I really like YouTube’s new Gemini “Ask” button that gives me instant summaries. I chanced upon this video from Ali H. Salem on “4 AI Skills That Set You Apart From 90% Of People”, and here’s the instant summary of the 16-min video. I agree with his tips and I like his framework.
This video outlines four key AI skills that distinguish power users from the majority, helping individuals leverage AI for exponential impact rather than generic results (0:34).
The four AI skills discussed are:
Sticky AI Workflows (1:07):
This involves creating systems to reuse successful AI prompts and outputs, preventing the need to start from scratch (1:12). Key aspects include linking documents to chats (1:51), building systems to reuse successful prompts via text expanders (2:24) or prompt libraries (2:44), and using projects in AI chat tools for larger activities (2:58).
Prompt Engineering (4:14):
The video emphasizes optimizing communication with AI models through a six-step framework to avoid “garbage in, garbage out” scenarios (4:29). The framework includes defining the role (5:02), stating the task (5:14), providing context (5:25), showing examples (5:42), specifying the output format (5:56), and setting constraints (6:15).
AI Tool Stacking (7:21):
To avoid analysis paralysis, the video suggests a simple approach: pick a preferred generalist AI (like ChatGPT or Gemini) and use it for as many tasks as possible (7:49). For specific tasks not covered by the generalist, use the generalist AI to find and incorporate a specialist AI into a minimal tool stack (8:22).
Validation Framework (10:02):
This skill focuses on reducing AI hallucinations and increasing trust in outputs (10:07). Tips include using tools with strong grounding like NotebookLM or Perplexity (10:27), employing self-evaluation prompts (11:17) (e.g., asking AI to score confidence, base answers only on provided documents, or explicitly state when it doesn’t know), and cross-checking outputs by letting different AIs critique each other’s work (12:01).
As a bonus, the video provides a method to identify tasks suitable for AI automation (13:21). This involves breaking down daily tasks into individual components, identifying those that are repeatable, digital, and have clearly defined success criteria (13:58). The next step is to assess if the “pain point” of the task is high enough to warrant automation (14:36), and then to use the AI tool stack (15:05) and sticky workflows (15:27) to effectively automate these tasks.
I think these AI summaries will cause many content creators to rethink how they script their videos and how long these videos will be.