AI for Everyone

AI for Everyone

This article was inspired by my experiences in learning to work with AI and help others do so, especially my mother. She not only uses ChatGPT, but Claude as well, both with sophisticated organization, and I wouldn't be surprised if she's delved into Perplexity. For the insatiably curious and those who love to learn and create, AI provides unprecedented opportunity.

Are you using ChatGPT as an improved search engine? Potentially you've seen how useful it can be with how-to advice for tasks, assisting with writing, or help with understanding by summarizing articles or other content. If not, it can certainly do all of those things, but it can do much more. This is about using AI as an expansion of your mental abilities, not a replacement for them.

Basic Organization

Use separate chats for separate conversations and topics, and name them yourself. AI will pick a default name, but you can do better, whether it's "Kitchen Renovation", "Job Search 2026", "Baking Cheesecake", or "Financial Planning". Then you can easily return to the same chat later. Note: you need to be logged in for this to work. I recommend always chatting logged in.

Reflection & Expansion

AI is remarkably good at active listening -- helping you think and work through something by engaging with and expanding your ideas instead of just giving you answers. Wrestling with a decision? Stuck on a creative problem? Having something that can reflect your thoughts back, ask clarifying questions, and help you find your own answer is extremely useful.

For this to work well, you need to give it relevant information. Don't ask it "what should I do to find a better job?", but tell it your situation, thought process, goals, and criteria. Just talk to it, like you are talking with someone who will never get tired of listening and never run out of energy. Have a crazy day and a head full of more thoughts than any human could ever listen to? Trying to make sense out of a possible career change? AI can process it all, reflect, and continue to iterate with you. Big enough it can't possibly all be resolved right now? Name your conversations (and use search) and you can pick it up later.

Pick It Up Later

One of the biggest expansion advantages. Most complex subjects, decisions, or processes aren't going to be resolved in one conversation. Usually they develop over time, and we need time to process them. But jumping back into something complex, whether it's "Life Goals 101", or "Molecular Biology of Frogs" is difficult for us.

Not only does AI have the whole conversation, but you can just ask it to help you. Tell it something straightforward like "Just getting back to this after quite a while, life threw some curves. Where were we, and what were we specifically working on?". This by itself almost completely removes a huge hurdle to juggling multiple complex projects as humans.

Time

Speaking of picking it up later, it's important to know that AI does not experience the passage of time, and its perception of time is completely different than ours. It may assume the current date is different, and base it on its last training date. It's very helpful to tell AI the date, especially if working on anything time-centric, for example current events.

If you're returning to a task, tell it. After the "what were we specifically working on", you could just add "Today is January 12, 2026". Use your favorite format, for example mine would be "20260112". This not only helps the AI understand the timeframes, it anchors it, reducing hallucination, like giving the wrong answer when asked things like "who is the president of the USA"? If in doubt, insert times more often.

Hallucinations and Verification

AI can give a wrong answer with 100% confidence, known as an hallucination. There's some "it's complicated" behind this but for now remember -- AI was trained on our data, writings, and recorded experiences. Like humans, it will avoid saying "I don't know" whenever possible.

Misunderstanding of time is one common cause that's easy to fix. You cannot prevent all hallucinations. There are clues, for example if it starts a response then seems to change its mind with phrases like "But wait..." or "No that's not right I need to think this through". Those are clues that it is guessing.

Asking it to verify or explain how it decided something is helpful, but asking another AI is often better. Try copying the answer from one (ex: ChatGPT), pasting it into another (ex:Claude) with an explanation like "This is the information I have about {subject} and I need to determine its accuracy. Please analyze, evaluate accuracy, and provide sources".

Tools -- Where to Start?

Starting with ChatGPT is fine. Add at least one more based on preferences: Claude for complex thought and natural conversation, Grok for current events, Gemini (Google) is another good generalist and particularly good at helping with Google apps. You don't need to get them all, but they do all have useful free accounts.

Get Perplexity. It is the hands-down best for many detailed research tasks and article analysis. Use its Deep Research mode when you want a thorough answer. Limited to a few uses per day on free account, so save that mode for when it counts.

They are all available on iOS, Android, and web.

Specific Tips:

  • Use multiple AIs. My suggestion: ChatGPT, Claude, and Perplexity. Expand when curious
  • The more information you give AI, the better it can help you.
  • Tell it your thought process. "Think out loud" to it.
  • Tell AI the time and date, frequently.
  • If AI seems uncertain, push back, verify, and ideally check elsewhere for consistency.

That's the foundation. There's a lot more, but this should help you get started.