AQI AIAQI AI Intelligence
⌕
Creator course referenceNot tested by AQI AI

Start learning AI: video guide and five exercises

Start with 卡普迪姆’s Chinese season-one compilation, then apply our original exercises to one small task.

Last checked: Event date:
A quick glossary

Models generate and reason. Agents use tools to work on tasks. A harness manages an agent’s environment. MCP connects tools and data. Skills are reusable task guidance.

Before you start

Prerequisites

Understand spoken Chinese; a phone or computer and notes suffice. The author describes 13 first-season lessons, displayed as 54:57. Eight-language guidance does not translate the video.

Estimated time

Video about55minutes, watch in sections; plan2–3practice sessions, not a measured completion time.

Implementation steps

  1. Choose one small real task

    Stage practice: After the author’s introduction, write the goal, input and acceptance criteria in your own words. Use public or fictional material; buying hardware is unnecessary.

    Sources for this stage: 卡普迪姆 · AI learning, season one compilation

  2. Record one prompt improvement

    Stage practice: Use the same fictional input twice: an open request, then one with goal, constraints and format. Save both and compare against the original criteria, not response length.

    Sources for this stage: 卡普迪姆 · AI learning, season one compilation

  3. Turn model selection into requirements

    Stage practice: List task, privacy, budget and equipment. Compare exact models here; use the hardware page for local options. Separate estimates and unknowns from proof of suitability.

    Sources for this stage: 卡普迪姆 · AI learning, season one compilation

  4. Map the workflow before choosing an agent

    Stage practice: Outline input, model judgment, program processing and human approval. Mark file access and outbound actions. This exercise is a paper plan; no plugin, account connection or API key required.

    Sources for this stage: 卡普迪姆 · AI learning, season one compilation

  5. Write a one-page review and choose a next route

    Stage practice: Submit one page with goal, two-output comparison, model-choice rationale, cost unknowns and one failure. Then choose a linked route: local chat, coding agents or web/Skill/MCP. One direction at a time is enough.

    Sources for this stage: 卡普迪姆 · AI learning, season one compilation

Validation checks

Practice

Submit one page with goal, two-output comparison, model-choice rationale, cost unknowns and one failure. Then choose a linked route: local chat, coding agents or web/Skill/MCP. One direction at a time is enough.

Cost

Not verified

The page opened at review. Playback, quality and regional rules belong to Bilibili; perpetual free access is not guaranteed. Paper practice and existing lawful tools suffice; paid models, APIs or equipment are optional extras.

Limits & boundaries

Reviewed the author’s description, chapters, duration and rights notice, not every minute or plugin. The page forbids unauthorized reposting. We link the author and write our own exercises; no copied mind map, video, subtitles or cover.

Original sources

  1. 卡普迪姆 · AI learning, season one compilation ↗

    www.bilibili.com · · Creator course reference