Data Collection

Since I started playing with LLMs I’ve been pretty against content generation. Picturing the internet filled with LLM generated nonsense composed of the same ridiculousness I would get reinserting Google translations over and over playing a linguistic game of telephone. With a perfect translation a translated text has its inverse (?). I assumed LLM content was low effort and shouldn’t be rewarded compared to something that took human time and effort to create and synthesize. Nonetheless, here I am creating content. In a nutshell, for a client I’m using their content to seed new content. To know more contact me :) My own feelings toward it have obviously changed since working on this project.

Reviewing some of the questions that came up are:

• Is previously generated content seeding new content acceptable?
  • I think it’s acceptable to use previously created content, obviously, with consent to generate new content.
    • It’s using the same tone, the content is activating weights, and the model can even regurgitate the seeded content to use in the new output.
  • Editorial review is required. Let’s be real, unchecked LLM outputs wandering the internet are irresponsible. We’ve seen enough cases already to know outputs can be a little unhinged.
    • Also what’s the difference between this and hiring someone less experienced?
• What about it becoming affordable for businesses who otherwise wouldn’t update their content or focus on it?
  • What I’ve realized is that small business might have the most to win from LLMs.
    • It drastically lowers the cost for things like this as an example. It’s decent, fact checked (or should be!), and can help provide; the initial blog structure, SEO improvements, guides, etc.

Does this take anyone’s job, no. Or at least I don’t think so. To all the doomers and AI gloomers if anything it’s created a job. This now has put something at the price point a small business can pay for. A business can now do content marketing in house.

One major remaining question for me is:

• What does this mean if everyone is doing it?
  • Being honest I’m currently not sure.