SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsstart-small-find-partner

start-small-find-partner

Design & creation

View repository

Updated 10-16 06:21

About this skill

Source Metadata

Original cangjie-skill frontmatter from the distillation run:

id: f22
title: "从小处着手+寻找伙伴 — AI入门的行动原则"
book: "AI for Everyone"
author: "Andrew Ng（吴恩达）"
source: p25
tags: [getting-started, action-principle, learning, collaboration]
related_skills:
  - slug: pilot-momentum-flywheel, relation: composes-with
  - slug: iterate-not-perfect, relation: composes-with

R — Reading (Quote from the original text)

"The project doesn't have to be too big... team up with engineer friends... invite colleagues to learn together... the second project will be better than the first" — p25

I — Interpretation (Restate in your own words)

For AI beginners, the most common trap is "waiting for the perfect plan" or "trying to get it right in one step." The action principles Andrew Ng gives are:

  1. Start small: The first AI project doesn't need to be big or very valuable; it only needs to besomething you can complete. A small project can help you build confidence, understand the process, and discover your own knowledge gaps.
  2. Find a partner: Don't fight alone. Find a technically savvy colleague, an engineer friend, or a like-minded learner to work with. Collaboration allows you to complement skills, motivate each other, and reduce the probability of giving up.
  3. Get moving first: Don't wait until you're "ready" to start. The second project will be better than the first, and the third will be better than the second—the key is tostart doing。

Why is it effective?

  • Learning AI is "learning by doing"—attending lectures and reading books cannot replace doing a project yourself
  • The cost of failure for a small project is low, and success can build confidence
  • A partner provides technical and emotional support, greatly lowering the barrier to entry

A1 — Past Application (Past application cases)

  • Individual learner: A product manager wanted to learn AI and did not wait until he had "finished learning all the math foundations." Instead, he directly found an engineer friend and spent two weeks building a simple customer feedback classification mini-project. The project was not big, but it helped him understand the complete AI project process and laid the foundation for larger projects later.

A2 — Future Trigger (Future trigger conditions)

Trigger scenarios:

  • An individual wants to get started with AI but doesn't know where to begin
  • A team has just encountered AI and is still in the "learning stage" without taking action
  • Someone worries that "I'm not capable enough to do AI"
  • An individual learning AI feels lonely or is prone to giving up when encountering difficulties

Language signals:

  • "I want to learn AI but I don't know where to start"
  • "AI is too hard, I'm afraid I can't do it"
  • "I'll start doing projects after I finish learning the basics"
  • "I want to find someone to learn AI with"
  • "What's a good first project to do"

Key distinctions:

  • Is the useran individual getting startedoran organizational transformation? This skill is aimed at individual getting-started scenarios.
  • Is the userseeking knowledgeorseeking action? This skill emphasizes "learning by doing."

E — Execution (Execution steps)

  1. Find a partner：

    • Prioritize finding a colleague or friend who understands AI technology (engineer, data scientist)
    • If there isn't one, find peers who also want to get started and learn together
    • You can form an AI study group within the company

  2. Choose a small project：

    • Choose a project that can be completed within 1-2 weeks
    • Choose a scenario where you have data or can obtain data
    • Don't chase "valuable"; chase "completable"

  3. Execute quickly：

    • Don't spend a lot of time learning theory; learn while doing
    • Discuss problems with partners promptly when you encounter them
    • Done is better than perfect

  4. Review and do a second project：

    • Summarize the lessons learned from the first project
    • The second project can be a bit larger and more valuable
    • Gradually increase the difficulty and impact

Completion criteria: Complete the first AI project (regardless of size), and find at least one AI learning partner.

B — Boundary (Boundaries of use)

Not applicable scenarios:

  • Organization-level AI transformation (need to use ai-transformation-playbook)
  • Teams with existing AI experience doing formal projects
  • Enterprise-level scenarios requiring strategic decision-making

Failure modes:

  • Choosing a project far beyond current abilities, leading to giving up midway
  • Trying to handle everything alone, with no one to help when encountering technical difficulties
  • Always in the "preparation" stage, never taking action

Author's blind spots:

  • Andrew Ng assumes you can find an "engineer friend"—but if you're in a traditional industry with no technical background at all, you may not be able to find a partner. In this case, consider: joining AI communities, finding external consultants, or starting with projects from online courses.
  • "Starting small" does not mean "only doing small things"—after getting started, you need to quickly transition to impactful projects.

Related skills

  • pilot-momentum-flywheel — composes-with: The individual-level "start small" is consistent with the organization-level "seek success first, then value" philosophy—both emphasize building confidence and credibility through small successes and gradually leveling up.
  • iterate-not-perfect — composes-with: "Start small" and "don't expect perfection on the first try" complement each other—the first project doesn't need to be perfect; what matters is starting and learning through iteration.