This post discusses our new approach to leveraging language models like ChatGPT and Claude and summarising tools like Perplexity to improve our reader experience. To be clear, none of the content on this blog is AI-generated. You can read our policy on AI content elsewhere on this blog.

Our purpose is to streamline our content to focus on our unique insights, arguments, and conclusions. To date, our posts have followed a basic structure as follows:

  • Introduce the topic
  • What brought us to this point? Often including a short story or insight
  • Definitions and existing theory overview
  • Our view on a topic
  • Conclusions
  • Next steps

The value we provide (we hope) is in our view on a topic, experience-driven insight and conclusions. The rest of the post simply provides structure to an argument.

We have to write the introduction as that provides an overview of why we believe the topic is important and what we would like to discuss. Obviously, we must present our view on a topic as that is the whole point. However, definitions and introductions to existing theories tend to be well-established. They are not controversial; in general, they are axioms.

We could (as we do now) link to the best definitions we could find on the WWW rather than clutter our main argument. The reader could then go and read that content if they wished. If, we could not find something we were happy with elsewhere, we wrote our version.

Now, we can ask Perplexity, ChatGPT or Claude for a definition or a short introduction to a topic. Deep learning-based tools will give a thorough and detailed response. If we have the capability (we do) to error-check that response, then why would we bother writing it ourselves?Why would we waste time (the search experience is awful) trawling the WWW?

So, to summarise our approach moving forward. In general, our posts will be shorter with links to Claude, ChatGPT or Perplexity output to provide context and background reading. We will publish our prompts and carefully check deep learning-based tool output before publication. The argument (the insight) provided in the post will be all ours.

Our objective is to deliver value and to provide the best reader experience. So, dear reader, if you do not like our new approach, please let us know.

This website uses cookies. By continuing to use this site you are agreeing to our cookie policy

Privacy policy