Getting Started with AI

No jargon. No hype. Just honest guidance on how to start using AI today.

The #1 rule: just use it

Seriously. The single best thing you can do is open Claude, ChatGPT, or Gemini and start a conversation. Don't worry about "prompt engineering" or doing it "right." Just talk to it like you'd talk to a smart colleague.

"Ask Claude/ChatGPT to ask you questions about yourself — your knowledge, experience, interests, and goals. Then ask it to suggest ways you can work together."

That single prompt will teach you more about what AI can do for you specifically than any article or course.

Basic Practices

Be specific about what you want

Instead of "write me an email," try "write a professional email to my team announcing that we're switching to a new project management tool, keep it upbeat and under 200 words." The more context you give, the better the output.

Iterate, don't start over

If the first response isn't quite right, don't retype everything. Say "make it shorter," "more casual tone," or "actually, focus more on the timeline." AI conversations are collaborative — build on them.

Use it for thinking, not just output

AI is great for brainstorming, exploring ideas, and pressure-testing your thinking. Ask it to play devil's advocate, list things you might be missing, or explain a concept from a different angle.

Verify what matters

AI can be confidently wrong. For anything factual — dates, statistics, legal information, medical advice — double-check against reliable sources. Use AI for the draft, your judgment for the final call.

Share context generously

Paste in the document you're working on, describe the project, explain who the audience is. AI doesn't know what you know unless you tell it. More context = dramatically better results.

Ideas to Get You Going

Personal

  • Plan a trip with specific preferences and constraints
  • Draft difficult emails or messages you've been putting off
  • Summarize a long article, paper, or book chapter
  • Create a study plan for something you want to learn
  • Get recipe ideas based on what's in your fridge
  • Practice for a job interview with mock questions

Professional

  • Analyze spreadsheet data and find patterns
  • Draft proposals, reports, or presentations
  • Create meeting agendas from rambling notes
  • Generate first drafts of documentation
  • Brainstorm solutions to a problem with structured frameworks
  • Review and improve your own writing

Creative

  • Write a short story or poem with collaborative iteration
  • Generate names, taglines, or brand concepts
  • Explore design ideas through detailed text descriptions
  • Build a simple website or app with AI-assisted coding
  • Create a newsletter or blog with AI as your co-editor

Technical

  • Debug code by pasting errors and asking for explanations
  • Generate unit tests for existing functions
  • Prototype APIs and data models from descriptions
  • Refactor messy code with explanations of the changes
  • Learn a new programming language by building small projects

Addressing Common Concerns

"Isn't AI going to take my job?"

AI is much more likely to change your job than eliminate it. The people who will thrive are the ones who learn to work with AI as a tool — the same way people who learned spreadsheets outperformed those who stuck to paper ledgers. Start learning now and you'll be ahead of the curve, not behind it.

"I'm not technical enough for this."

Modern AI tools are designed for conversation, not code. If you can write an email, you can use AI. Start with Claude or ChatGPT — they're literally built to be talked to in plain language. The L1 and L2 content on this site is specifically curated for non-technical people.

"Isn't everything AI generates wrong / hallucinated?"

AI can make mistakes, especially with specific facts. But it's remarkably good at drafting, brainstorming, explaining concepts, and processing information. Think of it as a brilliant intern — incredibly fast and capable, but you still review the work. The key is knowing when to trust and when to verify.

"It feels like cheating."

Using a calculator isn't cheating at math. Using a search engine isn't cheating at research. AI is a tool that amplifies your capabilities — the thinking, judgment, and creativity are still yours. The best output comes from people who bring real expertise and context to the conversation.

"I don't want AI to have my data."

This is a legitimate concern worth taking seriously. Most major AI providers have clear data policies — Claude and ChatGPT both offer options where your conversations aren't used for training. Read the privacy policies, use the privacy controls, and don't share anything you wouldn't put in an email.

"AI is terrible for the environment."

The environmental impact is real but often overstated. Data centers use about 1.5% of global electricity (IEA, 2025), and a ChatGPT query uses roughly the same energy as a Google Search (0.3-0.5 Wh per query, per Epoch AI). The bigger picture: the IEA estimates AI applications in transport, buildings, and industry could reduce emissions by 1,400 Mt CO2 by 2035 — 3-4x more than total data center emissions. See our detailed breakdown below.

AI & the Environment: What the Research Actually Says

Environmental concerns about AI are legitimate and worth understanding with real numbers. Here's what credible research tells us — the good, the bad, and the nuanced.

Energy Usage

  • Data centers consumed about 415 TWh in 2024 — roughly 1.5% of global electricity. Projected to reach ~945 TWh (3%) by 2030. (IEA, "Energy and AI," 2025)
  • A ChatGPT query uses roughly 0.3-0.5 Wh— comparable to a Google Search. The widely cited "10x more energy" claim is based on older hardware and models. (Epoch AI, 2025; MIT Technology Review)
  • AI-specific workloads account for about 14% of data center electricity today, projected to reach 40% by 2026. (Lawrence Berkeley National Laboratory)

Emissions & Renewables

  • Google, Microsoft, and Meta all match 100% of annual electricity with renewable energy purchases. Google's data centers run at a PUE of 1.09 vs. industry average 1.56. (Google 2025 Environmental Report)
  • However, absolute emissions are still rising — Microsoft's are up 23.4% since 2020, Meta's up 60%+. Annual renewable matching is not the same as 24/7 carbon-free energy. (Microsoft & Meta 2024 Sustainability Reports)
  • Data centers account for an estimated 2.5-3.7% of global GHG emissions, comparable to aviation (2.4%). (Carbon Brief, 2025)

Water Usage

  • US data centers consumed about 17 billion gallons of water in 2023, projected to double to quadruple by 2028. (Lawrence Berkeley National Laboratory / US DOE)
  • Roughly 20-50 ChatGPT queries consume about half a liter of water for cooling. Training GPT-3 used approximately 700,000 liters directly. (Shaolei Ren et al., UC Riverside, 2023)
  • Indirect water use from electricity generation is 12x greater than direct cooling — the energy source matters more than the data center itself. (Lawrence Berkeley National Laboratory)

The Other Side: AI for Climate

  • The IEA estimates AI applications in transport, buildings, and industry could reduce emissions by 1,400 Mt CO2 by 2035 — 3-4x larger than total data center emissions in the same timeframe. (IEA, "Energy and AI," 2025)
  • Google DeepMind reduced data center cooling energy by up to 40% using AI optimization, with ~30% average savings in ongoing deployments. (Google DeepMind)
  • Proven applications include grid optimization, climate modeling (DeepMind's GraphCast), predictive maintenance for energy infrastructure, and smart building management.
  • Caveat: The IEA notes "there is currently no momentum that could ensure the widespread adoption of these AI applications, so their aggregate impact could be marginal if necessary enabling conditions are not created."

Sources: IEA "Energy and AI" (2025), Google/Microsoft/Meta Sustainability Reports (2024-25), Lawrence Berkeley National Laboratory, Epoch AI, MIT Technology Review, Carbon Brief, UC Riverside.