My AI content policy
Published on
People instinctively understand that calculators are convenient, but they also recognize the risk: relying on them too much can erode mental math skills. The same applies to AI tools for text generation—if an AI does the writing, what’s left for the writer? What’s your role then? You wouldn't go to a gym to watch a robot push weights. This goes to the heart of why we write at all. Jordan Peterson summarises this well - you write to learn to think!
You need to learn to think
Pair that with Paul Graham's classic essay Putting Ideas into Words, "no one who hasn't written about a topic has fully formed ideas about it."
AI can obviously support and encourage writers. I also recognize its potential dangers - like overwhelming the internet with slop- junk that neither informs nor entertains.
However, I'm concerned about disappearing resources like Stack Overflow becoming inaccessible just because more advanced tools are available. If we're using AI-generated content responsibly, it can be a valuable tool for tidying up and simplifying writing tasks. I code a lot (for now!), and regularly encounter ideas and tricks that I could share, but the effort involved in polishing it up for publication stops me from doing so. I plan to leverage the ease of AI to help with that. It will hopefully be a useful reference, at least for me. There's also this idea that there's lots buried in the latent space - that the AI models know a lot of this stuff, from the way they've been trained on e.g. stack overflow content, but as we head towards a dead internet, a lot of the good resources will have disappeared. I don't want to add to this problem. To that end, I will do my best to sanity check everything. I don't want to be adding junk to the world.
For transparency, I’ll clearly mark posts where I’ve used AI assistance. Technical write-ups and simple explanations may lean more on AI for organization and clarity, while more complex topics—like the future of humanity—will be written entirely by me, with AI limited to minor edits like spellchecking.
This policy may be revised as needed, but my goal is to set a transparent standard for using AI-generated content in this context. Feedback is welcome.
(Side note: While this post is about AI-generated content, I won’t tag it as such, since I want it to appear in lists that filter out AI-generated posts.)