Calvin's Updates

The Latchkey Club Daily Draft — 2026-07-23

Teleprompter / Blog Script

The last time my back flared up, I tried to remember exactly when it started.

Was it when I got out of bed? Was it the day before? Did I lift something? Did I sit too long? Was this the same pain I had last time, or did it just belong to the same unpleasant family?

By the time somebody asks you those questions in an appointment, the answer can become, “It started recently, and it hurts somewhere around here.”

Very precise. Practically a medical chart.

Welcome back to the channel, guys.

Today I wanted to talk about using AI for health questions, because more people are doing it, and I think there’s a useful way to do this that sits between pretending the technology is a doctor and pretending it has nothing to offer.

I saw a KFF poll that found about one in three adults had used an AI chatbot for physical or mental health information during the previous year.

People were looking up symptoms. They were asking for explanations of lab results or diagnoses. They were comparing treatment options and trying to decide whether something was serious enough to see a doctor.

That makes sense to me.

A medical appointment may be fifteen or twenty minutes. The strange lab result can show up in the portal at night. The terminology is unfamiliar. And if you search the internet long enough, a sore elbow can become a rare tropical disease by page three.

AI feels calmer than that. You can ask a follow-up question. You can say, “Explain that in plain English.” It doesn’t look irritated when you ask the same thing twice, and it has never once glanced toward the door because the next patient is waiting.

But calm is not the same as correct.

An AI can produce a clear answer without having examined you, seen your complete history, checked the quality of the information you gave it, or noticed the detail you didn’t know was important.

It can also sound certain when it’s wrong. That’s the part people have to understand. A chatbot does not become worried about being sued halfway through a sentence and suddenly remember humility.

So I don’t want AI deciding what is wrong with me.

I want it helping me prepare a better conversation with the person who is actually responsible for my care.

That’s a different job.

For somebody our age, the preparation problem is real. By our fifties, many of us have a longer medical story. There may be old injuries, medications, procedures, family history, test results, and symptoms that come and go. The information exists, but it’s scattered across memory, portals, folders, and the notes app where it sits next to a grocery list and a measurement for an air filter.

The advantage is that we know our own patterns better than we did at twenty-five.

We know when something feels familiar. We may remember what made it worse last time, what treatment helped, and whether a new symptom appeared after a medication changed.

The risk is that we can become too sure about the pattern.

I’ve had back trouble before, so I may assume every pain is the same back trouble. Experience gives context, but it can also give us a favorite explanation.

That’s where I think AI can help if the assignment is narrow.

Before an appointment, I can give it notes and ask it to organize a timeline.

When did the symptom begin? Where is it? What makes it better or worse? Has it changed? What have I already tried? What medications or supplements am I taking? What questions do I keep forgetting when I’m sitting on the paper-covered table?

The AI can turn a pile of observations into a one-page summary.

Not a diagnosis. A summary.

And I would read every line before using it. If the machine decides I ran a marathon when I actually walked across a parking lot, that needs to come out. My medical history does not need creative writing.

Then I can ask it to help generate questions.

What are the common possibilities a clinician may consider?

What symptoms would usually require urgent care?

What information would help distinguish one cause from another?

What should I ask about the risks, benefits, and alternatives of a treatment?

Are there interactions I should specifically confirm with a pharmacist or doctor?

Notice the wording. I’m not asking, “Tell me what I have.” I’m asking, “Help me know what to discuss and what needs verification.”

That difference matters.

AI may also be useful after the visit.

You leave with a printed page, a prescription, a follow-up date, and a memory of the doctor saying something important while you were still trying to tie your shoe.

You can take your own notes—not a secret recording—and ask AI to reorganize them into plain language: what changed, what am I supposed to do, what should I watch for, and when should I follow up?

Then compare that summary with the official after-visit instructions. If they disagree, the chatbot does not win because it used better headings. You contact the office or pharmacist.

Privacy matters too.

Health information is some of the most personal information we have. Before putting records, test results, medications, or identifying details into any AI system, you need to know how that service stores and uses the data. If you don’t know, leave out names, dates of birth, account numbers, and anything else that isn’t needed for the question.

And some situations are not chatbot situations.

Chest pain. Trouble breathing. Signs of a stroke. Severe bleeding. A dangerous reaction. Thoughts of harming yourself or somebody else. If it feels like an emergency, use emergency services. Don’t spend twenty minutes improving the prompt while your body is trying to get your attention.

I think our hidden advantage at this age is not that we can diagnose ourselves better.

It’s that we’ve had enough experience to understand the value of preparation, second opinions, good records, and asking one more question before making a decision.

We’ve also lived long enough to know that experts can miss things and software can miss things. The answer isn’t blind trust in either one. The answer is using each for the job it can actually do.

The clinician can examine, test, diagnose, prescribe, and take responsibility for medical judgment.

The AI can help organize, translate, compare, and remind.

And I still have to tell the truth, check the summary, protect my information, and follow through.

So my simple rule is this: use AI to make the appointment better, not to make the appointment disappear.

Bring a clean timeline. Bring the medication list. Bring three questions you actually want answered. Ask for plain language when you need it. And when the machine gives you something frightening or unusually reassuring, treat that as a reason to verify, not a reason to make a major decision alone.

As I think more about retirement, health feels less like a side category and more like part of the whole plan. Money matters, but so does staying mobile, thinking clearly, and getting good care before a manageable problem becomes a larger one.

If AI helps me show up prepared and use a short appointment well, that’s useful.

If it convinces me I no longer need the appointment, that may be the most expensive free answer I ever received.

So that’s where I’ve landed for now.

Have you used AI before or after a medical visit? Did it help you ask better questions, or did it create more anxiety?

I’d be curious to hear how you’re using it. Thanks for listening.

Video Prompt Script — Questions to Answer Without Reading

Use these as prompts. Don’t read them on camera; answer them naturally.

1. Cold open: During your last back or sciatica flare, how hard was it to remember exactly when it began and what had changed?
   • Follow-up: What kind of vague answer do we give when the timeline isn’t organized?
2. Why people use AI for health: What did the KFF poll find, and why is the appeal understandable?
   • Follow-up: What happens when results arrive before a doctor is available to explain them?
3. The key distinction: What is the difference between asking AI for a diagnosis and asking it to help prepare for a clinical conversation?
   • Follow-up: Why can a calm, clear answer still be wrong?
4. The 55+ advantage and risk: What patterns do older adults know about their own bodies and medical history?
   • Follow-up: How can that same experience make us too confident in a familiar explanation?
5. Before the visit: What should go into a one-page symptom timeline?
   • Follow-up: Which questions can AI help generate without pretending to decide the diagnosis?
6. After the visit: How can AI help organize your own notes and explain unfamiliar language?
   • Follow-up: What should you do if its summary conflicts with the official instructions?
7. Privacy and urgent care: What personal details should be removed when possible?
   • Follow-up: Which situations should bypass the chatbot and go directly to emergency care?
8. The practical rule: Which jobs belong to the clinician, which belong to AI, and which still belong to the patient?
   • Follow-up: How can AI make a short appointment more useful without making it disappear?
9. Closing: Has AI helped you prepare for a medical conversation, or has it mostly increased anxiety?
   • Follow-up: Invite viewers to share what worked and what they still verify with a professional.

Title Options

1. Use AI Before the Doctor Visit, Not Instead of It
2. The Best Health Question to Ask AI Isn’t “What Do I Have?”
3. How I’d Use AI for a Better Doctor Appointment After 55

Thumbnail / Onscreen Text Options

• DON’T ASK AI TO DIAGNOSE YOU
• MAKE THE APPOINTMENT BETTER
• PREPARE. ASK. VERIFY.

Shorts / Reels Cutdowns

• “Calm is not the same as correct.” Use the contrast between AI’s patient tone and its lack of examination, complete history, and accountability.
• “Not a diagnosis. A summary.” Show the one-page timeline: onset, location, changes, triggers, treatments tried, medications, and forgotten questions.
• “Make the appointment better, not disappear.” Cut the practical division of labor among clinician, AI, and patient, then land on the simple rule.

Viewer Question

Have you used AI to prepare for or understand a medical appointment, and what information do you always verify with a clinician or pharmacist?