PerplexityA starting map of the literature with real links, not a black-box summary.
Track 2 · Beginner
2.1
Frontier chatbots
The frontier chatbots — ChatGPT, Claude, Gemini, Perplexity
1Overview
The frontier "consumer" cloud chatbots — hosted end-user apps you just open and prompt, no setup. ChatGPT (all-rounder), Claude (careful writing & analysis), Gemini (Google-grounded, multimodal), Perplexity (research with live sources). → Unlike the builder tools, these are for direct everyday use. Pick by what you want: cited web research → Perplexity; careful writing & reasoning → Claude; it already works inside your Google Drive & Gmail → Gemini; a broad all-rounder that runs Python on your data → ChatGPT.
1.1Before you start
1.2After this chapter you can
→Tell the four frontier chatbots apart by what each is actually best at
→Pick the right one for a given job instead of defaulting to the one you know
→Know what all four share — and what none of them do for you
→Read the comparison matrix rather than re-testing every tool yourself
1.3Which chatbot excels at cited research?
Use Perplexity when you need answers backed by live web sources; it specializes in pulling and citing up‑to‑date information for research tasks.
1.4When should I choose Claude?
Pick Claude for careful writing, nuanced analysis, and reasoning‑heavy prompts where a more deliberate, safety‑focused response is valuable.
1.5How does Gemini work with Google apps?
Gemini runs directly inside your Google Drive and Gmail, offering grounded answers and multimodal support while staying integrated with Google’s ecosystem.
2Matrix 7 rows · 4 tools
chatgpt
claude
gemini
perplexity
Usable free tier
yes
yes
yes
yes
Web search + live citations
partial
partial
yes
yes
Reads your files
yes
yes
yes
partial
Handles very long documents
partial
yes
yes
partial
Generates images
yes
no
yes
no
Strong careful reasoning
yes
yes
partial
partial
Best for
all-rounder
writing & analysis
Google + multimodal
research & sources
3Lessons 5
3.1 Generate a marketing email with ChatGPT
ChatGPT is the OpenAI conversational assistant described as a general‑purpose “all‑rounder”.
Create a ready‑to‑send promotional email for a new service.
- Open the ChatGPT web page (use the “Try ChatGPT” link on openai.com).
- Log in or start a free session as prompted.
- In the chat box type: “Write a short marketing email announcing a new AI‑assisted analytics service for small businesses.”
- Press Enter and wait for the response.
- Copy the generated text into a document.
- You'll see A complete, professionally worded email draft appears in the chat window.
- Takeaway ChatGPT can quickly produce polished copy for marketing tasks with minimal prompting.
3.2 Analyze a legal paragraph using Claude
Claude is Anthropic’s chatbot noted for “deep analysis” and high‑quality structured answers.
Obtain a concise summary and key issues from a sample legal text.
- Navigate to Claude’s web interface (search for “Claude AI” and open the official site).
- Start a new conversation.
- Paste the following paragraph: “The lessee shall maintain the premises in good condition, subject to reasonable wear and tear, and shall be liable for any damage caused by negligence.”
- Ask Claude: “Summarize this clause and list potential risks for the lessee.”
- Record Claude’s bullet‑point summary.
- You'll see Claude returns a short summary plus a list of identified risk points.
- Takeaway Claude excels at extracting structured insights from dense text, useful for document review.
3.3 Create a social post with image using Gemini
Gemini is Google DeepMind’s multimodal model that handles both text and images.
Produce a captioned graphic for a product launch that combines visual and textual content.
- Open Gemini’s web app (search for “Gemini AI” and select the official Google page).
- Select the multimodal option to upload an image.
- Upload a product photo of your choice.
- Prompt: “Write three catchy Instagram captions for this product, each with a different tone.”
- Download the generated text and pair it with the original image.
- You'll see Three distinct caption suggestions appear below the uploaded picture.
- Takeaway Gemini’s multimodal capability lets you generate coordinated visual‑text content in one step.
3.4 Gather cited market data with Perplexity
Perplexity is a chatbot focused on live web search and citation of sources.
Collect up‑to‑date statistics about the AI chatbot market and receive source links.
- Visit Perplexity’s website (search for “Perplexity AI” and open the official site).
- Start a new chat session.
- Enter: “What is the projected global revenue of AI chatbots in 2025? Provide sources.”
- Wait for the answer with inline citations.
- Copy the figures and their source URLs into a notes file.
- You'll see A numeric market forecast accompanied by clickable citation links to recent articles.
- Takeaway Perplexity streamlines research by returning up‑to‑date information together with verifiable sources.
3.5 Compare answer quality across the four chatbots
A side‑by‑side evaluation of ChatGPT, Claude, Gemini, and Perplexity on the same query.
Identify which tool best fits a given business need by analyzing their responses.
- In each previously opened chatbot (ChatGPT, Claude, Gemini, Perplexity) start a fresh conversation.
- Ask the identical question: “Explain the benefits of using AI assistants for small‑business customer support.”
- Record each response in separate sections of a document.
- Evaluate the answers against the criteria from page 4 (breadth vs depth, citation, multimodal content).
- Summarize which chatbot is optimal for general advice, deep analysis, visual examples, and sourced research.
- You'll see A comparative table showing each model’s strengths aligned with the business scenarios described in the source article.
- Takeaway Systematic testing reveals the practical trade‑offs among frontier chatbots, guiding tool selection for specific tasks.
4You’ll know it worked 120 checkable outcomes in this chapter
- ✓User sees inline links to original papers in the answer
- ✓A combined summary appears in the Drive sidebar after selecting documents
- ✓The posted JD contains no gendered pronouns and lists must-haves separately from nice-to-haves
- ✓Support agent receives a ready-to-send email that acknowledges the issue, outlines next steps and timeline
- ✓Patient can accurately state all medication doses and warnings after reading the rewritten instructions
- ✓The workflow calls all six skill files without errors
- ✓The browser displays the generated diagram or chart.
- ✓When you open a new conversation in the workspace, Claude references the information from Claude.md before responding
120 outcomes in all — one per recipe below.
5FAQ, Tips & How-to 236
one problem, one solution, one action
Research & data tools16
ChatGPTScript crashes with an error
A working script and an explanation you can learn from — no install needed.
A fast, checkable briefing instead of an unsourced guess.
Need a tidy CSV and quick insights
A computed answer plus a figure from your own data, no Python set up on your machine.
GeminiNeed a quick map of a topic while you work elsewhere
A broad first map of a topic with links, assembled while you do something else.
A quick, checkable verdict with links — not an unsourced yes/no.
Need to keep related questions and citations together
One organised, citation-rich thread instead of scattered one-off searches.
Need to track what’s changed with competitors or regulations
A lightweight, verifiable watch on the things that move, without a subscription service.
A cited, up-to-date set of inputs instead of an outdated number from memory.
Need a current regulatory brief
A cited briefing on the current status, effective dates, and key requirements, ready to pass to your auditor or legal team.
Need to narrow down dozens of resumes fast
A ranked shortlist with a one-paragraph rationale per candidate, turning a 2-hour CV screen into minutes.
Where can I find accurate salary benchmarks for a role?
A cited pay range you can present to a hiring manager or use in a pay-equity review.
Lots of free‑form survey responses
A thematic breakdown of qualitative feedback in minutes, instead of days of manual coding.
You reply with accurate information, not a guess — and you have the source if the customer escalates.
Need a fast, sourced market map of sector players
A first-pass market map with real sources in minutes, not a black-box summary you can't defend in IC.
Where can I find accurate salary benchmarks for a role?
A cited pay range to anchor the client conversation, not a number from memory you can't defend.
Knowledge & docs23
▸How-to 
Claude Scientist +1
Claude Scientist +1A structured critique of a long document in minutes, holding the whole paper in context.
▸How-to Claude Scientist
Claude ScientistNeed a full methods paragraph from bullet steps
A solid first draft of the hardest section to write, grounded in exactly what you did.
▸How-to Claude Scientist +2
Claude Scientist +2Want a private, critical look at your abstract
The obvious objections caught early, in private.
▸How-to Claude Small biz +2
Claude Small biz +2Need answers that follow my style guide every time
Consistent, on-context answers across many sessions from one shared workspace.
See agreements, disagreements and takeaways across docs
A synthesis across a stack of papers instead of reading each one cold, without ever leaving Drive.
▸How-to Claude Finance +1
Claude Finance +1Need an expense policy with clear limits and FAQs
A ready-to-share policy document in one pass, with no ambiguous language.
▸How-to Claude HR / People +1
Claude HR / People +1Need a ready‑made 30‑day plan for a new hire’s role
A structured 30-day plan you hand to a new starter on day one, with no blank-page effort.
▸How-to Claude Finance
Claude FinanceNeed a board‑pack narrative with variance explanations and outlook
A polished, board-ready narrative first draft that usually needs only minor edits.
Compare financial terms from two contracts without downloading PDFs
A structured comparison table of every financial clause across both documents, pulled right out of your inbox.
▸How-to Claude Finance
Claude FinanceNeed a tailored audit preparation checklist
A complete prep checklist tailored to your audit type, ready to assign in a spreadsheet or project tool.
▸How-to Claude HR / People
Claude HR / PeopleHiring bias in my job ad
A crisp, bias-reviewed JD that avoids coded language and focuses on what the role actually does.
Need a consistent interview process
A consistent, fair interview guide the whole hiring panel uses, not a set of improvised questions.
▸How-to Claude HR / People +1
Claude HR / People +1Need a UK remote‑work policy for 80 staff
A structured, ready-to-review policy draft in one pass — not a generic template filled with placeholders.
Need a custom offer letter fast
A complete, personalised offer letter in seconds, ready for legal review and sign-off.
▸How-to Claude HR / People
Claude HR / PeopleMy manager notes are messy
A review that is specific, fair, and actionable rather than vague or legally risky.
▸How-to Claude HR / People
Claude HR / PeopleNeed to give a redundancy notice
A structured script that keeps the conversation fair, on-track, and legally sound.
Recurring password‑reset tickets
A help-centre article written from real customer language, ready for your knowledge base.
A complete, accurate answer assembled from your own docs, not from the model's general knowledge.
▸How-to Claude Support
Claude SupportNew agents don’t know which tier to send tickets
A concise guide that reduces inconsistent escalations and helps new agents feel confident on their first week.
▸How-to Claude Investor
Claude InvestorLong earnings call transcript
A structured briefing on what matters in a long filing, ready to annotate or share with the team.
▸How-to Claude Investor
Claude InvestorNeed a quick one‑page memo from a pitch deck
A one-page first-pass memo with flagged gaps before you spend an hour on a call.
▸How-to Claude HR / People
Claude HR / PeopleNeed a client‑ready job description that starts with impact and uses inclusive language
A client-ready JD draft in one pass, without starting from a generic template.
▸How-to Claude Physician
Claude PhysicianBullet notes for a GP referral
A solid first draft of a routine letter in seconds, freeing time for the clinical judgement only you can provide — the AI drafts, you decide.
Dashboards & analytics3
Need a quick summary of my CSV data and a ready‑to‑use chart
A quick description plus a draft visual, using Gemini's multimodal strengths.
Need to highlight biggest budget gaps
A clear narrative for the numbers, ready for a board update or department head email.
Need optimistic vs downside cash flow view
Two scenario projections in seconds, with the key inflection point clearly called out.
Content & marketing7
Got bullet‑point notes and need an email
A solid first draft in seconds, then tightened to your voice.
Unstructured notes
A presentable skeleton in seconds that you refine, instead of staring at a blank deck.
▸How-to Claude Founder +1
Claude Founder +1Need a quick product overview that stays editable
A finished, editable document that evolves with the conversation.
Need a quick diagram from a text description
A draft visual in seconds you iterate on, instead of fighting a drawing tool.
Explain a finance model to non‑finance stakeholders
A jargon-free briefing note that lets non-finance stakeholders engage with the numbers.
Technical discharge notes confuse patients
Instructions a patient actually understands and follows, without diluting the medical content — you check the rewrite against the original before handing it out.
Patient handout needs a new language
A quick draft translation that helps a patient understand their care between visits — reviewed against the original before use, and via a certified translator for anything legally or clinically sensitive.
Internal tools & ops6
Team keeps rewriting the same prompt
A saved, shareable assistant your whole team opens instead of re-writing the same prompt.
Need a repeatable task without re‑typing prompts
A one-click specialist instead of re-typing the same setup prompt.
▸How-to Claude Support
Claude SupportNeed a warm, on‑brand first response to a support ticket
A polished, on-brand response ready to send or lightly edit — not a form letter.
Escalation thread full of messages
A manager or second-line agent can read the situation in 2 minutes instead of scrolling 40 messages.
▸How-to Claude Support
Claude SupportNeed to apologize for a wrong order and show care
An apology that sounds genuine and specific, not a corporate form letter — which is what actually rebuilds trust.
Support drafts that are too formal or blunt
A response that sounds like your brand, not a call-centre script — without lengthy guidelines re-training.
CRM & sales7
Need a cold‑outreach script for finance heads
A tested, persona-specific sequence ready to load into your outreach tool, not a generic template.
Need a fast, fact‑checked snapshot of a prospect
A 5-minute pre-call read that makes you sound prepared, with every fact checkable.
▸How-to Claude Sales
Claude SalesProspect uses their own wording and budget cues
A proposal that reads as if you listened carefully, not a copy-paste from a deck.
Need a competitor snapshot for calls
A one-page battlecard your reps can actually use on a call, built in minutes from public material.
Having to answer sales objections on the fly
A reference card reps can internalise before a call, not a script to read verbatim.
▸How-to Claude Sales
Claude SalesMessy deal debrief notes
A clean, shareable debrief that captures what actually drove the outcome — not a watered-down note in the CRM.
Need a personal outreach that cites their experience
A concise, specific message that reads as personally written and improves response rates.
▸How-to Claude Everyone
Claude EveryoneUnsure of your brand’s voice and audience
You get a structured brand essence markdown that guides design and copy decisions
▸How-to Claude Everyone
Claude EveryoneI have a markdown brand brief and need visual direction
You receive five distinct visual direction options that can be mixed and refined
▸How-to Claude Everyone
Claude EveryoneNeed a single source of truth for UI specs
You obtain a complete design system (colors, typography, layout, components) that can be handed to Claude Code
▸How-to Claude Everyone
Claude EveryoneRepeating the same work in every project
Sharing distilled wisdom across projects speeds up future work
▸How-to Claude Everyone
Claude EveryoneYour prompt file is bloated and leaks credentials
Regularly review the global Claude.md to maintain efficiency and security
▸How-to Claude Everyone
Claude EveryoneNewer Opus models generate more polished designs
▸How-to Claude Everyone
Claude EveryoneArtifacts can trigger scripts on your machine
▸How-to Claude Everyone
Claude EveryoneNeed to link your AI assistant to Higgs Field
Connect Claude to Higgs Field with a single URL and permission grant
▸How-to Claude Everyone
Claude Everyonedescribe a scene and get an image
Claude auto-creates a detailed prompt and selects the best Higgs Field model
▸How-to Claude Everyone
Claude EveryoneA still picture you want to move
Use the Animate button to add motion and audio to an image
▸How-to Claude Everyone
Claude EveryoneKnow when Fable 5 becomes paid and the per-token rates
!Tip Claude Everyone
Claude EveryoneWho can use Mythos 5 — only Glasswing partners and select cyber defenders
Understand who can access Mythos 5
!Tip Claude Everyone
Claude EveryoneMythos-class tier — one level above Opus in capability
Recognize that Mythos models are more capable than Opus
!Tip Claude Everyone
Claude EveryoneFable 5 vs Mythos 5 pricing parity — twice Opus rate
Both models cost double the price of Opus
▸How-to Claude Everyone
Claude EveryoneKnow how to start a routine from the cloud or external code
▸How-to Claude Everyone
Claude EveryonePrecise natural-language prompts reduce errors in automated routines
▸How-to Claude Everyone
Claude EveryoneUnread Gmail piling up
See a concrete example of an end-to-end routine
▸How-to Claude Everyone
Claude EveryoneWant a routine to run automatically on a set schedule
Configure when a routine should run automatically
▸How-to Claude Everyone
Claude EveryoneExecute a routine immediately and see its I/O in real time
▸How-to Claude Everyone
Claude EveryoneAgent can’t locate its skill files
Prevent `getSkill` failures by listing all skill files in the agent's prompt
▸How-to Claude Everyone
Claude EveryoneAgent skips required skills
Fix missing skill calls by configuring the agent to use only defined skills
▸How-to Claude Everyone
Claude EveryoneWant meaning-based search over markdown files
Enable meaning-based retrieval by indexing text chunks as vectors
▸How-to Claude Everyone
Claude EveryoneFind specific rules or excerpts quickly without full-file scans
▸How-to Claude Everyone
Claude EveryoneWant to see how topics relate to each other
Visualize how concepts interrelate and trace chains of information
▸How-to Claude Everyone
Claude EveryoneEdit markdown files in your vault from VS Code
You can run Claude from VS Code so it can edit markdown files in your vault
▸How-to Claude Everyone
Claude EveryonePull a PDF and web page together
One Claude prompt can read a PDF and a URL, create index entries, logs, and linked markdown pages
▸How-to Claude Everyone
Claude EveryoneProcessed pages all go into one folder
You can change the automatic folder structure from flat to hierarchical and re-run ingests
Use AI to generate ideas but retain control over the final decision for effective outcomes
▸How-to Claude Everyone
Claude EveryoneOnly Pro users can access the task-oriented desktop app
▸How-to Claude Everyone
Claude EveryoneNeed your AI to use Higsfield services
Learn how to link Claude with Higsfield using a simple three-step process
▸How-to Claude Everyone
Claude EveryoneI only have a text idea or reference photo
Turn natural language or an inspiration image into detailed image variations using Higsfield models
▸How-to Claude Everyone
Claude EveryoneNeed a video made from a single picture
Convert a saved image into a 5-second 16:9 video with in/out animations via Higsfield
▸How-to Claude Everyone
Claude EveryoneNeed a single place for project history, files, and instructions
Start a project to keep context and files together
▸How-to Claude Everyone
Claude EveryoneNeed to keep documents linked to a project
Keep PDFs, text, repos, or Drive docs permanently linked to the project
▸How-to Claude Everyone
Claude EveryoneWant custom behavior just for one project
Define Claude's role, output style, and file references for the project only
▸How-to Claude Everyone
Claude EveryoneClaude remembers personal details and facts you mention, reducing repetition
▸How-to Claude Everyone
Claude EveryoneNeed answers from dozens of big uploads
Use Retrieval-Augmented Generation to query many large documents while keeping context low
▸How-to Claude Everyone
Claude EveryoneNeed a quick way to create an account
Create a Claude account quickly using your existing Apple or Google credentials
▸How-to Claude Everyone
Claude EveryoneLocate the main controls for interacting with Claude
▸How-to Claude Everyone
Claude EveryoneNeed structured answers
Structure prompts to produce precise, high-quality responses
▸How-to Claude Everyone
Claude EveryoneNeed up‑to‑date facts
Enable real-time, up-to-date information retrieval
▸How-to Claude Everyone
Claude EveryoneGet visual feedback on screenshots or photos directly in Claude
▸How-to Claude Everyone
Claude EveryoneWant a quick ROI calculator without writing any code
Generate interactive tools like ROI calculators without writing code
▸How-to Claude Everyone
Claude EveryoneWant the same tone in every chat
Set rules that apply to every new conversation for consistent personality or style
▸How-to Claude Everyone
Claude EveryoneChoose the appropriate model for task complexity and cost
Can’t test upcoming AI model
You can preview GPT-5.6 behavior by selecting GPT 5.5 in Pro mode and asking for tasks like landing page
▸How-to Claude Everyone
Claude EveryoneAvoid wrapper frameworks that add latency and regression risk; the core model delivers most gains
▸How-to Claude Everyone
Claude EveryoneNeed to change your app’s libraries or services
Swap libraries or providers by editing a single markdown file in the skill's references folder
▸How-to Claude Everyone
Claude EveryoneNeed extra libraries or a payment gateway?
Add new packages or services by creating a markdown file in the references folder
▸How-to Claude Everyone
Claude EveryoneMy Start‑an‑App skill defaults are outdated
Updating the skill keeps your stack current, eliminating repeated decisions for new projects
▸How-to Claude Everyone
Claude EveryoneUse practitioner, academic, skeptic, economist, and historian perspectives to spot gaps others miss
▸How-to Claude Everyone
Claude EveryoneParallel agents needed for STORM analysis
Run parallel agents, map contradictions, synthesize findings, and peer-review before outputting an HTML briefing
▸How-to Claude Everyone
Claude EveryoneUnsure which lenses disagree
Visualize where agents disagree to focus verification efforts
▸How-to Claude Everyone
Claude EveryoneNeed a research brief template
Output research in an HTML file with summary, ranked findings, and source status
▸How-to Claude Everyone
Claude EveryoneStorm uses ~12 agents, avoiding rate limits and cost; deep research uses ~103 agents
▸How-to Claude Everyone
Claude EveryoneI have an app idea but no blueprint
A coding agent can draft a full architecture diagram and store the plan for later use
▸How-to Claude Everyone
Claude EveryoneNeed to link an AI assistant to n8n
You can link Claude to your n8n instance within a minute by adding the n8n connector
▸How-to Claude Everyone
Claude EveryoneNeed control over logging or blocking execution data
You can choose how execution data is logged or blocked to meet compliance needs
Providing background information yields a more useful first answer
Assigning a role to ChatGPT guides its tone, perspective, and level of detail
Providing the relevant draft or data lets ChatGPT produce feedback that is specific and useful
Specifying an exact output format (bullets, table, item count) forces ChatGPT to deliver answers that are ready to use
A targeted prompt yields a concise, organized summary of a research article
Need a statistical test and chart from my data
Describing the desired statistical test and plot in natural language makes ChatGPT write and run real Python on your file
Always double-check the statistical values ChatGPT reports to ensure they are correct
Avoid uploading unpublished or patient-level data on the free tier to protect privacy
I have a qPCR CSV and need stats and a plot
A ready-to-use prompt lets you quickly get statistics and a plot from your qPCR data
Getting explicit URLs lets you check the original material instead of trusting a summary
Confirming that the source actually says what ChatGPT claims prevents propagation of errors
Want to keep chat data private
Prevent your consumer chats from being used to train the model by turning off the setting in Data Controls
Want AI to always use precise scientific terms and SI units
A concrete prompt you can paste into Custom instructions to enforce precise scientific style in all chats
Even on the free tier you can run any public custom GPT, so no upgrade is needed just to try them
Need to run several analysis steps on a file
You can ask a custom GPT to perform several analytical steps on an uploaded file and return cleaned CSVs or figures for download
Upload a CSV
Providing a clear, step-by-step prompt lets the GPT know exactly how to process your data and format the final result
You can quickly gather up-to-date sources and know which ones need verification before citing
Need a results paragraph and want to flag uncertain citations
You receive a ready-to-use results write-up that only references your own data, while external sources are clearly marked for later verification
▸How-to Claude Everyone
Claude EveryonePosing an actual research or work problem yields useful, actionable answers from Claude
▸How-to Claude Everyone
Claude EveryoneStating the exact output format (checklist, bullets, paragraph) yields a correctly-structured answer on the first try
▸How-to Claude Everyone
Claude EveryoneWrapping each part of the prompt in explicit tags (<context>, <task>, <data>) keeps Claude from mixing them up
▸How-to Claude Everyone
Claude EveryoneYou can get specific answers from the uploaded document without copying its text
▸How-to Claude Everyone
Claude EveryoneAdding a request for a direct quote lets you verify any important statement
▸How-to Claude Everyone
Claude EveryoneNeed a ready‑to‑build diagram or table
You can get a live, editable artifact (flowchart, table, code) by explicitly asking Claude to create something buildable
▸How-to Claude Everyone
Claude EveryoneThe generated artifact opens in its own panel beside the chat and updates in real time
▸How-to Claude Everyone
Claude EveryoneWant to change a diagram on the fly
You can iteratively modify the same artifact simply by describing changes in chat
▸How-to Claude Everyone
Claude EveryoneCan't run code or generate files in artifact panel
Turning on "Code execution and file creation" lets Claude run code inside the artifact panel and generate downloadable files
▸How-to Claude Everyone
Claude EveryoneClaude can generate SVGs, charts, and code but does not produce photographic images
▸How-to Claude Everyone
Claude EveryoneI need a quick question that won’t be saved
Allows one-off queries that won't affect the Project's ongoing summary
▸How-to Claude Everyone
Claude EveryoneNeed a custom thesis methods project setup
Shows a ready-to-use command for creating a specialized Project with detailed instructions
▸How-to Claude Everyone
Claude EveryonePrompting Claude to act as the "toughest reviewer" surfaces the most critical weakness, mirroring real peer-review pressure
▸How-to Claude Everyone
Claude EveryoneNeed to hand off a whole job instead of chatting
You can switch Claude into a task-delegation workflow instead of a turn-by-turn chat
▸How-to Claude Everyone
Claude EveryoneNeed a quick look at my Downloads folder without any changes
You can ask Claude to inspect files and report findings without modifying anything
Want free‑tier AI models but don’t have an account
You can start using Gemini instantly with any Google account and get immediate access to its free-tier models
Don’t know which tone to use
Specifying a role and target audience changes Gemini's tone and depth dramatically
Refine a reply without starting over
Continuing the conversation lets Gemini keep prior context, so each refinement builds on the previous answer
Can't copy text from a screenshot
Images are converted to searchable text, letting you query screenshots of figures or tables
- **Toggle** — turns the Google Workspace connection on for this chat.
- **@Gmail / @Google Docs / @Google Drive** — the three data sources Gemini can search once connected.
- **Example query** — a real "find it in my inbox" use case about a Grand Canyon hiking-trip email.
When you need to know where info comes from
Adding a request for page or section forces Gemini to cite the exact spot in the document
You receive answers that include clickable citations you can verify yourself
- **Arrow chip** — appears at the end of a paragraph when Gemini has related sources to show.
- **Related content and sources panel** — expands below the answer with numbered source cards.
- **Source cards** — link to the actual page (favicon + domain) so you can verify it.
You confirm whether the answer's claim matches what the original page actually says
You catch paraphrase errors or hallucinations before trusting the summary
Need a quick diagram or infographic from a single sentence
You can turn a single sentence into a labelled diagram or infographic instantly
- **Style tiles** — optional presets (Monochrome, Sketch, Cinematic, …) to steer the look before you type.
- **Describe your image** — the prompt box where you describe what you want generated.
- **Create image chip** — confirms this message will generate an image rather than text.
Need to change just a detail in a picture
You can change specific details without regenerating the whole picture, preserving correct elements
Drafting a long document but want real‑time AI help
Use Canvas to draft and revise long documents with live Gemini feedback instead of scrolling through chat
- **Premade by Google** — ready-made Gems you can try immediately, no setup required.
- **Brainstormer / Career guide / Learning coach** — each card shows the Gem’s name and its opening line.
- **Sidebar +** — where you click to create a custom Gem.
Draft Methods need journal‑style formatting
A pre-configured Gem can automatically reformat Methods sections to match high-impact biology journal conventions
Need to turn a research report draft into something interactive
Convert a research report draft in Canvas into quizzes, infographics, or mini web pages for teaching or onboarding
Need a research outline you can check before it runs
You control the research scope by approving or editing the AI-generated plan before it runs
- **Deep Research** — the mode selected in the prompt bar before entering a research query.
- **Sources dropdown** — choose one or more places Deep Research is allowed to search.
- **Search / Gmail / Drive / Chat** — Search is on by default; the rest opt in your own Workspace data.
When I need a detailed research report with citations
The agent delivers a ready-to-use multi-section report with citations and optional audio
Need a ready‑to‑use slide deck from a research report
Exporting to Canvas lets you repurpose raw research into slides, infographics, or quizzes
Need a deep dive on mRNA vaccine platforms for solid tumours
A well-crafted prompt guides the agent to produce a focused, sourced report on a complex topic
Hovering over a citation shows the underlying webpage without leaving the answer
Need a response in a specific format
Requesting a specific answer format (table, bullet list) yields structured results that are easier to verify
Want to tweak an answer without starting over
Using follow-up questions keeps context, sharpens answers, and saves free-tier quota
Can't tell which topics lack reliable sources
Appending a request for items you could not find highlights gaps in the literature before you trust the results
Need recent CRISPR off‑target data
A concrete prompt demonstrates how to combine scope, recency, and shape for a tidy answer
Need only scholarly articles
Select Academic to pull only peer-reviewed papers and preprints, ensuring citable sources
Want every answer to follow your style and citation rules
Ensures every answer follows your preferred style, citation policy, or expertise level
Need a fast market or competitor snapshot
Produces a quick, source-linked overview of commercial offerings
Want to tweak your search results without restarting
Lets you narrow or expand results efficiently by asking targeted follow-ups
Is ChatGPT free, or do I have to pay for it?
ChatGPT has a free tier that gives you access to its current standard model with a cap on the most-capable model before falling back to a lighter one. Paid plans add higher limits and more features, starting around $20/month for ChatGPT Plus. The free tier is fully usable for studying, writing help, and general questions without a credit card.
Does ChatGPT remember what we talked about in a previous session?
It depends on your settings. ChatGPT has a memory feature that, when enabled, stores facts you tell it or that it picks up from conversations — for example, your study focus area — and uses them in future sessions. You can view, edit, or delete these in Settings → Personalization → Manage memories. Memory is separate from conversation history. In a 'Temporary Chat', nothing is saved or remembered.
What is its knowledge cutoff — does it know about recent discoveries in biology?
ChatGPT's training data has a cutoff date, meaning it has no built-in knowledge of events after that point. For anything before the cutoff it can discuss established science; for very recent papers it may be unaware or wrong. The web-search feature (available to all users) bridges this gap by pulling live results, but treat those as leads to verify. For cutting-edge research, use PubMed or Google Scholar directly.
What are custom GPTs, and should I use them?
Custom GPTs are specialized versions of ChatGPT configured with specific instructions, a knowledge base, or connected tools — for example, one tuned for genetics or academic writing. Anyone can use public ones; only paid subscribers can create new ones. You can find subject-specific GPTs in the GPT Store (in the ChatGPT sidebar) that may give more focused help than the standard assistant. Quality varies, so check who built it and test its accuracy.
What is Deep Research in ChatGPT, and can I use it for free?
Deep Research is an AI agent inside ChatGPT that autonomously searches many web sources for several minutes and produces a detailed, cited report — similar to what a research assistant would compile. It is available across plans with monthly query limits that scale with your plan. OpenAI warns it can still make factual errors, so treat its output as a draft that needs verification. For a literature review it can identify themes and papers, which you then check individually.
Can I talk to ChatGPT by voice instead of typing?
Yes. ChatGPT has a Voice Mode that lets you have spoken conversations, and it responds in a natural-sounding voice. On mobile you can also share your camera or screen while talking. Free users get a limited monthly allowance; paid users have higher limits. Tap the microphone or voice icon in the ChatGPT app. It can be handy for reviewing biology concepts aloud while commuting or revising.
What is the difference between the free plan and ChatGPT Plus?
Plus (around $20/month) gives significantly higher message limits, access to reasoning 'Thinking' modes, Deep Research, image and video generation, custom GPT creation, and priority access. Free users are switched to a lighter model once they hit their cap. For occasional homework help, the free plan is usually sufficient; Plus makes sense if you draft long reports daily or need Deep Research.
Can ChatGPT browse the internet and give me up-to-date information?
Yes — ChatGPT Search is available to all users, including free and logged-out users. ChatGPT will automatically search the web when your question involves recent events or time-sensitive data, or you can click the search icon to force a search. Results include links to the original sources. Always check the linked sources yourself, because the model can still misread or misrepresent retrieved pages.
ChatGPT gave me a list of scientific papers to read — are those real?
Not necessarily. Studies show ChatGPT frequently invents academic citations that look entirely real — with plausible author names, journal titles, volume numbers, and DOI-style links — but the papers do not exist. Before attempting to read or cite any paper ChatGPT suggests, search for it directly on Google Scholar or PubMed. If it doesn't appear there, assume it is fabricated.
Can I upload a PDF or image to ChatGPT, for example a journal article or a diagram?
Yes. Free users can upload a limited number of files per day; paid users get higher limits. Supported formats include PDF, DOCX, XLSX, JPEG, and PNG. You can ask ChatGPT to summarize a paper, answer questions about a figure, or extract data from a table. This works well for breaking down a dense methodology section, but verify any factual claims ChatGPT makes about the uploaded content.
?FAQ Claude Everyone
Claude EveryoneWhat are Projects, and should I use them for my coursework or research?
Projects are self-contained workspaces inside Claude with their own chat history and file storage. You upload your own documents (lecture notes, protocols, papers) to a project's knowledge base, and Claude can reference them throughout all conversations in that project. This is useful for organizing materials by course or research topic so you don't have to re-upload files every session.
?FAQ Claude Everyone
Claude EveryoneClaude cannot generate photographs or illustrations the way image-generation tools like DALL-E or Midjourney do. What it can do is create diagrams, charts, and interactive visuals using HTML and SVG code, which appear directly in your chat. It can also analyze and describe images you upload to it. If you need a polished scientific figure, you would create it in a dedicated tool and then use Claude to help you interpret or caption it.
?FAQ Claude Everyone
Claude EveryoneYes, Claude can 'hallucinate' — generating information that sounds authoritative but is not grounded in fact. The Anthropic Help Center explicitly warns that Claude may fabricate quotes, lack up-to-date data, or produce plausible-sounding but incorrect details. For scientific work, treat Claude as a starting point or writing aid, not a primary source: always check its claims against textbooks, original papers, or databases like PubMed. Do not cite Claude as a source in academic work.
?FAQ Claude Everyone
Claude EveryoneHow is Claude different from ChatGPT — which one should I use?
Both are capable AI assistants at a similar price point ($20/month for paid tiers). Claude tends to produce more natural, collaborative writing and is particularly strong at working with long documents — its large context window means you can paste in an entire research paper and it holds the full content in mind. ChatGPT has broader built-in tools including image generation and a web browsing agent. Neither is universally better; many people use both for different tasks. For reading and discussing lengthy scientific papers, Claude's long-context strength is a practical advantage.
?FAQ Claude Everyone
Claude EveryoneCan I upload PDFs, research papers, or images to Claude?
Yes. Claude accepts PDF, DOCX, TXT, CSV, XLSX, and other document formats, plus JPEG, PNG, GIF, and WebP images. In a regular chat you can upload up to 20 files at once, with a maximum of 500 MB per file. Claude can read and analyze text and visual content in PDFs under 100 pages (including charts and figures); for very long PDFs it extracts text only. You can also paste or upload images directly and ask Claude to describe or interpret them.
?FAQ Claude Everyone
Claude EveryoneBy default Claude works from its training data, which has a knowledge cutoff. However, Claude includes a web search feature you can switch on by clicking the slider icon in the chat input and toggling 'Web search.' When enabled, Claude searches the live web and includes direct source citations in its response. This means you can ask about recent publications, news, or current data and get up-to-date answers.
?FAQ Claude Everyone
Claude EveryoneThe context window is the total amount of text Claude can 'see' at once within a single conversation — your messages, its replies, and any uploaded files all count toward it. On paid plans, most models support a 200,000-token context window, which is roughly 500 pages of text. This means you can paste in an entire journal article or lengthy lab report and Claude will keep the full content in mind while responding.
?FAQ Claude Everyone
Claude EveryoneClaude does not automatically carry memory between separate conversations by default — each new chat starts fresh. However, a Memory feature lets Claude automatically summarize key insights from your past chats and use them as context in future conversations. You can turn this on in Settings > Capabilities. If you want a persistent workspace where Claude always has access to specific documents and notes, you can use Projects for that purpose instead.
Want to keep your chats private
You can prevent your chats from being used to train future models
!Tip Claude Everyone
Claude EveryoneArtifacts — a separate panel for large Claude outputs
You get a dedicated view to manage big content generated by Claude
▸How-to Claude Everyone
Claude EveryoneNeed to work with a whole research paper
Allows you to paste a full research paper and keep all content available for reference
▸How-to Claude Everyone
Claude EveryoneDon’t want your chats feeding AI models
You can decide if Claude uses your conversations for training
▸How-to Claude Everyone
Claude EveryoneClaude removes personal identifiers from chats before using them for training
▸How-to Claude Everyone
Claude EveryoneDon’t want your chats used for training
You can change your training preference at any time
▸How-to Claude Everyone
Claude EveryoneWant a private chat that never trains the model
Even if you opt-in, incognito conversations are excluded from training
ChatGPT excels at breaking down difficult concepts into clear, multi-step explanations
What is the context window — how much can I paste into Gemini at once?
The context window is the total amount of text Gemini can read in one conversation. Free users get a sizable window (enough for a lecture handout or a short paper). Paid subscribers get a much larger window — up to around a million tokens, roughly a thousand-plus pages — letting them feed in entire textbooks or large datasets in one session.
Can Gemini help me analyse images — a diagram from a textbook or a microscope photo?
Yes, Gemini is multimodal: you can upload an image and ask questions about it in plain language. Photograph a diagram from your textbook, a chart from a paper, or a microscope image and ask Gemini to describe what it shows, label structures, or explain a process. For highly specialised microscopy images it may still make errors, so cross-check any identifications with your course materials.
Which Gemini model should I use — the fast one or the more powerful one?
The 'Flash' model is faster and optimised for quick, everyday questions. The 'Pro' model is slower but performs much better on complex, multi-step reasoning — detailed essay analysis, interpreting experimental data, or understanding a complicated mechanism. For routine questions in class, Flash is fine; for deep analytical work, Pro gives noticeably better results. Free users get the fast model by default; the Pro model requires a paid subscription.
Can I upload a PDF, image, or document and have Gemini analyse it?
Yes. You can upload multiple files per prompt, including PDFs, images, spreadsheets, and most document types. Gemini can summarise the content, answer questions about it, and generate charts from data within the files, on both web and mobile. The very large context window (enough for many hundreds of pages) is available on paid plans.
How does Gemini integrate with Google Docs, Gmail, and Drive?
You can connect your Google Workspace account to Gemini (with 'Keep Activity' enabled and Smart features on). Gemini can then summarise emails and documents, find information in your Drive files, and create Calendar events — all by asking in plain language. Note that it cannot access images or comments inside files, and school or work accounts require administrator permission first.
What is Deep Research and can I use it for free?
Deep Research is a feature where Gemini automatically browses many websites, cross-references the results, and writes a structured, multi-page report with citations — similar to hours of manual research. Free users get a few Deep Research reports per month; paid subscribers get more and can also search their own Google Drive and Gmail content. The report can be exported to Google Docs and links back to all sources.
▸How-to Claude Everyone
Claude EveryoneClaude consistently outperforms competitors in writing tasks by maintaining natural phrasing, matching requested tones, and avoiding over-formatting. It follows stylistic instructions precisely, making it ideal for client work, emails, and content creation where voice matters.
Gemini demonstrates superior reasoning and coding consistency, especially with large codebases and complex dependencies. It handles multi-step analytical problems better than competitors and maintains context across larger projects without degrading output quality.
Gemini natively supports a 1-million-token context window that accepts text, images, audio, video, and YouTube links simultaneously. This allows you to ask cross-modal questions without switching tools or manually transcribing content.
▸How-to Claude Everyone
Claude EveryoneClaude excels at research when you provide the source material directly. It stays strictly within uploaded documents, provides traceable citations, and flags uncertain information, preventing hallucination and web-search drift.
Only have a short idea and images
Uploading OpenAI’s official Sora 2 documentation into ChatGPT trains the LLM on the exact terminology, formatting, and structural expectations of the video model. ChatGPT then acts as a prompt architect, converting brief concepts and reference images into highly detailed, timecoded scripts that specify camera movement, lighting, pacing, and mood. This bridges the gap between vague text inputs and the model’s actual rendering logic.
Need a prompt to copy an existing picture
By feeding ChatGPT Images 2.0 an image and asking it to "Analyze this image, then write me a prompt that would get ChatGPT to actually create this image," the model returns a detailed textual description that can be used as a generation prompt. The returned prompt captures layout, colors, lighting, typography, and composition, enabling you to generate a near‑identical image in a new chat.
▸How-to Claude Everyone
Claude EveryoneNeed UI screens from a feature list
A concise prompt describing the main feature set (camera, gallery, identification, history) lets Claude generate fully compiling code that integrates into your boilerplate, turning ideas into functional UI instantly.
Need to combine web and Drive data into one report
Gemini’s research mode automatically searches the web, attaches Google Drive files or Notebook LM notebooks, and synthesizes them into a single editable canvas with source citations. This works because it combines live web retrieval with a structured output layout, saving hours of manual cross-referencing.
▸How-to Claude Everyone
Claude EveryoneNeed a multi‑step task run across your apps automatically
Claude’s agentic ecosystem (Co-work for desktop, Dispatch for mobile, and Claude Code for terminals) allows you to send complex, multi-step tasks that run autonomously across your files and applications. This works by chaining tool calls and maintaining context over long sessions, letting the AI handle execution while you step away.
Need to make or understand images and short videos in a chat
Gemini natively generates images and short videos directly within the chat interface and can analyze uploaded media or YouTube links. This works because Google’s infrastructure tightly integrates its diffusion models and video synthesis tools with the conversational UI, eliminating the need for external generators.
▸How-to Claude Everyone
Claude EveryoneHaving to repeat project purpose and rules every time
Claude.md is a markdown file that Claude automatically reads at the start of every session in a project workspace. By writing project context, user profile, rules, and folder structure here you give the agent lasting guidance without re‑prompting each time.
▸How-to Claude Everyone
Claude EveryoneWant the AI to show its step‑by‑step plan first
Planning mode forces the agent to present a written step‑by‑step plan and wait for your approval, preventing it from committing to wrong assumptions early. This small pause saves time correcting downstream errors.
▸How-to Claude Everyone
Claude EveryoneNeed a full research report from a single topic
By writing a plain‑English SOP (research_agent.md) that tells Claude how to clarify scope, plan, research, synthesize, and save the output, you turn a single prompt into a repeatable autonomous agent for any topic.
▸How-to Claude Everyone
Claude EveryoneOne long script needs bite‑size versions
A single SOP describing input location, content extraction steps, platform‑specific copy creation, and PDF conversion lets Claude automatically produce multiple deliverables from one source file, saving hours of manual repurposing.
▸How-to Claude Everyone
Claude EveryoneWant to change only the executive summary in a report
Because Claude retains full project context, you can ask it to modify only a targeted part of a previously generated document (e.g., trim executive summary) without re‑creating the whole file, enabling fast incremental improvements.
Need answers from only my company docs
A Gemini ‘gem’ lets you feed an entire NotebookLM plus extra files into a custom LLM that answers queries using only your proprietary data, providing a secure, domain‑focused retrieval‑augmented generation (RAG) layer.
ChatGPT is the most obedient model; it reliably executes every step in a detailed prompt without skipping. Testing with an optimizer prompt shows its longer, more thorough output compared to other models.
Meeting video, slides, and whiteboard photo?
Gemini can ingest mixed media (video, audio, images, text) in a single request thanks to its multimodal capability and 1 million‑token context window, allowing it to synthesize information from all sources at once.
▸How-to Claude Everyone
Claude EveryoneWhen your code must work on the first try
Claude consistently produces higher‑quality first drafts of code, often working without needing revisions. Supplying clear problem description lets Claude output ready‑to‑run scripts.
Need multiple pictures from one description
By adding a brief header that defines the number of images and then listing specific details for each slide, the model will output multiple distinct images in one request. This works because the model parses the whole prompt as a batch instruction.
Need a specific image shape
The model no longer has a separate UI control for aspect ratio, so including phrases like “16:9” or “9 by 16” in the textual description tells it to render at that size. The model interprets common ratio formats and adjusts canvas dimensions accordingly.
Images missing correct logos or period details
The Intelligence dropdown (instant, medium, high) controls how much reasoning the model applies. Selecting “high” forces the model to draw on its internal knowledge base, producing more accurate contextual details like logos or historical settings.
Want to adjust a photo’s pose, lighting or background via text
By uploading a reference image and describing desired changes (pose, lighting, background, aspect ratio), the model treats the request as an “image‑to‑image” operation, preserving core features while applying edits.
Need to turn a photo into a magical mini‑me scene
Templates pre‑fill a structured prompt (e.g., “turn this photo into a magical mini‑me world”) and automatically add an upload slot, saving time on formatting. You can modify the template text to suit your needs.
▸How-to Claude Everyone
Claude EveryoneNeed code generated and validated on its own
Define a loop that repeatedly calls a builder model to produce code and a judge model to verify it against a 'done' condition. By separating the cheap builder from the more expensive judge you keep token usage low while ensuring quality without human input.
The same set on /recipes, filtered by tool and role.
6Videos 10
The ULTIMATE ChatGPT Guide 2026: How to Use ChatGPT For Beginners
The deeper-dive companion once the basics click. Good for getting more than surface-level value out of the tool.
Gemini Workspace Intelligence Is Here (Connect Gmail, Drive, Sheets, Docs…)
The single most current Gemini-for-work video. Watch this before deciding whether Gemini fits a research or comms workflow you already do in Google Workspace.
Claude AI Tutorial for Beginners (Step-by-Step)
Kevin Stratvert is one of the clearest tech explainers on YouTube. Start here if you've never opened Claude — you'll be productive in 20 minutes.
How to Use Perplexity Comet Browser (2026 Full Guide)
The angle most newcomers are curious about — agentic browsing. Watch this and you'll know whether to switch your daily browser.
ChatGPT Complete Beginners Guide 2026 (How To Use ChatGPT)
A clean current intro. Watch the Data Controls part closely — that's the responsible-use setting most people miss.
Google Gemini Complete Beginners Guide 2026 (How To Use Gemini)
The plain beginner intro. Start here, then watch the Workspace and for-work videos to go deeper.
FULL Claude Tutorial for Beginners in 2026! (Become a PRO!)
The course-style intro if you learn by following along. Watch after Kevin's video to go from 'I can use it' to 'I use it well'.
NotebookLM + Perplexity = Insane RESEARCH Workflow You NEED in 2026
If you write or study, this is the workflow that ties the two tools together. Watch when you're tired of switching between Perplexity tabs.
This Perplexity + NotebookLM Workflow Is Insane!
Pick Lipsky if you want the tight 12-minute version; pick Sadie above if you want the 'why this works for research' framing. Two takes on the same powerful pattern.
Master Gemini 3.1 for Work in 12 Minutes (2026)
The signal-over-noise video from a top productivity creator. Watch to learn what's worth using, not everything that exists.
7FAQ 36
Is ChatGPT free, or do I have to pay for it?
ChatGPT has a free tier that gives you access to its current standard model with a cap on the most-capable model before falling back to a lighter one. Paid plans add higher limits and more features, starting around $20/month for ChatGPT Plus. The free tier is fully usable for studying, writing help, and general questions without a credit card.
Does ChatGPT remember what we talked about in a previous session?
It depends on your settings. ChatGPT has a memory feature that, when enabled, stores facts you tell it or that it picks up from conversations — for example, your study focus area — and uses them in future sessions. You can view, edit, or delete these in Settings → Personalization → Manage memories. Memory is separate from conversation history. In a 'Temporary Chat', nothing is saved or remembered.
What is its knowledge cutoff — does it know about recent discoveries in biology?
ChatGPT's training data has a cutoff date, meaning it has no built-in knowledge of events after that point. For anything before the cutoff it can discuss established science; for very recent papers it may be unaware or wrong. The web-search feature (available to all users) bridges this gap by pulling live results, but treat those as leads to verify. For cutting-edge research, use PubMed or Google Scholar directly.
What are custom GPTs, and should I use them?
Custom GPTs are specialized versions of ChatGPT configured with specific instructions, a knowledge base, or connected tools — for example, one tuned for genetics or academic writing. Anyone can use public ones; only paid subscribers can create new ones. You can find subject-specific GPTs in the GPT Store (in the ChatGPT sidebar) that may give more focused help than the standard assistant. Quality varies, so check who built it and test its accuracy.
What is Deep Research in ChatGPT, and can I use it for free?
Deep Research is an AI agent inside ChatGPT that autonomously searches many web sources for several minutes and produces a detailed, cited report — similar to what a research assistant would compile. It is available across plans with monthly query limits that scale with your plan. OpenAI warns it can still make factual errors, so treat its output as a draft that needs verification. For a literature review it can identify themes and papers, which you then check individually.
Can I talk to ChatGPT by voice instead of typing?
Yes. ChatGPT has a Voice Mode that lets you have spoken conversations, and it responds in a natural-sounding voice. On mobile you can also share your camera or screen while talking. Free users get a limited monthly allowance; paid users have higher limits. Tap the microphone or voice icon in the ChatGPT app. It can be handy for reviewing biology concepts aloud while commuting or revising.
What is the difference between the free plan and ChatGPT Plus?
Plus (around $20/month) gives significantly higher message limits, access to reasoning 'Thinking' modes, Deep Research, image and video generation, custom GPT creation, and priority access. Free users are switched to a lighter model once they hit their cap. For occasional homework help, the free plan is usually sufficient; Plus makes sense if you draft long reports daily or need Deep Research.
Can ChatGPT browse the internet and give me up-to-date information?
Yes — ChatGPT Search is available to all users, including free and logged-out users. ChatGPT will automatically search the web when your question involves recent events or time-sensitive data, or you can click the search icon to force a search. Results include links to the original sources. Always check the linked sources yourself, because the model can still misread or misrepresent retrieved pages.
ChatGPT gave me a list of scientific papers to read — are those real?
Not necessarily. Studies show ChatGPT frequently invents academic citations that look entirely real — with plausible author names, journal titles, volume numbers, and DOI-style links — but the papers do not exist. Before attempting to read or cite any paper ChatGPT suggests, search for it directly on Google Scholar or PubMed. If it doesn't appear there, assume it is fabricated.
Can I upload a PDF or image to ChatGPT, for example a journal article or a diagram?
Yes. Free users can upload a limited number of files per day; paid users get higher limits. Supported formats include PDF, DOCX, XLSX, JPEG, and PNG. You can ask ChatGPT to summarize a paper, answer questions about a figure, or extract data from a table. This works well for breaking down a dense methodology section, but verify any factual claims ChatGPT makes about the uploaded content.
What are Projects, and should I use them for my coursework or research?
Projects are self-contained workspaces inside Claude with their own chat history and file storage. You upload your own documents (lecture notes, protocols, papers) to a project's knowledge base, and Claude can reference them throughout all conversations in that project. This is useful for organizing materials by course or research topic so you don't have to re-upload files every session.
I need a lab‑report figure but can’t draw it — get an SVG chart displayed in chat
Claude cannot generate photographs or illustrations the way image-generation tools like DALL-E or Midjourney do. What it can do is create diagrams, charts, and interactive visuals using HTML and SVG code, which appear directly in your chat. It can also analyze and describe images you upload to it. If you need a polished scientific figure, you would create it in a dedicated tool and then use Claude to help you interpret or caption it.
When you need scientific facts from an AI — it may hallucinate, so double‑check everything
Yes, Claude can 'hallucinate' — generating information that sounds authoritative but is not grounded in fact. The Anthropic Help Center explicitly warns that Claude may fabricate quotes, lack up-to-date data, or produce plausible-sounding but incorrect details. For scientific work, treat Claude as a starting point or writing aid, not a primary source: always check its claims against textbooks, original papers, or databases like PubMed. Do not cite Claude as a source in academic work.
How is Claude different from ChatGPT — which one should I use?
Both are capable AI assistants at a similar price point ($20/month for paid tiers). Claude tends to produce more natural, collaborative writing and is particularly strong at working with long documents — its large context window means you can paste in an entire research paper and it holds the full content in mind. ChatGPT has broader built-in tools including image generation and a web browsing agent. Neither is universally better; many people use both for different tasks. For reading and discussing lengthy scientific papers, Claude's long-context strength is a practical advantage.
Can I upload PDFs, research papers, or images to Claude?
Yes. Claude accepts PDF, DOCX, TXT, CSV, XLSX, and other document formats, plus JPEG, PNG, GIF, and WebP images. In a regular chat you can upload up to 20 files at once, with a maximum of 500 MB per file. Claude can read and analyze text and visual content in PDFs under 100 pages (including charts and figures); for very long PDFs it extracts text only. You can also paste or upload images directly and ask Claude to describe or interpret them.
My AI only knows old info — enable web search for up‑to‑date answers
By default Claude works from its training data, which has a knowledge cutoff. However, Claude includes a web search feature you can switch on by clicking the slider icon in the chat input and toggling 'Web search.' When enabled, Claude searches the live web and includes direct source citations in its response. This means you can ask about recent publications, news, or current data and get up-to-date answers.
Need an AI to read a multi‑page file in one go — it keeps up to 200 K tokens in view
The context window is the total amount of text Claude can 'see' at once within a single conversation — your messages, its replies, and any uploaded files all count toward it. On paid plans, most models support a 200,000-token context window, which is roughly 500 pages of text. This means you can paste in an entire journal article or lengthy lab report and Claude will keep the full content in mind while responding.
My AI forgets earlier chats — it now references past conversation details automatically
Claude does not automatically carry memory between separate conversations by default — each new chat starts fresh. However, a Memory feature lets Claude automatically summarize key insights from your past chats and use them as context in future conversations. You can turn this on in Settings > Capabilities. If you want a persistent workspace where Claude always has access to specific documents and notes, you can use Projects for that purpose instead.
What is the context window — how much can I paste into Gemini at once?
The context window is the total amount of text Gemini can read in one conversation. Free users get a sizable window (enough for a lecture handout or a short paper). Paid subscribers get a much larger window — up to around a million tokens, roughly a thousand-plus pages — letting them feed in entire textbooks or large datasets in one session.
Can Gemini help me analyse images — a diagram from a textbook or a microscope photo?
Yes, Gemini is multimodal: you can upload an image and ask questions about it in plain language. Photograph a diagram from your textbook, a chart from a paper, or a microscope image and ask Gemini to describe what it shows, label structures, or explain a process. For highly specialised microscopy images it may still make errors, so cross-check any identifications with your course materials.
+ 16 more in the library.
8Glossary 63 terms
Show the 63 terms
ChatGPT
o3- OpenAI's most capable reasoning model available in ChatGPT, designed for difficult problems in coding, mathematics, and science that require extended multi-step thinking before responding.
prompt- The message or question you type into ChatGPT — it is the instruction that tells the AI what you want it to do.
context window- The total amount of text (measured in tokens) that ChatGPT can read and hold in a single conversation; content beyond this limit may no longer be considered when generating a response.
token- A small chunk of text — roughly three-quarters of a word — that ChatGPT uses internally to read and generate language; token limits determine how long a conversation or document can be.
Custom Instructions- A settings feature that lets you tell ChatGPT persistent preferences — such as your profession or preferred response style — so you don't have to repeat them in every new chat.
Memory- A feature where ChatGPT saves useful facts you share across conversations (like dietary preferences or your job) so future chats feel more personalised; you can view, edit, or delete saved memories at any time.
Projects- A workspace inside ChatGPT that groups related chats, uploaded files, and shared context under one goal, useful for ongoing work that spans multiple sessions.
Canvas- A side-by-side editing workspace that opens automatically for longer writing or coding tasks, letting you directly edit the output and ask ChatGPT to revise specific sections.
GPTs- Custom versions of ChatGPT built with specific instructions, uploaded knowledge, and selected tools for a particular purpose — for example, a GPT focused on cooking advice or legal summaries.
GPT Store- A public directory at chatgpt.com/gpts where anyone can browse and use GPTs created by OpenAI, partners, and the wider community, organised by category.
Deep Research- A ChatGPT tool that autonomously searches the web, reads multiple sources, and compiles a long-form referenced report on a topic you specify.
Advanced Voice- A mode that lets you speak to ChatGPT and hear it respond in a natural-sounding voice, available on the ChatGPT website, iOS, Android, and Windows app.
Data Analysis- A built-in ChatGPT capability that lets you upload spreadsheets or data files and ask questions about the data, create charts, or run calculations — no coding required.
File Uploads- The ability to attach documents such as PDFs, Word files, or spreadsheets to a ChatGPT conversation so the model can read, summarise, or answer questions about their contents.
Image Generation- A ChatGPT capability that creates original images from a text description you provide, or edits an existing image based on your instructions.
Web Search- A ChatGPT tool that looks up current information on the internet in real time, allowing it to answer questions about recent events or facts beyond its training data.
ChatGPT Plus- A paid subscription tier that gives individual users access to more powerful models, higher usage limits, and features like Deep Research and Advanced Voice.
ChatGPT Agent- A feature that lets ChatGPT perform multi-step tasks on your behalf — such as browsing the web, filling forms, or running code — with minimal input from you.
Claude
<context>- An XML-style tag you wrap around background information in a Claude prompt so Claude can clearly distinguish it from your instructions or question.
<task>- An XML tag recommended by Anthropic for wrapping the specific instruction you want Claude to carry out, keeping it distinct from background context in the same prompt.
<data>- An XML tag used to enclose raw data (such as a table or CSV snippet) inside a prompt so Claude can tell it apart from your instructions and context.
Artifact- A document, diagram, chart, or runnable code snippet that Claude generates in a side panel where you can refine, copy, download, or share it.
Project- A named workspace in claude.ai that holds its own documents and instructions so every chat inside it starts with shared context without re-pasting.
Memory- An optional capability where Claude builds a running summary of your role and ongoing work so new chats start already informed about you.
role- A persona or expert identity you give Claude at the start of a prompt (e.g. 'a careful biostatistics reviewer') to steer the depth and tone of its answers.
web search- A live internet lookup Claude can perform when a question needs current information, returning cited sources you can open and verify.
Gemini
Gemini- Google's personal AI assistant that can answer questions, write text, analyze files, generate images and videos, and connect with Google apps like Gmail and Drive.
prompt- The message or question you type to Gemini to tell it what you want it to do.
context window- The maximum amount of text, files, and conversation history Gemini can read and hold at one time — like its working memory for a single session.
multimodal- Gemini's ability to work with multiple types of content at once — text, images, audio, video, and documents — all in the same conversation.
token- A small chunk of text (roughly a word or part of a word) that Gemini uses to measure how much it has read or generated; larger context windows allow more tokens.
Deep Research- A Gemini feature that automatically searches many sources, spends roughly 5–10 minutes analyzing them, and produces a detailed written report on any topic you ask about.
Deep Think- An advanced reasoning mode exclusive to Google AI Ultra subscribers where Gemini takes extra time to reason through complex or difficult problems before responding.
Gems- Custom AI assistants you build inside Gemini by giving it a name, specific instructions, and optional reference files, so it always behaves like a specialist for a particular task.
Canvas- A side-by-side workspace in Gemini where you can collaboratively create and edit documents, code, slides, apps, and more while chatting with the AI in real time.
Gemini Live- A feature that lets you have a natural, back-and-forth spoken conversation with Gemini using your voice, and optionally share your camera feed so it can see what you see.
Audio Overview- A podcast-style audio conversation that Gemini generates from a document, research report, or notebook so you can listen to the content instead of reading it.
Notebook- A dedicated project space in Gemini where you upload sources (PDFs, Drive files, websites, etc.) and have ongoing conversations that always remember your documents and past discussions.
Connected Apps- Google services (such as Gmail, Drive, Calendar, and Tasks) that you can link to Gemini so it can search your real emails, documents, and events when you ask questions.
Memory- An optional setting where Gemini learns details from your past conversations and uses that context to give more personalized answers in future sessions.
Imagen- Google's AI image-generation model, available inside Gemini, that creates photorealistic or illustrated images from a text description you provide.
Google AI Pro- A paid subscription tier for Gemini that gives 4× higher usage limits and a 1-million-token context window (supporting up to 1,500 pages of text).
Google AI Ultra- The highest Gemini subscription tier, offering the largest usage allowances and exclusive access to features such as Deep Think and enhanced Deep Research visuals.
NotebookLM- A separate Google AI tool focused on document research; Gemini's Notebook feature is integrated with it, so sources and changes sync automatically between both products.
Perplexity
Answer Engine- What Perplexity calls itself — instead of returning a list of links like a search engine, it reads the web in real time and writes a direct answer with cited sources.
Thread- A single conversation in Perplexity where each follow-up question keeps the context from previous turns, so you never have to repeat yourself.
Citations- Numbered source links that appear inline in every Perplexity answer so you can click through and verify the original web page or paper.
Focus- A filter you apply before searching that tells Perplexity which part of the web to look in — for example Academic restricts results to peer-reviewed papers, Social searches Reddit and forums, and Video pulls from YouTube content.
Pro Search- A deeper search mode that breaks your question into multiple sub-queries, consults more sources, and synthesizes a more thorough answer than the default Quick Search.
Quick Search- The default, faster search mode suited for simple factual questions where a brief answer with a few sources is enough.
Deep Research- An autonomous research mode that spends several minutes performing dozens of searches across hundreds of sources and produces a structured, multi-section report — equivalent to asking a research analyst to investigate a topic for you.
Spaces- Collaborative workspaces inside Perplexity where you can group related threads, write custom instructions that apply to every search, and invite teammates to contribute.
Pages- A publishing feature that turns any Perplexity research thread into a formatted, shareable article with a public URL you can distribute or embed.
Connectors- Integrations that link Perplexity to your external apps — such as Gmail, Slack, Notion, or Google Drive — so it can pull live data from those tools into its answers.
Model Council- A premium feature that runs your query through three AI models simultaneously, then uses a fourth 'chair' model to synthesize their responses into one combined answer — available to Max subscribers.
Scheduled Searches- Automated queries you set up once that Perplexity re-runs on a daily, weekly, or monthly schedule and delivers to you as a notification.
File Upload- A feature that lets you attach PDFs, spreadsheets, images, or documents to a conversation so Perplexity can read and answer questions about their contents; free users get a daily limit, while Pro unlocks unlimited uploads.
Sonar- Perplexity's own family of AI models, optimised for fast, accurate web-grounded answers; the base Sonar model is the default search model, while Sonar Pro and Sonar Deep Research handle progressively more complex tasks.
Comet- An AI-native web browser made by Perplexity, built on Chromium, with a built-in assistant for summarising pages and automating multi-step browsing tasks, available on desktop and mobile.
Perplexity Pro- The paid subscription tier (around $20/month) that unlocks unlimited Pro Search, Deep Research, file uploads, model selection, and API credits.
Perplexity Max- A higher-tier subscription above Pro that adds features like Model Council, priority access to new models, and additional advanced capabilities.
Voice Mode- A conversational interface that lets you speak your questions aloud and hear Perplexity's answers spoken back, using a real-time speech model.
Sonar API- A developer interface that lets programmers embed Perplexity's web-grounded search capability into their own applications, using the same request format as the OpenAI API.
9See also
💬 Discuss this chapter
Ask, share, or report — over on the Heidelberg AI community forum.