An AI digital caregiver is probably the wrong dream for our generation.
In recent months, I have noticed something curious, slightly disturbing, and, in its own way, almost tender.
Many people over 50, and often over 60, look at artificial intelligence as if they had finally found the perfect caregiver: patient, polite, always available, always ready to answer any question, correct a text, write an email, summarize a document, suggest a recipe, reassure their ego and, ideally, never contradict them too much.
A kind of total assistant, with no holidays, no bad moods and no payroll taxes.
The problem is that this view of AI is, in most cases, completely wrong.
Or rather: it is small. Very small.
Because using AI only to get help with marginal tasks is a bit like buying a Ferrari to take the recycling down the street. Technically, it works. But perhaps that was not exactly the point of the investment.
The 1980s, computers, and the people who only played games
I started working with computers in the 1980s.
I saw computers enter homes, offices and companies. I saw the boom of the 1990s in Italy, when the computer suddenly became the magical object that was supposed to “change everything”.
And it did change everything.
But not for everyone in the same way.
There were those who sat in front of the screen watching animations, games, graphic effects, demos and colorful screensavers. They were fascinated by the machine, but they remained spectators.
Then there were those who tried to actually use the software. Maybe badly, maybe slowly, maybe making a thousand mistakes. But they were trying to build something useful: a customer database, an invoicing system, a price list, a warehouse archive, a better way to manage work.
The difference was already clear.
On one side, the computer as entertainment.
On the other side, the computer as a productive tool.
Today, with artificial intelligence, the exact same thing is happening again. Only much faster.
Today we still have spectators and entrepreneurs
Today, many people use AI as an advanced toy.
They ask it to write a nicer sentence, summarize an article, invent a joke, fix an email, explain something they will probably not read until the end.
All legitimate, of course.
But that is not where real value is created.
Real entrepreneurs, real professionals, the ones who are trying to understand what is actually happening, do something different.
They do not ask AI: “Can you help me do this small thing?”
They ask: “How can I transform this process? How can I sell better? How can I reduce time, errors and costs? How can I create a new product? How can I make what I already do more useful, more scalable or more marketable? How can I get there before my competitors?”
That is the real difference.
AI is not interesting because it can write a polite email.
AI is interesting because it can change an operating model, a service, a workflow, a commercial process, a digital product, an internal procedure, a way of selling or a way of supporting customers.
But to get there, one thing is required that AI cannot provide by itself: direction.
The problem is not using AI. The problem is using it without knowing where you are going
Many people are using AI obsessively.
They open ChatGPT, Claude, Gemini, Copilot, Perplexity, Grok and any other tool with a futuristic-sounding name, and then they start asking for everything.
- “Write this for me.”
- “Correct that for me.”
- “Create a script.”
- “Summarize this.”
- “Give me a strategy.”
- “Build me a business.”
- “Make me rich by Friday, preferably without leaving the house.”
The point is that AI will answer. In fact, it always answers. That is exactly the problem.
It answers even when the question is confused.
It answers even when the objective is weak.
It answers even when the context is incomplete.
It answers even when the person using it has no real intention of turning that answer into action.
And so we get a tragicomic phenomenon: people who are convinced they have become more productive, while in reality they are simply spending more time talking to a machine.
In the past, people wasted time in useless meetings.
Now they waste time asking AI to prepare the minutes of meetings that should never have happened in the first place.
The perfect example: FileMaker, scripts and the illusion of intelligent automation
In my field, Claris FileMaker development, this situation is particularly visible.
Today there are developers using AI compulsively to help create FileMaker scripts. They copy pieces of logic, ask for translations, ask for optimizations, ask for formulas, ask for JSON structures, loops, validations, error handling and privilege checks.
So far, so good.
Actually, very good.
The problem begins when AI is used without a real architectural understanding behind it.
Because if a developer does not understand the data model, does not understand the user workflow, does not understand the relational context, does not understand what should happen in case of an error, does not know whether the script should run on the client or on the server, does not understand the performance impact, and does not understand the real meaning of privilege management, then AI does not solve the problem.
It simply makes the problem look more elegant.
Maybe it produces a cleaner script, better comments, a more readable structure.
But if the original logic is wrong, it remains wrong.
Only with better formatting.
I have seen people spend hours asking AI to improve a FileMaker script that should not have been improved at all. It should have been thrown away and rewritten from scratch.
I have seen endless attempts to generate complicated formulas to work around problems that were caused by a bad relationship graph, a field used for three different purposes, or a table created in an emergency in 2011 and then never touched again out of fear.
And here comes the ironic part.
Because, let’s be honest: after seeing some FileMaker scripts written by human developers, it is hard not to admit that even the worst AI could probably do better.
- Scripts with 400 lines and not a single useful comment.
- Scripts that open windows as if they were hotel doors.
- Scripts that change layout five times just to calculate a sum.
- Scripts that use global variables like sticky notes on a refrigerator.
- Scripts where an error is not handled; it is simply ignored with optimism.
- Scripts where the button name says one thing, the script name says another, and the final result is a third undocumented possibility.
In those cases, yes: even a blindfolded AI connected to a toaster would probably produce something more readable.
But that is still not enough.
Because the value is not in generating code.
The value is in knowing which code should exist in the first place.
The real question is not “Can AI do this?”
The real question is:
Does this thing make sense?
This is where many people go wrong.
They ask AI to improve something useless.
They ask AI to speed up a broken process.
They ask AI to automate a procedure that should not exist at all.
They ask AI to create content nobody will read, scripts nobody will maintain, reports nobody will use and strategies nobody will apply.
This is the great illusion of artificial productivity.
You look busy.
You look modern.
You look up to date.
You may even be paying for the premium plan, so you feel almost like a pioneer.
But in the end, you have not created a better product, sold more, reduced a significant cost, improved a customer experience or built a competitive advantage.
You have simply talked a lot with AI.
Which is pleasant, sometimes useful, but not necessarily productive.
AI does not replace strategic thinking
Artificial intelligence can be extremely useful.
It can accelerate analysis, writing, development, testing, documentation, marketing, customer support, research, design, debugging, estimation, code review and many other activities.
But it does not replace the ability to understand what really matters.
It does not replace experience.
It does not replace judgment.
It does not replace responsibility.
It does not replace the courage to make a commercial decision.
It does not replace the ability to look at a system and say: “This should not be optimized. This should be redesigned.”
And this is exactly where experienced professionals should have a huge advantage.
Those who lived through the computer revolution of the 1980s and 1990s should know very well that every new technology creates two categories of people:
those who passively enjoy it;
and those who turn it into value.
The risk for mature professionals: confusing familiarity with competence
Many mature professionals have a real advantage: they know real problems.
They have seen disorganized companies, badly designed software, absurd processes, difficult customers, slow databases, broken management systems, and terrible decisions that were defended for years as if they were cultural heritage.
That experience is gold.
Because AI needs real context.
The problem is that many people, instead of using their experience to guide AI, allow AI to guide them as if they had finally found a young assistant who smiles politely and “knows how to do everything”.
But AI does not know how to do everything.
AI combines, suggests, rewrites, accelerates and proposes.
If the question is mediocre, it often produces a mediocre answer in a very convincing form.
And this may be the most dangerous thing of all: well-written mediocrity.
Before, confusion was visible.
Now it arrives nicely formatted.
Using AI properly means asking uncomfortable questions
A serious use of AI is not:
“Write me a post.”
It is:
“Analyze my market, find a positioning, identify an offer that can actually be sold, build a contact sequence, help me test it and measure what works and what does not.”
It is not:
“Create a FileMaker script for me.”
It is:
“This process currently requires 12 manual steps, generates errors, blocks users and slows down the system. Help me redesign it, identifying what should stay inside FileMaker, what can run server-side, what can be exposed through an API, what must be logged and what must be controlled by privileges.”
It is not:
“Write me a more professional email.”
It is:
“This email must generate a reply, overcome a commercial objection and move the client toward a concrete decision. How should we structure it?”
The difference is enormous.
In the first case, AI is a secretary.
In the second case, AI is a strategic amplifier.
The point is not to look modern. The point is to monetize, improve, sell and build
In the professional world, using AI should not be a hobby.
It should have a purpose:
- Reduce time.
- Increase quality.
- Generate leads.
- Improve a product.
- Create a new service.
- Make an existing skill more marketable.
- Automate a repetitive task.
- Find errors before they become expensive.
- Document better.
- Support customers better.
- Create a concrete advantage.
Everything else is technological entertainment.
Beautiful, sometimes fun, but still entertainment.
And this takes us back to the 1990s.
There were people watching colorful screensavers and saying: “Amazing what computers can do.”
And there were people building databases, management systems, workflows, accounting tools, catalogs and automations.
Today it is exactly the same.
Only now the screensaver talks back.
Conclusion: AI is not here to take care of us
AI is not the digital caregiver of our generation.
It is not here to make us feel less alone in front of a computer.
It is not here to write sentences for us while we continue to avoid making decisions.
It is not here to automatically transform a confused idea into a business.
AI is an extremely powerful tool.
But it is still a tool.
And like every powerful tool, in the wrong hands it produces noise, waste and fake efficiency.
In the right hands, it can become a huge lever.
For software developers.
For consultants.
For companies.
For people working with databases, processes, customers, documents, automation, APIs, legacy systems and real problems.
So the question is no longer:
“Do you use AI?”
Everyone uses AI now.
The real question is:
Are you using it to create value, or just to keep yourself company while wasting time more efficiently than before?
Are you using AI, or is AI using your time?
This is the uncomfortable question every professional should ask before opening another AI chat and requesting another perfectly formatted answer to a poorly defined problem.
If your FileMaker system is slow, messy or difficult to maintain, you can contact Lolli Group here for a technical review.