Should we still write software engineering blog posts?

We initiated this blog with the realization that we were consuming a lot of content on the internet without actively contributing. As software engineers, both as a hobby and professionally, we have become accustomed to relying on search engines to find solutions and complete our software projects. Often, our queries led us to blog posts, Stack Overflow threads, or forum discussions where we discovered the answers we sought.

Hence, we naturally created this blog as a way to document our own software engineering discoveries, hoping to assist fellow engineers who may encounter similar challenges. Initially, it worked well, serving as a valuable resource for others. However, as we enter the era of artificial intelligence, the way people seek software engineering solutions is changing. Traditional search engines are being replaced by AI language models like Google Bard and ChatGPT, which provide direct answers. There is no longer a need to sift through blogs, forums, or personal websites to find solutions.

Given this new reality, one might question the purpose of continuing to write these concise blog posts that aim to address specific engineering issues. Initially, it may seem unnecessary. After all, why maintain personal notes when the answers are just a chat away?

However, upon deeper reflection, we realize that these blog posts likely contribute to the training data of these AI language models. Thus, the original intention of sharing collective knowledge remains valid, albeit in a more indirect manner without receiving any credit. Even for bloggers who do not monetize their content, visitor counts and occasional thank-you emails serve as valuable feedback, affirming that they are making a positive impact. Unfortunately, this feedback loop is broken in the new era.

If we were to cease writing these blog posts, it could potentially diminish the quality of the answers provided by AI language models over time. But if our only readership will eventually be AI training crawlers, what would be the point? While I'm not an expert in the field, it appears there may be an upcoming issue concerning the future training of AI language models. Perhaps we are currently witnessing the peak performance of these models, and there will come a time when blog posts like ours regain their utility.

Personally, I have not yet decided whether I will continue writing these posts. My motivation for doing so has certainly waned. Only time will tell.