Video summary

How To Master Google Gemini in 2026 (Free Course)

Main summary

Key takeaways

Technology

Summary of the subtitles (tech concepts, features, setup, and workflow)

1) What Gemini is and why results differ

  • Gemini is presented as a highly capable AI chatbot.
  • The main claim is that users get either strong or underwhelming results largely based on how well they understand and configure Gemini.

2) Where to start + account access

  • Access Gemini via gemini.google.com.
  • Being signed into a Google account (e.g., Gmail/Calendar) on that browser automatically signs you into Gemini.

3) Pricing plans (and how the speaker uses them)

  • Gemini plans: Free, Plus, Pro, Ultra.
  • Ultra is described as for “super users,” and the speaker recommends ignoring it.
  • Practical approach:
    • Start on Free
    • Move up later as needs grow
  • During the walkthrough, the speaker switches between:
    • Free (white background)
    • Pro (black background) because some features differ by plan.

4) Critical setup: “Instructions for Gemini” (custom instructions + memory)

In Settings (gear icon), the speaker focuses on Instructions for Gemini, which includes:

  • Custom instructions: how Gemini should respond (e.g., detailed answers with examples/analogies).
  • Personal information Gemini should remember: career, preferences, device types, dietary restrictions, etc.

Key idea:

  • The more detail provided, the better Gemini can tailor responses.

Example memory:

  • “I have two dogs and no cats” is stored and appears in an “info” area.

5) Pro-only enhancement: “Personal intelligence” (better memory + connections)

On Pro, there’s Personal intelligence, described as bundling:

  • Past Gemini chats
  • Instructions
  • Connected apps

With Personal intelligence enabled (speaker recommends toggling it on):

  • Gemini can use info from past chats + connected apps + instructions more effectively than on the Free plan.

Connected apps include services like:

  • Gmail, Docs, Google Keep, Google Drive, Google Calendar, YouTube search, etc.

Note:

  • On Free, connected apps exist, but are said to work less powerfully than Personal intelligence.

6) Chat management and privacy approach: Temporary chat / model training notes

Gemini navigation includes:

  • New chat
  • Temporary chat (similar to incognito)

Temporary chat behavior:

  • Doesn’t show up in recent chats
  • Won’t be used for Personal intelligence (on paid plans)
  • Not used to train the models (as stated)

Speaker’s limitation/annoyance:

  • Gemini generally doesn’t let users opt out of model training the way some other bots might.
  • If you opt out in a way that prevents training, chats may not be saved (speaker calls this “nerfs” the experience).

Recommendation:

  • Use Temporary chat for more private content.

7) Core chat usage + model selection (fast vs thinking vs pro)

Core chat:

  • Works like a standard AI chatbot for tasks like writing emails and researching hikes.

Input options:

  • Microphone input (speaker enables voice input to avoid typing)

Model dropdown:

  • fast, thinking, pro

Guidance:

  • Pro reserved for advanced math/coding (speaker uses less often)
  • Free plan:
    • Unlimited fast
    • Limited thinking
  • Paid plans:
    • More thinking and access to pro

Speaker’s preference:

  • On paid accounts, use thinking for better quality.
  • Use fast for speed, but on Free you can hit limits for thinking.

8) Tools (Free vs Paid), and how to use them effectively

A) Deep Research

  • Generates dense, long reports (with charts/graphs).
  • Usage limits:
    • Free: only a few times per month
    • Paid: more frequent
  • Speaker framing:
    • Often overkill unless there’s academic or business need.

B) Create images (and recommended workflow)

Image generation can be triggered by:

  • Turning on tools → create images, or
  • Saying “create an image” in the chat (speaker says this is enough)

C) Canvas (highlighted as underutilized but powerful)

  • Canvas provides an interactive workspace for writing/coding with versioning.

Example workflow shown:

  1. Generate content in chat
  2. Reopen/redo in Canvas
  3. Edit specific sentences by highlighting and asking for changes
  4. Use previous/next versions
  5. Apply formatting (bold/italics, etc.)
  6. Share/export to Google Docs and generate share links

Canvas “create” conversions:

  • Web page, infographic, quiz, flashcards, audio overviews

Important drawback:

  • Copying infographic content inside Canvas may copy code, making it hard to download the actual image.
  • Recommendation:
    • If you want a downloadable infographic, generate it in a normal chat (not Canvas).

D) Guided Learning

  • A tutor-like learning mode:
    • Gemini asks what the learner wants to study
    • Provides back-and-forth explanations and comprehension checks

E) Visual Layout (Free “labs” feature)

  • Produces more visual, carousel-like results instead of plain text.
  • Includes:
    • Image snippets, summaries, follow-up questions
    • Sometimes references to YouTube videos
  • Downside:
    • Takes longer to load
  • Speaker usage:
    • Rarely used unless a visual answer is needed.

F) Create videos (Pro-only capability)

  • On Pro, Gemini can generate videos via a “create video” request.
  • Optimization tip:
    • Create the image first (start frame), then generate video from it
  • Hard limit:
    • Only a small number of videos per day (speaker example: “two more videos today”).

G) Add files / multimodal inputs

  • Upload:
    • Images, screenshots, PDFs, Word docs
  • Also supported:
    • Google Drive items, Photos
    • Notebook LM notebooks (as knowledge sources)

9) Advanced: Notebook LM integration (knowledge organization)

Notebook LM is presented as a way to organize sources:

  • Create notebooks per topic
  • Store and categorize materials

Workflow:

  • Use Notebook LM to store research
  • In Gemini, select a Notebook LM notebook so Gemini can pull the information
  • Example:
    • Create a “training guide” using a Notebook LM topic source
    • Optionally refine it using Canvas

10) “Gems” (repeatable custom assistants + reusable instructions/tools)

  • Gems are reusable AI “apps”/configurations.

Categories:

  1. My gems (user-created)
  2. Gems made by labs (experimental; available to the speaker’s free account)

Main use case:

  • Avoid rewriting the same instructions repeatedly (e.g., an “email helper”)

Gem configuration includes:

  • Custom instructions
  • Default tools to always enable for the gem (e.g., Canvas for email formatting)
  • Knowledge upload:
    • Upload SOP PDFs, help docs, or files so Gemini answers based on that material

Gem testing and iteration:

  • Submit example input in the gem preview
  • Adjust instructions (e.g., “never use numbered lists”)
  • Save the final configuration

Gem management:

  • Edit via three dots
  • Pin gems so they don’t disappear (only a limited number show in the left list at once)

11) Scheduled actions (Pro-only)

  • Recurring tasks example:
    • Daily at 6:00 a.m. Mountain Time, scan the latest AI news from the last 24 hours
  • When enabled:
    • Appears under scheduled actions
    • Can be edited or toggled off
    • Triggers notifications in the mobile app.

12) App on mobile (additional capabilities)

Mobile Gemini is similar to web, with notable extras:

  • Camera
    • Take a picture and ask “What is this?”
    • Can translate signs while traveling
  • Templates for videos (if you have Pro/video access)
  • Real-time chat experience
    • Speaker suggests it’s great for driving/walking/cooking
  • Camera can also be activated during the real-time session

After the session:

  • The conversation content is retained/shown.

Main speakers / sources

  • Single main speaker (tutorial host/author): The narrator throughout the subtitles (no other named sources are present as speakers).

Original video