.png)
4 months ago
3
Reddit and Hacker News threads about AI and LLMs are plastered with comments like “I asked ChatGPT to do this very simple thing, and it gave me this completely irrelevant answer.” The responses usually fall into two camps: “AI is overhyped” or “you’re prompting it wrong.” (Just this morning @karpathy posted “+1 for context engineering over prompt engineering. )
I find such comments baffling but believable. AI is composing symphonies, discovering new drugs, and beating humans at complex strategic games. Yet brilliant people are failing to make use of these tools for basic tasks, we’re looking at another potential division among us: AI literacy.
As The New York Times recently reported, AI chatbots are increasingly being used for mental health support during our ongoing psychological crisis. Millions more rely on these tools for everything from coding help to creative writing. Yet most of us don’t understand the fundamental limitation that shapes every AI interaction: how these systems handle conversational memory.
It’s vital to understand how to use AI effectively. This isn’t just about prompt engineering or knowing the right commands — it’s about understanding the fundamental constraints that shape every interaction. This article focuses on one of the most important and least understood aspects of…
