Google Gemini Bisa Gini?! (18 Fitur Lengkap)

Main Ideas / Concepts Conveyed

• The speaker argues that Gemini is a highly capable AI (described as one of the “most powerful AIs” in Indonesia).
• The speaker explains why someone might consider upgrading to Gemini Pro, instead of relying only on other AI tools (e.g., GPT Chat / Cloud).
• Gemini is positioned as especially strong because it is integrated with Google’s ecosystem—for example:
  • Gmail, Google Sheets, YouTube, Google Drive, and more.
• The video’s core goal is to enumerate 18 Gemini features and demonstrate practical examples for each, helping viewers judge whether Pro is worth it.

Methodology / List of Features (18 Total)

The speaker claims to be showing “18 features available in Gemini,” presented roughly in this order:

1. New Gemini Interface + Model Choice

• Gemini offers a choice between:
  • 3.5 Flash and 3.1 Pro
• The interface is described as having a “new look.”
• “Fastest” Flash is also mentioned.

2. Multi-Model / Multimodal Understanding

• Gemini can understand more than text, including:
  • Images, videos, audio, and other uploaded file types.
• Example concept: upload materials, then ask questions one by one based on what you provide.

3. Pro vs Flash/Normal Tradeoff

• Pro (3.1) is described as smarter but slower.
• It uses more compute and takes longer to think.
• Analogy: the smarter the AI, the more expensive it is and the longer it takes.
• Scenario used: planning vacation leave so the user stays within year-end leave rules (including date/count recommendations).

4. Example: Policy/Holiday Questions Answered Precisely

• Example in an Indonesian context:
  • national holidays + remaining leave
  • answers appear in structured form (e.g., selecting options like A/B).
• Emphasizes Gemini’s ability to produce structured answers based on the question.

5. Multimodal Input (Text + Microphone/Audio)

• Gemini supports voice input.
• Example: microphone tests like “Test, test, test…”
• Key point: you don’t have to rely only on typing.

6. Analyze YouTube Content

• Gemini can interpret a YouTube video when the content is provided in a way Gemini can access.
• Example outcome:
  • ask what the video is about
  • generate learning outputs from it (based on the video content/metadata).

7. Gems (Gemini Custom Mini-Apps / Sidebar Tools)

• Gems are described as mini applications accessible via a sidebar.
• Examples include:
  • A tool to create a podcast (“personal podcaster”) from a topic
  • A mini tool that generates multiple angles of an object from an image (front/side/back/top-down/isometric/macro), similar to earlier “canvas”-type concepts

8. Deep Research

• A “research mode” designed for more detailed, step-by-step research.
• The video claims it can take a long time (example: ~30–40 minutes, potentially up to tens of minutes).
• Example workflow:
  • Gemini provides an initial plan/steps
  • the user can review and edit
  • topic example: “how to clean an egg,” structured into methods and steps

9. LM Notebook

• Presented as a tool that creates learning resources based on what the user uploads.
• Example contrast:
  • Gemini provides answers/resources based on uploads
  • LM Notebook is framed more as a learning/source tool that organizes output (e.g., turning a video into slides).

10. Scheduled Actions

• Users can set automated recurring updates/reminders.
• Example:
  • “Every Monday morning” send AI news in Asia + exchange rates (USD/IDR/SGD)
• Delivery can be via email and/or inside the Gemini/Google app on the phone.

11. Creative Studio (Image Generation/Editing)

• Gemini can generate/edit images using image models (the narration mentions something like “Nano Banana 2”).
• Example includes generating realistic-looking imagery (e.g., banana/eating/turtle-style scenes).

12. Video Generation (Text-to-Video / Image-to-Video)

• Mentions a video model such as Veo (and refers to “Omni/Google Omni” in narration).
• Demonstrated by turning a prompt (and/or image) into a moving video, described as realistic (National Geographic-like).

13. Canvas (New Chat/Workspace for Building Mini Apps + Slides)

• Canvas is presented as a workspace to:
  • create mini applications
  • create slides
  • build outputs that can be exported into Google tools
• Example:
  • insert a turtle image and ask for slides about a concept
  • export/route to Google Slides (tied to the Google ecosystem)

14. Guided Learning

• A learning-oriented mode aimed at teaching concepts.
• Example:
  • explain compound interest “like I’m 5 years old”
  • uses an interactive metaphor/simulation style approach
• Positioned as useful for teaching kids or younger siblings with simpler explanations.

15. Gemini Live (Phone-Only Real-Time Interaction)

• Described as a more conversational, “live” experience.
• Includes voice/video-call style interaction:
  • user can flip the camera
  • Gemini responds contextually
• Example:
  • user is confused about a car dashboard/buttons and asks what they are for.

16. Workspace Integration

• Gemini is described as integrated into Google work tools:
  • Google Drive (working with files, supporting workflows with Sheets)
  • Google Sheets
  • Gmail
• Example:
  • Gemini helps identify which emails are important to read vs not.

17. Personal Intelligence

• Framed as similar to “memory” in other AI systems.
• Gemini can use knowledge about the user and what apps you use to improve responses.
• The speaker’s argument:
  • Google’s advantage is access to many services (Maps, Search, Drive, YouTube, etc.), enabling deeper personalization.

18. Agent Mode (Assistant That Can Manage Across Apps)

• Presented as an advanced capability.
• The speaker notes they personally didn’t have access; it’s said to be available in the US.
• Concept:
  • Gemini can be granted access to multiple Google apps (Photos, Drive, Gmail, etc.)
  • enabling Gemini to behave more like a personal AI agent
• Speaker framing:
  • it could eventually manage many parts of the user’s life/workflow from one place.

Main Lessons / Takeaways

• Gemini’s key differentiator is integration with Google’s ecosystem, expanding what it can do (e.g., learning from YouTube, working with Drive/Sheets/Gmail).
• Model choice matters:
  • Flash: faster
  • Pro: smarter for complex tasks, but slower
• Pro is portrayed as most worth it for:
  • deeper reasoning
  • lengthy research
  • more complex planning
• The speaker encourages viewers to try before buying, including:
  • free access and a 1-month Pro trial

Speakers / Sources Featured

Speaker: William

• Background mentioned:
  • previously worked at Bidance and Alibaba
• Current project mentioned:
  • Belajar GBT (an AI education platform)

Tools/Services Referenced (Not the Video’s Original Speakers)

• Google ecosystem/services:
  • Gmail, Google Sheets, Google Drive, Google Slides, YouTube, Google Maps, Google Search, Google Photos
• Gemini features/models/tools:
  • Gemini (3.5 Flash, 3.1 Pro)
  • Gems, Deep Research, LM Notebook, Scheduled Actions
  • Creative Studio (image)
  • Veo/Google Omni (video)
  • Canvas, Guided Learning, Gemini Live
  • Workspace Integration, Personal Intelligence, Agent Mode
• Comparator AI tools mentioned (in general terms):
  • “GPT chat,” “Cloud chat” and “ChatGPT”