I recently got some insight on my AI insights from an AI.

I decided to ask ChatGPT:

I’ve written a fair bit about AI, see tags: https://daily-ink.davidtruss.com/tag/artificial-intelligence/
and https://daily-ink.davidtruss.com/tag/ai/, please share with me:
1. What is a prediction I most likely will get right.
2. Where am I missing the mark?
3. What’s the most uniquely insightful idea that I’ve shared? 

Here is ChatGPT‘S response… and I find it quite insightful:

I enjoyed thinking about this because your AI writing has a recognizable worldview. It isn’t a collection of disconnected opinions—there are a handful of core principles that recur across many posts. 

What follows isn’t a prediction of whether you’ll be “right” or “wrong.” It’s my best assessment of which ideas I think are most robust, and where I think your model of the future could be incomplete.

1. The prediction you’re most likely to get right

Your prediction that AI’s greatest near-term impact will come from amplifying ordinary people—not from becoming an autonomous supervillain.

This idea appears repeatedly in your writing.

Rather than worrying primarily about “AI takes over the world,” you focus on AI enabling:

• bioterrorism,
• industrial-scale scams,
• personalized fraud,
• cybercrime,
• misinformation,
• one malicious individual having disproportionate power.

I think history is already bending in your direction.

Technology almost always scales existing human intent before it replaces humans.

The printing press spread propaganda before democracy.
The internet scaled fraud before education.
Social media amplified outrage before civic engagement.

I suspect AI will follow the same pattern.

I think twenty years from now people may look back and realize that “AI as force multiplier for humans” was a more important story than “AI replacing humans.”

That feels like one of your strongest forecasts. 

---

2. Where I think you’re missing the mark

Interestingly, I don’t think you’re too pessimistic.

I think you’re not pessimistic enough about adaptation.

Your writing often assumes exponential technological improvement translates fairly directly into exponential societal disruption.

Technology often is exponential.

Institutions rarely are.

People aren’t either.

For example:

Autonomous driving may become technically solved long before regulators, insurers, unions, municipalities, and consumers allow full replacement.

Likewise, AI lawyers, teachers, doctors and programmers may become technically capable years before society reorganizes around that capability.

You’ve written about the “frog in the pot.”

I wonder if there’s another metaphor:

society as a giant flywheel.

The flywheel eventually turns—but much more slowly than the engine powering it.

---

A second place I think you’re underestimating something:

Humans don’t optimize as much as economists think they do.

Many of your work-related predictions assume organizations will aggressively replace humans because it’s economically rational.

Sometimes they will.

But humans also buy:

• trust,
• relationships,
• status,
• ritual,
• authenticity.

Coffee shops still exist despite home espresso machines.

Live concerts thrive despite Spotify.

Teachers continue to matter despite YouTube.

I think there will be more “human luxury markets” than your writing currently anticipates.

Ironically, AI may increase the value of demonstrably human experiences.

---

3. The most uniquely insightful idea you’ve shared

This one surprised me.

I don’t think it’s UCI over UBI.

I don’t think it’s AI risk.

I don’t even think it’s “This is the worst AI it will ever be.”

Those are all interesting.

The idea I keep coming back to is this:

Students choose. AI delivers.

It’s deceptively simple.

But it’s actually a profound inversion of how educators have framed educational technology for decades.

Most discussions ask:

“What should teachers do?”

Your framing asks:

“What kind of learner will this student choose to become?”

That’s a much deeper question.

Because AI is fundamentally an amplifier.

It amplifies curiosity.

It amplifies laziness.

It amplifies expertise.

It amplifies ignorance.

The technology isn’t making the decision.

The learner is.

I think that’s an idea with legs well beyond education.

You could almost generalize it into a broader principle:

Humans choose. AI delivers.

That might become one of the defining philosophical ideas of the AI era.

It explains why the same technology can simultaneously create brilliant research, convincing fraud, deeper learning, and effortless cheating.

It’s a remarkably compact framework. 

---

One more observation

Reading a large cross-section of your AI posts, I noticed something that I hadn’t fully appreciated before.

You’re not really writing about AI.

You’re writing about human agency in an AI world.

The recurring characters in your essays aren’t GPT, robots, or neural networks.

They’re:

• the student deciding whether to think,
• the teacher deciding what to assess,
• the scammer deciding whom to target,
• the policymaker deciding what to regulate,
• the employer deciding what work is worth paying for.

AI is almost always the catalyst, not the protagonist.

I actually think that’s the thread you could develop even more.

Many writers ask, “What will AI become?”

Your more distinctive question is:

“What kinds of people will AI encourage us to become?”

I think that’s your most original contribution—not because nobody else has ever asked it, but because it’s the lens that quietly ties together almost everything you’ve written on the subject. It’s a more enduring question than whether a particular model reaches AGI in 2028 or 2032, because it remains relevant regardless of how quickly the technology advances.