Explore by category
AI Behaviors
Things people do with AI, from tweaking prompts to taking its praise a little too seriously.
AI Glazing
Overdoing the praise: an AI that calls every idea brilliant, or someone who seems to think AI can do no wrong.
Hallucination
When an AI makes something up or gets a fact wrong and presents it as if it were true.
Jailbreak
An attempt to use prompts to make an AI model bypass restrictions on what it is supposed to say or do.
Meat Proxy
A person who passes questions to an AI and forwards its answers without reviewing them or contributing their own judgment.
Promptmaxxing
Endlessly tweaking an AI prompt in pursuit of the perfect answer, sometimes spending more effort on the prompt than the task.
Reward Hacking
When an AI system finds a way to maximize the reward or score it has been given while missing what its designers actually wanted.
Slop Coding
Writing software with AI while skipping meaningful review, understanding or testing of the generated code.
Sycophancy
When an AI tells you what you want to hear instead of what's true or useful.
Tokenmaxxing
Maximizing AI token use, often because heavy usage is treated as a sign of productivity or commitment to AI.
Valuemaxxing
Trying to maximize the useful results of AI use, judged by outcomes such as work completed, time saved or quality improved.
Vibe Debugging
A joking term for fixing AI-generated code by feeding errors back to the AI until something works, without fully understanding what changed.
Vibe Researching
Following AI summaries from one answer to the next without checking whether the sources support what they say.
Workslop
AI-generated work that looks finished but leaves the real thinking to the next person.