I’ll have a longer post about this, but we’re publicly stating LLMs and their apps should be used in all the wrong ways for all the wrong reasons. The media hype is around search, image generation, and video generation, but all of those things pale in comparison to the utility they open up.
This blog post is being written by me, entered into Claude on a virtual OS, and uses a WordPress MCP to publish directly to blog. It drops in the categories and tags appropriate for this post based on the instructions I gave the project. This was done in seconds. I never entered my admin. I didn’t draft the text in the editor. I simply pasted this post into my LLM, it recognized it had write access, and boom: Now you’re reading this.
A second Golden Age of blogging isn’t happening because content can be churned out. It’s because the promise of Web 2.0 and dynamic content is now truly democratized.
If you have a few minutes, Ed Zitron posted a good article about LLM’s being overvalued, and I’m not arguing with his points (especially on IP theft). We need to start saying that the biggest impacts they’re having is on data management, portability, interoperability, and development. Mass-producing content when audiences want raw authenticity from a writer, publisher, or company is a flop. This is one of the most exciting times in tech and creative technology I can remember. We need to guard it against the engagement baiters, crypto grifters, IP thieves, and VCs pumping up valuations only to have the business models fail.
If you’re building something big or small with code assistants and LLMs, please DM me on Bluesky or Threads. Would love to hear about it.
Edit 1: LLMs default to using classic editor formatting. If you want them to use the block editor (I did), you need to say so in the project instructions.
Edit 2: The tags are all wrong. I need to figure out why ‘paper’ was added as a tag. I had to manually restore them.