Sean and Claude Explain AI To His Dad

What follows is a direct conversation I (Sean, a human author) had with Claude (Opus 4.5) about how to explain AI to my dad. The below is a copy and paste of that exact conversation.

Key Notes and Clarifiers:

• Everything I (Sean) write is in normal font.

• Everything Claude (v Opus 4.5) writes is in italics.

• This is not just an answer to the “how do I explain AI to my dad” question, but also an example of how I converse with Claude, including where I iterate and where I push back.

• Claude is clearly mirroring me in tone and style. There are definitely differences, but my Claude sounds more like me and it’s a good example of how AI looks and feels different to every individual user.

• This is the only time I’ve had Claude write in this blog. I never have it edit my personal blog posts. You can tell when it’s me and when it’s a facsimile of me. So I’m being explicit about the AI inputs in this unique post.

Sean: Hey Claude, I want to have a conversation with you to help me explain AI to my dad. He’s 69 years old, not technical, and runs a medium sized in-person and online training company with a lot of trainers and then standard things like sales, marketing, operations, etc. I think he represents a pretty large audience — people who know AI is or could be important but don’t really get what it is or how to use it beyond just a glorified search engine or copy editor. Let’s talk through how I’d actually explain different concepts to him, and what advice I’d give him as both a consumer and a business owner. This conversation might get published on my blog, where I’d copy and paste everything so people could see the answers and just as importantly, how you and I interact. Ready to go?

Claude: Ready. This is a great framing — your dad’s situation is exactly where a lot of business owners are right now. They’ve heard the noise, maybe tried ChatGPT once or twice, and walked away thinking “okay, so it’s a smarter Google.” Which undersells it dramatically, but also isn’t their fault. The interfaces and marketing have done a poor job of explaining what’s actually different here.

Let’s start wherever you want. We could go conceptual first (what even is this thing, at a non-technical level), or practical first (here’s what it can actually do for your business tomorrow), or some blend. What feels like the right entry point for your dad?

S: Ok I’m going to bounce around a bit. You’re an expert, I’m learning and experimenting, and we’re talking over beers and barbeque. So first question: Why does AI even exist? Why am I now getting inundated with ChatGPT ads and social media posts and all this stuff? Is it really that big of a deal? Because I’ve heard versions of this before (virtual reality, metaverse, etc. will be the future) and it fizzles.

C: Great question, and healthy skepticism. You’re right to be wary — there’s a lot of hype in tech, and most of it doesn’t pan out the way the breathless headlines suggest.

Here’s why this one’s different, in plain terms:

For decades, computers could only do exactly what we told them to do, step by step. Write code, follow rules, execute instructions. They were fast calculators. What changed — and this happened gradually, then suddenly — is that we figured out how to make systems that learn patterns from massive amounts of data instead of following explicit instructions.

The breakthrough that kicked off this current wave was around 2017–2018, when researchers figured out a new architecture (called transformers, if you ever hear that term) that let these systems get dramatically better at understanding and generating language. Then it turned out that if you fed them essentially the entire internet’s worth of text and scaled up the computing power, something unexpected happened: they got generally capable. Not just good at one narrow task, but able to write, reason, explain, code, analyze — across domains.

ChatGPT launched in November 2022 and hit 100 million users faster than any product in history. That’s why the ads are everywhere now. The business opportunity is enormous and every company is scrambling.

Is it as big a deal as people say? Honestly, probably yes — but not in the way most of the hype suggests. It’s not magic, it makes mistakes, and it won’t replace most jobs wholesale. But it’s a genuine step-change in what computers can do. The closest analogy might be the internet itself, or maybe spreadsheets. Tools that didn’t replace humans but changed what a single person could accomplish.

The VR/metaverse comparison is apt, though. Those were solutions looking for problems. This is different because it plugs into things people already do every day — writing, researching, analyzing, communicating. The use case is obvious; the execution is what’s still being figured out.

S: I’ve heard the spreadsheet analogy before. I actually think that’s a good one. Can you explain that a bit further? And try and keep the answers a bit shorter. Dad’s got a short attention span when we start talking tech :)

C: Ha, fair enough.

The spreadsheet analogy: before spreadsheets, if you wanted to do financial modeling or accounting, you needed specialists. Actual humans with green eyeshades crunching numbers by hand or with calculators. Spreadsheets didn’t eliminate the need for financial thinking — they just meant one person could do in an afternoon what used to take a team a week.

AI is similar. Tasks that used to require specialized help — drafting a contract, writing marketing copy, summarizing a 50-page report, analyzing customer feedback for patterns — one person can now do that themselves, reasonably well, in a fraction of the time.

It doesn’t mean the work is gone. It means the leverage changed. Your dad probably doesn’t think of Excel as revolutionary technology anymore. It’s just… how work gets done. AI is heading there, faster than most people expect.

S: Good answer. And when Excel started to replace those teams of people, what happened? There wasn’t just a massive unemployment wave. How did people and the job market adapt? How can they adapt in the AI age to stay securely employed? Yes I know this is crystal ball stuff. Just asking you to do your best speculation.

C: Right, the spreadsheet transition is actually reassuring if you look at what happened.

Accounting didn’t disappear — it evolved. The people who thrived were the ones who stopped seeing themselves as “number crunchers” and started seeing themselves as “people who understand what the numbers mean and what to do about them.” The tool handled the mechanical part. The human handled judgment, context, relationships, decisions.

Same pattern with ATMs and bank tellers, by the way. Everyone predicted teller jobs would vanish. Instead, banks opened more branches (because they were cheaper to run), and tellers shifted toward sales and customer service. Different job, same humans.

For AI, my best guess at who stays secure:

The people who learn to use the tools rather than compete with them. If you’re a copywriter who refuses to touch AI, you’re probably in trouble. If you’re a copywriter who uses AI to produce first drafts in ten minutes and spends your time on strategy and client relationships, you’re more valuable than before.

Also: anything involving physical presence, human trust, or genuine accountability. Your dad’s trainers standing in a room reading people’s body language, adjusting on the fly, building rapport — that’s not getting automated soon.

The danger zone is pure information work that doesn’t require much judgment. Summarizing, reformatting, basic research, simple drafting. That’s where the squeeze comes first.

S: Alright, so as we’re in this adjustment phase, how do you recommend people in the job market learn AI well enough to make it useful for them? And remember, these people have jobs and kids and busy lives. They can’t dedicate hours upon hours to learn AI, especially when it changes so rapidly. So how does a normal person put themselves in a position to at least understand and apply it in a healthy, useful way?

C: Honestly? Just start using it for real work. Not tutorials, not courses, not YouTube videos about AI. Actual problems you already have.

Next time you need to write an email you’re dreading, open ChatGPT or Claude and say “help me write this.” Next time you’re staring at a spreadsheet trying to figure out what story the data tells, paste it in and ask. Next time you’re preparing for a difficult conversation, talk it through with the AI first.

That’s it. Fifteen minutes here and there, on things you were already going to do. You’ll learn what it’s good at, what it’s bad at, and how to talk to it — all through direct experience rather than abstract instruction.

The people I see getting the most value aren’t the ones who took a course. They’re the ones who just started treating it like a coworker they can hand things to. “Here’s a messy situation, help me think through it.” “Here’s a draft, make it better.” “Here’s a document, tell me what I’m missing.”

The rapid change thing actually argues against heavy upfront learning. Anything you study today might be different in six months. But the skill of knowing how to collaborate with these tools — that transfers no matter what changes under the hood.

One concrete suggestion: pick one task you do repeatedly that you find tedious. Try using AI for it three times. By the third time, you’ll know whether it’s useful or not.

S: I notice a lot of your examples are the human drafting or writing something, and then handing it to AI to help improve it. What do you think is the best way to use AI in that type of situation, and how can you do it in a way where the AI has some level of memory? Memory of the topic sure, but also the style, voice, values of the human it works with? How do you “partner” effectively but again, with a finite amount of time and knowledge.

C: Actually, I’d flip your framing. The examples I gave mostly go the other direction — AI drafts, human refines. That’s usually faster and, counterintuitively, often produces better results than starting from scratch yourself.

The workflow that seems to work best: give the AI context and let it take a first swing. Then you react. It’s easier to edit than to create, and your judgment about what’s wrong with a draft is often sharper than your ability to produce the right thing cold.

On memory and voice — this is where the tools are evolving fast. Right now, Claude (me) has memory within a conversation and can build up context over time if you’re using the same account. So one approach is to have an ongoing thread for a particular domain. Your “marketing voice” conversation. Your “strategic thinking” conversation. The context accumulates.

Another practical move: create a short document that captures your style and values. Doesn’t need to be elaborate. “I write in short sentences. I avoid jargon. I value directness over diplomacy. My audience is small business owners.” Paste that in when you start a working session, or save it somewhere you can reference. That’s your shortcut to consistency without re-explaining yourself every time.

The “partnering” mindset you mentioned is the key unlock. Most people either over-delegate (paste something in, accept what comes back) or under-delegate (only use it for minor polish). The sweet spot is genuine back-and-forth. “This is close but the tone is too formal.” “Give me three different angles on this.” “What am I not considering here?”

That’s the skill worth building — learning to direct and react, not just prompt and accept.

S: I get what you’re saying, but I’m still really nervous about trying to learn your style and voice as things continue to grow. I worry that if I have you take the first draft, then I’m adapting to you. And over time, my voice fades and yours remains, even though they would theoretically be similar. A short instruction on “this is how i write” doesn’t seem sufficient. Am I overthinking this? A simple example is marketing. I could have you write all my Instagram posts. But if I did that, I’d always be editing your voice. And then it would be inauthentic. Get what I’m saying? How do we manage this?

C: You’re not overthinking it. This is a real tension and I don’t think there’s a clean answer yet.

Here’s the honest trade-off: AI-first drafting is faster, but you’re right that it pulls toward a mean. These models are trained on enormous amounts of text, so they naturally gravitate toward patterns that are common. Competent, clear, but… smoothed out. The weird edges that make a voice distinctive get sanded down.

A few ways to manage it:

For high-stakes voice stuff (your blog, your Instagram, things that are you), write first. Use AI to punch it up, catch errors, suggest alternatives — but start from your words. The raw material stays yours. This is slower but preserves authenticity where it matters.

For lower-stakes stuff (internal docs, routine emails, first passes on proposals), let AI lead. The efficiency gain is worth it and nobody’s reading those for your distinctive voice anyway.

Feed it your own writing as examples. Not a description of your style — actual samples. “Here are five emails I’ve written that I think sound like me. Write the next one in this voice.” That works better than instructions.

And maybe most importantly: stay in the editing chair. The moment you start accepting outputs without real scrutiny, that’s when the drift happens. The discipline of reacting, pushing back, rewriting chunks — that keeps your voice in the loop.

You’re identifying the right risk. The answer isn’t “don’t worry about it.” The answer is being intentional about where you let the AI lead and where you insist on leading yourself.

S: Alright, write my dad a quick explainer of what AI is, how it can benefit him as a person and business owner, and how he can learn it in a simple, useful way.

C: Here’s a draft:

What AI actually is

You know how computers have always needed exact instructions? Do this, then this, then this. AI is different. It learned patterns from reading basically the entire internet, and now it can write, analyze, and reason in ways that feel surprisingly human. It makes mistakes and it’s not magic, but it’s a genuine shift in what computers can do.

Think of it like spreadsheets. Excel didn’t replace accountants — it just meant one person could do what used to take a whole team. AI is doing that for writing, research, analysis, and communication.

What it can do for you

As a person: draft emails you’re dreading, summarize long documents, think through decisions, prepare for difficult conversations.

As a business owner: create training materials faster, analyze customer feedback at scale, generate marketing content, onboard new hires more efficiently, get a thought partner for strategy questions at 11pm when no one else is around.

How to learn it

Skip the courses. Just start using it for real work. Next time you’re stuck on an email, open ChatGPT or Claude and ask for help. Next time you’re preparing a presentation, hand it your notes and see what comes back.

Fifteen minutes here and there on problems you already have. That’s it. You’ll figure out what it’s good at through experience, which is worth more than any tutorial.

Want me to adjust anything?

S: That’s good. So here’s a follow up question about the learning piece. How does he know he’s doing it “right”? Any advice there? I know that’s subjective, but it’s a real concern. You don’t want to spend time playing with it and just going in circles and not realizing it.

C: Practical answer: you know it’s working if it saves you time or makes the output better. That’s it. If you spent twenty minutes wrestling with the AI and could have just written the email yourself in five, that wasn’t the right use. If you got a first draft in two minutes that you edited into something good in three more, that was.

A few signs you’re spinning your wheels:

You’re prompting the same thing over and over hoping for a different result. If the third attempt isn’t getting closer, change your approach — give it an example of what you want, or break the task into smaller pieces, or just do it yourself.

You’re accepting outputs that feel “off” because you don’t want to waste the effort. That’s backwards. The AI is free. Your judgment is the valuable part.

You’re using it for things you actually enjoy doing. If writing is how you think and you find it satisfying, don’t outsource it just because you can.

Signs it’s working:

You’re getting to “good enough” faster. Not perfect, but usable, with less friction.

You’re using it reflexively for certain tasks. The same way you don’t think about whether to use Google — you just search. When AI becomes that automatic for specific tasks, it’s landed.

You’re getting better at directing it. Your prompts get shorter and the results get closer to what you want on the first try.

Honestly, tell your dad to give himself permission to abandon attempts that aren’t working. That’s not failure — that’s learning the boundaries.

S: Anything else you think my dad would ask that we should answer here before I close this convo?

C: A few things I’d expect him to wonder:

“Is this safe? What happens to the stuff I put in there?” Fair concern. Don’t paste sensitive client data, financials, or anything confidential without understanding the tool’s privacy policy. For general business writing and brainstorming, it’s fine. But treat it like talking to a stranger in a coffee shop — don’t share what you wouldn’t want floating around.

“Which one do I use? There are like fifty of these things.” Start with ChatGPT or Claude. Both are free to try. Pick one and stick with it for a month rather than bouncing around. The skills transfer.

“What about my trainers? Should I be worried they’ll use this to phone it in?” Maybe worth a conversation with your team about how you want AI used. Some companies embrace it, some set guardrails. The risk isn’t that they’ll use it — they probably already are. The risk is not having a shared understanding of where it helps versus where it undermines the value you’re selling.

“Is this going to make my business obsolete?” Training is fundamentally about human presence, trust, and transformation. AI can help with curriculum development, content creation, follow-up materials — the stuff around training. But a room full of people with a skilled facilitator reading the room? That’s not going anywhere soon. If anything, AI might make that human element more valuable, not less.

Anything else you want to cover, or does that feel like a reasonable place to land?

S: That’s it. Good chat.