Most people get these 3 AI terms wrong

The video begins by tackling hallucination, the most widely misunderstood concept in AI. The author points out that many users mistakenly treat hallucination as purely negative, but it is actually the flip side of AI creativity — without this creative latitude, the AI couldn't write text or offer novel perspectives. The practical takeaway: assess the nature of your task. If you want creativity, accept that hallucinations may occur; if you need accuracy, enable web search to add grounding constraints. This balanced perspective helps users make smart choices about whether to limit the AI's creative freedom depending on the task at hand.

This chapter compares the AI's context window to a finite whiteboard. When a conversation gets long enough, the whiteboard fills up and the AI can no longer process new information or starts forgetting earlier context. The author gives a highly actionable recommendation: the moment you notice the AI giving strange or logically inconsistent answers, the simplest fix is to start a new chat window. This clears out the accumulated memory load and restores the AI's accurate reasoning ability, keeping interactions high-quality.

The video closes by explaining tokens — the smallest units of text the AI breaks input into for computation, and the underlying reason free AI tiers have usage limits. Every token represents compute cost, which means leaner, more efficient prompts can dramatically cut token consumption. By optimizing your prompting strategy, you get more precise results and can handle more tasks before hitting free-tier limits. Understanding this mechanism is essential for professionals who want to maximize AI output within a limited quota.