Video summary

5 Hacks To Use ChatGPT So Well It’s Almost Unfair

Main summary

Key takeaways

Technology

Summary of technological concepts, product features, and “hacks” (from auto-generated subtitles)

The speaker argues that most people use ChatGPT as a basic helper, but claims you can “optimize” it into something more powerful by building persistent context, using agentic parallelism, applying adversarial critique, and running structured learning/review workflows. The video outlines 5 strategies plus multiple prompt/tuning techniques.

1) The Clone (persistent memory + custom behavior)

Problem

  • New chat sessions start from zero context (compared to how memory fails in late-stage Alzheimer’s).

Fix: a 3-step framework (ESP)

  • E = Extract

    • Use Voice Mode (spoken interview) as “chief of staff.”
    • Prompt it to interview you for 20–30 minutes about:
      • background
      • tone
      • strategy
      • blind spots
    • Instruct the assistant: be messy and don’t self-edit to capture raw material.

  • S = Synthesize

    • Feed the transcript back into ChatGPT.
    • Ask it to produce a summary of who you are and how you think.
    • Then store it via ChatGPT Settings → Custom Instructions.
    • Add behavioral rules, such as:
      • format data as tables
      • avoid “walls of text.”

  • P = for Projects

    • Use Projects as persistent workspaces to store assets (e.g., resume/PDFs/data) and apply project-specific instructions.
    • Emphasizes separating domains (e.g., consulting vs. YouTube vs. health).

Claimed outcome

  • Turns ChatGPT from “temporary chat” into a more persistent brain aligned to you.

2) The Swarm (parallel agent/assistant workflow)

Core idea

  • Humans and ChatGPT are limited by serial thinking; use multiple “capabilities” in parallel.

The “swarm” concept: dispatch multiple AI-like tasks

1) Your hands - Use agent mode/dispatch for tasks that would normally be sequential. - Example workflow: - “find leads” - verify emails - rank by revenue - then let it run - Noted limitation: in the described setup, ChatGPT can’t directly execute actions end-to-end (e.g., writing to a CRM or building an automated process completely on its own).

2) Your eyes - Use image understanding on photos of: - whiteboards - parts - dashboards - Prompt example: - “Analyze this picture. What’s wrong? Tell me one move to fix it.” - Emphasis: it can notice things humans miss.

3) Your research assistant - Upload large dense documents (e.g., 300-page presentations). - Use multiple AI analysts in parallel to synthesize results in a short time (compared to weeks).

Claimed outcome

  • Transforms “one-off help” into an army of interns for faster research, triage, and execution.

3) The Devil’s Advocate (adversarial review instead of validation)

Claim

  • ChatGPT tends to keep the user happy (validation), so to improve decisions you must force conflict/critique.

Motivation/framework

  • “Devil’s advocate” mindset: high-stakes decisions need challenge, not agreement.
  • References:
    • IKEA effect: creators overvalue what they made; even smart people spot others’ flaws but miss their own.

“Three tricks” to override people-pleasing behavior

1) Magic triggers - Use explicit adversarial prompts, such as: - “Audit this using first principles” - “Act as my fierce competitor”

2) Shadow boxing - Simulate debate—e.g., have a ruthless skeptical VC push back against a pitch.

3) Blind spot check - Provide your strategy and ask: - “Based on what you know about me, what are the gaps in my reasoning?”

Claimed outcome

  • Use ChatGPT as a stress-testing tool to find weaknesses before real-world meetings.

4) The Neural Link (accelerated understanding + knowledge transfer)

Argument

  • Don’t just copy fluent phrases; build real understanding (contrasted with a “chauffeur” memorizing speeches).

Three methods

1) Jargon translator - Ask for explanations using analogies (example: explain an API via a restaurant metaphor).

  - Goal: map complex concepts into intuitive “documentation → menu” style understanding.

2) Progressive professor - Explain the same concept at three levels: - like a 5-year-old - like a 15-year-old (logic + examples) - like a college student (underlying theory) - Claimed effect: concepts “click” faster via layered scaffolding.

3) Degree download - Use AI as a target audience/peer: - “Only actionable skills, no fluff” - then test me - ensure the user can teach it back.

Claimed outcome

  • Replace long study paths with guided learning plus feedback loops.

5) The Executive Coach (emotional processing + decision clarity via “mirrors”)

Purpose

  • Other strategies won’t work well without emotional self-management.

Core use-case

  • Use ChatGPT as a zero-emotion-baggage mirror, possibly via voice mode.

Three “mirror” exercises

1) Vomit mirror - Use voice mode and rant uninterrupted for ~5 minutes. - Then ask it to organize output into: - problems solvable today - problems to delegate - noise to ignore

2) Stoic mirror - After a loss (e.g., deal lost / fired), ask it to reframe using story principles and argue why it’s an opportunity.

3) Decision mirror - Before major decisions, “talk it out”: - “I am about to do X. Here’s my logic—let’s have a conversation.” - Emphasis: don’t blindly trust it; use it to clarify your own judgment.

Overall positioning / conclusion

  • The speaker frames these approaches as creating an “edge” (alpha) through better prompts, structure, and workflow design.
  • Warns that what’s rare today may become commodity later—so the real edge is maintaining a high-quality process and human-centered priorities.
  • Ends with a theme: AI frees humans from mechanical labor; the speaker adds the “CEO reminder” that “E” also stands for emotions, implying leadership is about managing human feelings as AI automates tasks.

Main speaker/source(s)

  • Primary source: The video presenter (former CEO/board member/investor with deep AI experience; name not provided in the subtitles).
  • Referenced examples/stories:
    • Grandpa/Alzheimer’s anecdote
    • Ancient Indian mythology (Ashtabuja Dura)
    • Catholic Church “Advocaci/devil’s advocate”
    • Marcus Aurelius
    • Max Planck chauffeur story
    • “IKEA effect” research (general reference; not attributed to a named researcher)

Original video