version 23 content here. Since then I’ve been waiting for the full on-prem release so I can publish the remaining stuff, some of which was written 22 months ago. That’s nearly two years without being able to hit the publish button!

In that time I could have started doing videos to go with the existing 23ai posts, but after being burnt by the name change (23c to 23ai) I decided to hold back until the official on-prem release dropped. That has left me in a dead zone as far as content creation is concerned.

Proving functionality with daily use

The new features posts are pretty easy to write from a straight “how-to” perspective, but it’s hard to give true insight until you have been using products as part of real systems on a daily basis. There is that joke about “everything works perfectly in a PowerPoint presentation”. To a certain extent we could say everything works perfectly in a laptop demo. Until you can start running realistic workloads, it’s all a bit unknown.

Over time I’ll go back and revise posts as I learn more about the functionality, but that requires regular use in real situations.

Inspiration from daily use

In addition to proving functionality, real use will often inspire entirely new posts. In doing your job you come across things you wouldn’t have noticed by reading the docs or playing on your laptop. In some cases these types of posts may be more useful than the straight new features posts.

I guess it’s possible you might stumble on this stuff by accident, but the chances are a lot higher when you are using a product for real, and you have colleagues asking questions that approach things from a different angle.

Without daily use, the content feels a little lifeless.

My current state of apathy (#retired)

When I first got involved in the 23c/23ai thing I was really enthusiastic, churning out content and enjoying the process of learning new stuff. The stagnation caused by waiting for the on-prem release has made that seem like a distant memory. I feel totally apathetic about it all now.

I’m hoping they announce the on-prem release of 23ai at CloudWorld, and I’m hoping that will reignite the content creation spark in me, because at the moment I feel like I’m #retired. I need to start using this stuff in my job to actually care about it again.

Cheers

Tim…