Get unstuck without losing the learning

You do not need to be the fastest installer, strongest coder or most confident speaker. The goal is to make useful work and explain why it can be trusted. Ask for help early; do not pay for an unapproved tool or fabricate a result to catch up.

One small loop

Predict: name one expected result and its requirement before asking AI to explain it.
Try: run one safe, scoped check using fictional data.
Compare: record what actually happened and what you still cannot conclude.
Explain: tell a partner in your own words; show the source rather than reading the AI chat.
Improve: choose one correction or next question. Apply the idea to your own product during the existing homework budget.

Use the short recall prompt at the top of each published lab during arrival/discussion—not as another exam. At the next session, revisit one earlier claim before reopening its evidence. Needing to look it up is useful information about where to practice, not a failing score.

Choose support, not a different standard

New to this: have a coach demonstrate one step, repeat it yourself, then do the next step with less prompting. Use the handout’s supported route and supplied templates.
Comfortable with the basics: complete the core lab independently, then ask someone else to reproduce one observation.
Ready for more: attempt the existing Stretch only after core evidence is recorded. Stop when the timebox ends; stretch work is not required for course completion.
Tooling blocked: record what could not run and use the documented alternative if the lab supplies one. Paired observation is not solo execution. Mark unverified work honestly and arrange the missing hands-on practice with the instructor.

These routes change scaffolding, not the approved project criteria. No route automatically awards completion or certification.

Ask for help with a useful packet

Send only the minimum safe context through the instructor-approved channel:

Session / activity:
My goal and requirement:
Expected result:
Observed result or exact sanitized error:
OS / shell / browser and relevant version:
One thing I already tried:
Relevant relative file path or safe artifact:
The specific help I need:

Keep passwords, tokens, recovery codes, personal absolute paths and private data out. Do not paste entire logs or the whole repository into a model. For the model-evaluation lab, input-only cases and reviewer labels must remain separate. For Session 00, the explanation task uses only fixtures/setup-example.mjs.

Catch up after missing a session

Open the lesson and lab for the missed session. Read the mission and deliverables first. Ask your instructor which demonstration or paired activity to reproduce, what support is available and when feedback can happen. Reproduce the core evidence with approved tools; viewing slides alone is not evidence of hands-on completion. Resume your current project milestone after agreeing on any scope or schedule adjustment.

Make the course work for your circumstances

Use the platform guide for Mac, Linux and Windows. Written handouts, downloadable materials, paired roles and the Markdown reader in the site navigation offer different ways to access the material. Tell your instructor privately if keyboard use, vision, hearing, language, scheduling or hardware makes an activity difficult. Agree on accessible evidence formats without requiring you to publish a diagnosis or other personal information.

The course has no public speed rankings or bug-count leaderboard. Feedback should identify an observation and a next action, not label the learner. Decisions and private records stay with the instructor; public practice work uses the safe submission process.