← AIAP Tool GuidesAll programsHomeSearch
AIAP Tool Guides·Claude·16:37

Claude Overview: the Long Context Window, Plan-Gated Models, Changing Its Tone with Styles and Preferences, and Projects

Karan Rana Brand/marketing manager, YouTuber, podcaster and course creator ("a non-tech guy"); presents the AI for Marketing, NotebookLM and Manus tool guides
▶LMS lesson
← PreviousPerplexity Discover, Personalisation, Multimodal Uploads and Voice: Context About You, Ima
Lesson 1 of 4 in Claude
Next →Claude Multimodal and Long Context: Analysing Images, Summarising a 64-Page Report, and Mi

The short version

  1. Claude's headline strength at recording was its ~200k-token window ("about 1.5 lakh words") and handling long PDFs without making things up; tokens and confident hallucination explained (0:00-0:04).
  2. Anthropic (2021, ex-OpenAI founders); "constitutional AI" - helpful, honest, harmless; model selector shows Sonnet 4 for everyday work and Opus 4 for complex tasks, which burns usage limits faster (0:04-0:06).
  3. Override the calm default tone with Styles or Settings preferences; weaknesses: few integrations, no agentic booking; dashboard with Research, starred chats and Projects (0:06-0:16).

At a glance, three clicks deep

Skim here first: the closed row is the glance, open is the study card with the key points and timestamps, and the ↓ link drops to that concept's full write-up below.

01Tokens, the long context window, and confident wrong answersText is counted in tokens (roughly a word each);›

Text is counted in tokens (roughly a word each); a large window lets the model read a whole long document at once; the dangerous error is a fluent, confident wrong fact.

Tokens and window (0:01-0:04)

Models and plans (0:04-0:06)

↓ Full write-up of this concept

02Change Claude's tone with Styles or personal preferencesPick a Style in the chat input (or create one from a writing sample), or write standing preferences in Sett…›

Pick a Style in the chat input (or create one from a writing sample), or write standing preferences in Settings that override the default tone.

Styles and preferences (0:06-0:09)

Weaknesses (0:09-0:11)

Dashboard (0:11-0:16)

↓ Full write-up of this concept

The concepts in full

01

Tokens, the long context window, and confident wrong answers

Garbage output you can spot; a confident lie "you will never be able to catch".

He presents the 200k window as "the single most important strength of Claude" for non-technical users with long PDFs, pitch decks or big spreadsheets. Model choice matters for limits: on Pro he saw Opus 4 ("for complex tasks") and Sonnet 4 ("smart and efficient for everyday use"), plus older Haiku 3.5, Opus 3 and Sonnet 3.7.

Why it matters

Picking the right tool and model for big inputs.

This idea elsewhereparallelAI Studio: tokens and context windows
02

Change Claude's tone with Styles or personal preferences

how-to

Funny mode: "most plants need light like your dating prospects."

Claude's default is calm and empathetic, "like a thoughtful coach". Styles menu > create and edit styles lets you upload a writing example. Settings preferences ("use a funny tone instead of calm") applied to every answer until he set "do not use a funny language". Also shown: the Research toggle, starring chats, and a Project holding related chats.

Do it in this order
Why it matters

Outputs that sound like you, not like a chatbot.

Tools referenced

ToolCoverageMomentContext
Claudedemonstrated
Claude Sonnet 4explained
Claude Opus 4explained
Geminimentioned

Session materials

Archived locally on V: — click to open. Companion pages link to the LMS.

lms
Course Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Audio Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform
lms
Overview
companion page on the platform

Action items

    Resources mentioned

    Resources
    • docLesson page on the platform

    Extraction notes

    This page was built from an auto-generated transcript, which garbles product and people's names. Those were corrected silently in everything above and logged here for transparency. The warnings flag claims that were true on the recording day but change fast.

    Transcript corrections applied

    The transcript saysThe trainer actually means
    cloudClaude
    circum and stanchioncircumstantial
    google keyGoogle Keep

    True on recording day — verify before relying

    ← PreviousPerplexity Discover, Personalisation, Multimodal Uploads and Voice: Context About You, ImaNext →Claude Multimodal and Long Context: Analysing Images, Summarising a 64-Page Report, and Mi