Keep this in mind: software generated content is not content, it is only data. If we think of AI as software, we must treat AI generated content as data.
In the case of Large Language Model, the data you see is nothing but patterns that have high probability in that particular model, plus the prompts generated by you and hidden prompts generated by the system. The plausibility is just an illusion.
This is a hand edited partial list of AI generated content (data) for me to read offline on Kindle or for my local bots to analyze. You are welcome to read, analyze, and share them anyway you like.
Always assume AI generated data are nonsense unless you’re convinced some of them make sense to you.
Critical reading in the age of AI generated content
Do not mistake a sophisticated echo for the truth
Sense, nonsense, and common sense
Passionate sympathetic contemplation
Fiction is harder than nonfiction
Philosophy in the age of compute
How can we improve Nix in the age of LLM?
Scaling curve of resource utilization
Skill: Nix Flake Migration Engineer
Can a "Less Powerful" Model Analyze a "Superior" One?
The Architecture of Insight: Optimizing AI for Theory over Action
The Watchman and the Actor: Synthesis, Semantics, and the Architecture of Insight
Perfectly Polished, Completely Empty 光鲜之壳，空洞之芯
The Spleen Algorithm | L'Algorithme du Spleen | 忧郁算法