DEPTH · GEMINI · 01 OF 04

What Gemini actually is

Google's AI — strongest where Claude and ChatGPT are weakest

11 min read·1 of 4 in the Claude track

Gemini is Google's chat AI, and it's the most underrated of the three. People skip it because they already have Claude or ChatGPT — and miss that Gemini is actively better than both at specific things: live information, very long documents, and deep integration with the rest of Google.

If you've read the Claude or ChatGPT trust pieces, the fundamentals carry over. What's worth your time here is understanding where Gemini is genuinely different — and where those differences make it the right tool.

If you haven't read the Claude Trust piece, start there for the fundamentals (brilliant friend, no memory, no eyes, no live feed). This page focuses on what's specifically different about Gemini.

PART 01

What's different from Claude and ChatGPT

1. Gemini is grounded in live search by default

Claude and ChatGPT only see live information if you turn on web search and they decide to use it. Gemini hits Google by default. Ask it 'what happened in the news yesterday' and you'll get a current answer, with citations, automatically.

This is the single biggest functional difference. For anything that depends on current information — current events, current prices, current opinions, current status of anything — Gemini is the tool. The other two are guessing from training data unless explicitly told to search.

2. Gemini handles very long inputs

Gemini's context window — how much text it can take in at once — is the largest of the three by a wide margin. You can upload a 300-page PDF and ask questions about specific paragraphs. You can paste in entire books, multiple research papers, hours of meeting transcripts.

This matters more than it sounds. Most useful AI work hits the ceiling of how much context you can give the tool. Gemini's ceiling is higher.

3. Gemini is integrated with Google Workspace

If you use Gmail, Docs, Drive, Sheets, or Calendar, Gemini lives inside those — not as a separate chat, but as a sidebar you can pull up and ask about whatever you're looking at. 'Summarize this email thread.' 'What did I commit to in this doc?' 'Find me last quarter's report.'

This is convenience-level integration, not magic. It works because Google owns both ends. If you're a Google household, Gemini becomes the easiest AI to actually use day-to-day.

4. Gemini has a different personality

Claude tends to be articulate and slightly long. ChatGPT tends to be punchy and enthusiastic. Gemini tends to be more cautious — it hedges more, refuses more, and produces slightly more structured (sometimes corporate-feeling) responses.

The cautious tone is a feature when you want grounded answers and a bug when you want creative writing or confident sparring. Match the tool to the task.

PART 02

How to read what Gemini says

Same three habits: separate shape from content, treat specifics as guesses until you check, notice the agreement reflex. With Gemini, the calibration is slightly different.

Because Gemini cites sources by default, you can check claims by clicking the citations — much faster than verifying Claude or ChatGPT. Do this often when accuracy matters. Citations don't always say what Gemini says they say.

Gemini's yes-man reflex is more muted than ChatGPT's and roughly equivalent to Claude's. You'll still need to ask for pushback explicitly, but it's less of a fight.

PART 03

What Gemini can't do (today)

  • Match Claude's writing voice for long-form work. Gemini's prose tends to be more structured and less alive.
  • Match ChatGPT's image generation breadth. Gemini generates images but with fewer options.
  • Run as easily on non-Google ecosystems. If you don't use Google Workspace, you lose the biggest convenience advantage.
  • Handle 'tell me everything you remember about me' the way ChatGPT can. Gemini's persistence is more limited.
  • Take actions on your behalf in apps outside Google. The agent-mode trend is real, but Gemini's version is mostly within Workspace.

PART 04

Free vs Pro vs Ultra

Gemini has a free tier and paid tiers (the exact names shift; it's currently Pro and Ultra, but Google reshuffles these often). What matters as a beginner:

Free is enough for

  • Trying Gemini for the first time
  • Live-information questions (news, prices, current events)
  • Working with reasonably-long documents
  • Most casual writing, thinking, and learning tasks
  • Light image generation

Paid is worth it if

  • You want access to the strongest Gemini models for harder reasoning
  • You want deeper Workspace integration
  • You want Deep Research mode for long-form information gathering (a Gemini standout, covered in Climb)
  • You hit usage limits often on free
Honest advice: if you already pay for Claude or ChatGPT, you probably don't need to also pay for Gemini. The free tier covers most uses where Gemini is genuinely better. Pay for Gemini specifically when you need Deep Research or heavy Workspace integration.

PART 05

When Gemini is the right tool

Gemini is strong at: anything requiring current information (live search baked in), working with very long documents (longest context window of the three), Workspace tasks (Gmail, Docs, Drive integration), and grounded factual answers with citations.

Gemini is weak (relative to the others) at: alive prose with a strong voice (Claude wins), creative breadth and multimedia (ChatGPT wins), being a confident sparring partner (it hedges more).

A useful rule of thumb across all three: Claude for thinking and writing, ChatGPT for breadth and speed, Gemini for live information and long documents. You don't need to pick one. You can use the right tool for each task.

Try it now

Three prompts that test Gemini's distinct strengths. Free tier handles all three.

Practice 1 — Use live search on purpose

What's happened in the last week with [SOMETHING YOU FOLLOW — a sport, a country, a company, a topic]?

Give me the three most significant developments with sources. For each, tell me whether it's a big deal or a small story. Cite the source URLs so I can check.

Practice 2 — Test the long-context strength

(Upload a long document — a research paper, a long article, a PDF book, a meeting transcript.)

I want you to do three things with this:

1. Summarize the main argument in three sentences.
2. Find the three most important specific claims (with location in the document).
3. Tell me what's missing from this — what would I want to know that it doesn't cover?

Practice 3 — Use Gemini inside Google Workspace

(This is an exercise, not a prompt — only works if you use Workspace.)

Open Gmail or Google Docs. Find the Gemini icon in the sidebar. Ask it to do something specific to the document or email you're looking at:

- "Summarize this thread."
- "Draft a reply that [does X]."
- "What did I commit to in this document?"

Notice the speed advantage over copy-pasting into a separate chat tool. That's the integration story.