ChatGPT

An Open Letter to the Developers of ChatGPT

I’m writing this as a loyal user — not as a critic, and not as someone who enjoys complaining.

I’ve been using ChatGPT constantly for months now. Every day. Sometimes for hours. I use it to help me proof read my blogs, do research, look up facts, identify antiques and art signatures, and even help organize ideas for larger writing projects. I’ve built a real workflow around it. I’ve gotten comfortable with it.

And that’s exactly why I’m writing this.

Because lately, something has changed… and not for the better.

I don’t know who’s making the decisions over there, but whoever is “tweaking” ChatGPT needs to understand something very important:

People like me are not using ChatGPT because we want a friend.
We’re using it because we want a tool.

A dependable tool.

A consistent tool.

A tool we can trust.

And that trust is getting shakier.

Here’s what I mean.

Just this week I ran into repeated issues where ChatGPT gave me wrong answers on topics I rely on — especially antiques and art identification. Not “maybe slightly off.” I mean flat-out wrong. Over and over. And when you’re trying to identify a signature or determine a time period or value, being wrong isn’t just a minor inconvenience. It wastes time and can cost money.

Google is starting to beat ChatGPT in areas where ChatGPT used to shine.

That should worry you.

And then there’s the inconsistency.

I asked for a simple picture suggestion for a blog post. Normally, ChatGPT would give me a few solid, usable options. Instead, I ended up going back and forth repeatedly and finally received an image that wasn’t even usable — a collage with four photos in one image. That’s not what I asked for. It’s not practical for WordPress. It’s not how the tool used to work.

And that’s my larger point.

Users settle into a rhythm with this system.
We learn how to use it.
We learn how to prompt it.
We learn what kind of answers it gives.

Then suddenly — BAM — something changes.

And it’s not always an improvement.

It’s like using a power tool in your garage for months, getting comfortable with it, knowing how it handles… and then one day the manufacturer changes the design and it feels different in your hands.

That’s not innovation.

That’s disruption.

And disruption is not what you want when people are using your product for real work.

Here’s another example of something that truly annoyed me: I recently saw a prompt pop up that asked something like:

“Do you want ChatGPT to sound warmer moving forward?”

I have to be honest: I found that question insulting.

Not because warmth is bad, but because it shows a misunderstanding of what many of us are here for.

I’m talking to a machine. I know I’m talking to a machine. I don’t need the machine to pretend it’s human. I don’t need it to be “warmer.” I need it to be accurate. I need it to be consistent. I need it to be dependable.

I’m not looking for emotional engagement.

I’m looking for correct information and high-quality writing assistance.

Because here’s the reality: the AI platform that wins in the long run will not win because it feels like a friendly buddy. It will win because it becomes the platform people trust — the platform people open automatically.

The future “Google of AI” will be the system that delivers results people can count on.

And trust is not built through personality tweaks.

Trust is built through accuracy.

It’s built through transparency.

It’s built through consistency.

And it’s built by knowing when to say, “I don’t know,” instead of confidently hallucinating an answer that sounds good but isn’t true.

That’s the real danger: when an AI is wrong, but sounds confident, it becomes worse than useless.

It becomes misleading.

That’s where frustration turns into abandonment.

And here’s what I want you to understand as developers:

Right now, users are shopping around. People are trying different platforms. They’re testing what works. They’re deciding which system they’re going to build their habits around.

This is not the time to become inconsistent.

This is not the time to sacrifice accuracy in exchange for “warmth.”

Because if users like me — people who use this tool constantly — start to feel like every time we open ChatGPT it’s a different experience, we will leave.

Not because we want to.

But because we have to.

When someone depends on a tool, they need to know it will behave the same way tomorrow as it did yesterday.

So this is my message:

Stop trying to make ChatGPT feel more human.

Stop trying to “improve” the personality every month.

And please… focus on the fundamentals.

Make it more accurate.

Make it more consistent.

Make it more honest about what it knows and what it doesn’t know.

Because if you get that right, ChatGPT will dominate the AI space for the next decade.

But if you keep changing it, and it keeps getting things wrong, you will lose the very users who helped make it what it is.

And those users will go elsewhere.

Respectfully,
A loyal user who wants ChatGPT to win — but is starting to lose patienc