If you’re a coach, therapist, or NLP practitioner, guess what? You’re not just keeping up in the age of AI, you’re actually ahead of the game! Today, we’re diving into the idea that AI, particularly large language models, is just a super-sized version of what NLP has been teaching us for decades. We’ll explore how language shapes behavior, and how the way we communicate can directly influence the outputs we get from AI. So, whether you're looking to enhance your skills or just curious about how these tech tools can work alongside your practice, this episode is all about recognizing the connection between NLP and AI. Grab your favorite snack, kick back, and let’s unravel the fascinating ways AI can help us think better and communicate smarter!

Takeaways:

1. This episode dives into how NLP practitioners are actually ahead of the AI curve.
2. The podcast explores how AI mirrors human thinking, making it a perfect training ground.
3. We discuss the importance of clear prompts, as AI responds better to specific input.
4. The key takeaway is that AI is a reflection of our own clarity and depth of thinking.
5. Understanding language as a programming tool means NLP practitioners are already fluent in AI.
6. The episode invites listeners to notice interactions with AI to improve their communication skills.

Chapters:

1. 00:04 - Understanding the Emotions Behind AI
2. 02:59 - The Cognitive Mirror of AI
3. 04:38 - Exploring NLP Concepts in AI
4. 08:15 - The Shift in AI Understanding
5. 11:39 - Transitioning from Consumer to Creator Mindset in AI Interaction
6. 14:03 - Understanding AI as a Reflective Tool

Companies mentioned in this episode:

1. ChatGPT
2. Claude
3. Gemini

Welcome back to the Deep Dive.

We are doing something a little different today.

Usually when we talk about artificial intelligence, the underlying emotion, let's be honest, is low grade panic.

Or high grade panic.

Yeah.

Or maybe high grade, depending on your Twitter feed.

It is a frantic landscape.

The narrative is very much adapt or get left behind.

Exactly.

It's all I need to learn Python or I have to master prompt engineering like yesterday.

Right.

The algorithms are coming from my job.

But we have a stack of research today that suggests, well, the exact opposite for a specific group of people.

Okay.

If you are a coach, a therapist, a trainer, or specifically a practitioner of nlp, neuro linguistic programming, the argument here is that you haven't fallen behind at all.

You're actually ahead of the curve.

You're ahead of the curve, you just don't know it yet.

It's a fascinating premise.

The source of material we're breaking down is this newsletter article called.

If you're trained in nlp, you're not learning AI, you're recognizing it.

I love that title.

And the core thesis is that large language models, you know, LLMs like ChatGPT, Claude, Gemini, they're all operating on principles that NLP has been teaching for 40 years.

So the mission for this Deep Dive is to really explore that idea that AI is just scary scaled up nlp.

But I want to be careful here because that sounds like a nice bumper sticker.

Does it actually hold water?

Technically, that's the question.

Because on the surface, you know, a therapist asking about feelings in a quiet room and a giant server farm crunching terabytes of data, they feel like very different worlds.

They feel completely different.

Absolutely.

Yeah.

But the source argues the mechanism of processing is identical.

How so?

It's about how language programs behavior, it's about how structure shapes the outcome.

And ultimately it's the fact that we.

What you reinforce gets repeated.

Whether that's a neural pathway in a.

Human brain or a weight in a neural network, it's the same principle.

Okay, so let's set the stage.

If this parallel is so obvious, why isn't everyone talking about it?

Why is my feed just engineers and VCs talking about parameters and compute speed?

Well, you just answered your own question.

Look at who holds the microphone?

The engineers and the VCs.

Exactly.

The engineers built the things, so they ask engineering questions.

What can this do?

How.

How fast can it go?

And then the business people are asking, how do we make money with this?

How do we productize this?

How do we scale it?

Which, to be fair, are necessary questions.

You can't just philosophize if you're burning cash on servers.

Absolutely.

But the source argues they are incomplete questions.

They're structural, not functional.

At least not when it comes to cognition.

Okay.

The author actually draws on their own history, starting in it back in the 80s.

Oh, interesting.

And they saw the same movie before computers entered the workplace and everyone was obsessed with storage capacity.

Megabytes, processing speed.

Right.

The hardware.

The hardware.

Almost nobody was asking, what is this teaching us about how humans process information?

So we're just repeating the 80s, obsessing over the machine and missing the cognitive mirror.

We are missing the mirror entirely.

And the author makes this really sharp point here.

Until the machines that couldn't think started revealing how people actually think, nobody paid much attention.

And now we have AI, which directly mirrors our thought processes.

And we're still treating it like a fancy calculator.

We're asking how fast can it write my email?

Instead of how is it deciding what to write?

That's the pivot point.

Let's get into the mechanics, because for this to be useful, we need to move beyond it's a mirror and talk about how it actually works.

Do it.

The source lists five or six specific NLP concepts that map directly onto AI.

Let's walk through the big ones.

Okay.

The first and arguably the most fundamental is pattern recognition.

In the AI world, that's, that's pretty standard.

The model just predicts the next word based on patterns in its training data.

Right, but in nlp, this is described as what gets repeated becomes the rule.

Human beings are pattern matching machines.

I mean, if you tell yourself a story every day for 10 years, I'm bad at math, I'm not creative.

That story becomes your reality.

It becomes your predictive model.

So when an LLM hallucinates, when it just makes up a fact because it sounds plausible, that's basically the same as a human's cognitive bias.

Precisely.

It's completing a pattern based on probability, not on truth.

It's just doing the next most likely step in the dance.

An NLP practitioner would see that self limiting pattern in a client and intervene to break the loop.

The AI is just running that same loop at a mathematical scale.

Okay, that tracks.

What about the concept of state in nlp?

State management is huge critical.

If I'm angry, I process things differently than if I'm calm.

This is state dependent learning.

In nlp, we teach that context dictates meaning.

You can't separate the words from the state of the person.

Right.

So what's the state for an AI, it doesn't have emotions, it's the prompt.

But specifically, it's the context window.

It's short term memory.

Exactly.

It's everything that's happened in the conversation so far.

If you fill that window with chaotic, conflicting information, the state of the AI becomes confused, the output degrades.

That is a really helpful analogy.

People treat AI like it has this persistent brain.

But if the context window is cluttered, just like a human brain under stress, the IQ drops.

You have to set the state for the AI, just like you'd set the state for a coaching session.

You wouldn't dive into life purpose if someone just ran in screaming about traffic.

Exactly.

You regulate the state first.

And this leads right to the next concept.

Language precision.

Garbage in, garbage out.

We hear that all the time, we say it casually.

But in nlp, it's a technical discipline.

There's something called the meta model.

Okay.

It's a set of tools to challenge vague language.

So if a client says, everyone hates me, the meta model response is everyone, specifically who?

You challenge the generalization to find the.

Data underneath, you challenge the fuzziness.

The source argues AI suffers from the exact same thing.

Give it a fuzzy prompt, you get a fuzzy answer.

That generic vanilla AI corporate speak we.

All hate because it's trying to average out the vagueness.

So prompt engineering isn't really engineering in the coding sense.

It's just being articulate.

It's clean language.

That's the term.

It's stripping away ambiguity.

So the system, human or machine, has a clear target.

Right.

And the last big mechanic is reinforcement.

Which in AI is RLHF.

Reinforcement, learning from human feedback.

That's the thumbs up, thumbs down system.

Yeah, that just sounds like dog training to me.

How is that nlp?

It is behaviorism, yes.

But in nlp, we talk about neural pathways.

Neurons that fire together, wire together every time you run a thought pattern and get a reward.

Even just feeling safe or righteous, that pathway gets thicker faster.

So when AI is learning, it's just deepening the grooves of the pathways.

That got a thumbs up.

Exactly.

It isn't thinking, it's optimizing for reinforcement.

And the takeaway here is really empowering.

If you understand language as a programming language, which is the definition of nlp, you are already fluent.

You don't need Python, you just need precision.

I want to see this in action because the source includes this amazing story about an interaction with Claude that really illustrates this.

The map versus territory moment.

Oh, this is the centerpiece of the article.

It's a great story.

Because it shows the frustration we all feel.

So set it up.

That the author was trying to get Claude to write some marketing copy.

Right.

For a coaching program.

And like we all do when we're rushing, the author was using map language.

Okay, define map language.

For someone who hasn't taken an NLP.

Course, the map is the generalization, the abstraction.

Clients struggle with confidence.

Coaches need to find their niche.

It's the high level summary.

The territory is the reality.

John froze up when asked his price.

That's the territory.

Got it.

The menu versus the meal.

So the author is feeding Claude all this high level map fluff.

Write me something about how hard it is to be a coach.

And usually an AI just plays along.

It churns out generic platitudes in today's fast paced world.

Exactly.

But this time Claude stopped.

It effectively pushed back.

Wait, it pushed back the AI?

It didn't refuse, but it paused the output and asked clarifying questions.

It said, what actually happened?

Give me a specific moment.

Where did you experience this directly?

It was demanding the territory.

It was.

And the author's first reaction was frustration.

Why are you being so picky?

Just write the thing.

Which is exactly how a client reacts when a coach pushes them on a vague statement.

I just want to feel better.

Stop asking me when I felt bad.

100%.

Don't psychoanalyze me.

But then the realization hit.

The AI wasn't broken.

It was being a skilled practitioner.

Wow.

It was doing exactly what the author would have done to a client.

It was saying, I cannot build a model on fluff.

I need data.

And what happened when they complied?

When they gave this specific story?

That's the key unlock.

When the author gave the specific anecdote, the territory.

The AI didn't just write a better email.

It learned the structure of the author's thinking.

So it started mimicking their voice, their nuance, their voice.

It stopped sounding like a robot and started sounding like the author.

This feels like a major shift.

We treat them like search engines.

Type a query, get a result.

This suggests we should treat them like intelligent partners that need context.

We need to treat them like mirrors.

Yeah.

And this brings us to what the source calls the mirror effect.

Okay.

The quality of AI output is directly proportional to the quality of your thinking.

I think this is where people get defensive.

Because if ChatGPT gives me a bad answer, I want to blame the tech.

Oh, it's getting dumber, right?

It feels better to blame the tool.

It protects the ego.

But the source argues no AI is a mirror.

If you Bring lazy, vague, disembodied thinking to it.

It reflects that back as generic fluff.

That's harsh, but it's true.

If you don't give it the territory, it has to make up a map.

And the inverse is also true.

If you bring embodied, specific, structured thinking.

You get nuanced, surprising high value output.

Exactly.

So we have to take responsibility for the input.

We can't just be idea guys throwing vague concepts at the machine.

We have to own the client role.

And this is the practical application part, what the source calls the the client mindset shift.

Explain that.

Because we usually think of ourselves as.

The user, think about how a great coach treats a client versus how we treat AI.

Okay, scenario A.

A clown walks in and says, I feel stuck.

Okay, I would never just say, here are five ways to get unstuck.

That's terrible coaching.

I'd ask, stuck.

How?

Where do you feel it?

When did it start?

Right.

You probe, you investigate, you demand structure.

Now, scenario B.

You open ChatGPT.

You type, write a blog post about feeling stuck.

I hit enter and you hit enter.

You accept the first draft.

You treat it like a vending machine.

Put coin in, get candy out.

And then we complain that the candy tastes synthetic.

Exactly.

We're letting Al make decisions we would never outsource to a client.

Think about brand voice.

People say, I want this to sound witty and professional.

The classic make it sound like Ryan Reynolds.

But witty isn't a setting on a dial.

Witty is a result of a specific worldview.

If you don't provide the structure of that wit, the AI gives you a caricature.

It gives you dad jokes and corporate zest.

The source says, your voice isn't something AI generates.

It is something you decide, and then you train the system to recognize it.

So we need to stop being consumers of AI and start being consumed creators or designers of the interaction.

That's the shift.

The consumer asks, how do I use this faster?

The creator asks, how do I use this?

Well, one makes you a user, the other makes you a master.

And if you're a coach or an NLP practitioner, you already know how to.

Ask the second question.

You've built a career on it.

I love that it takes the anxiety out of it.

It's not a new skill, it's a new application of an old skill.

Exactly.

Now, for the listeners who want to put this into practice, the source provides a specific tool, a diagnostic question.

Yes, this is something can use today.

Before you hit enter on a prompt, or right after you get a response that feels Off.

Just ask yourself this, okay.

If this were a coaching conversation with a client, would I accept what just happened?

That's a high bar.

It is, but let's play it out.

If a client gave you a vague answer about their goals, would you accept it?

No, I dig deeper.

I wouldn't let them off the hook with I just want to be successful.

Then don't accept it from your prompt.

If a client gave you a surface level solution to a deep problem, would you high five them and end the session?

No way.

I'd challenge them to find the root cause.

Then don't accept the surface level bullet points from the AI.

If you wouldn't let a human get away with lazy reasoning, don't let the algorithm get away with it.

Because the AI is just mirroring the depth of your inquiry.

If you probe deeper, the AI goes deeper.

It's interesting.

We talk about AI replacing coaches, but this suggests that AI is actually the ultimate training ground for coaches.

That's a beautiful way to put it.

It's the ultimate dojo for communication.

Think about it.

A human fills in the gaps for you.

If I speak vaguely to you, you use our shared culture to guess what I mean.

The AI can't do that.

It has no intuition.

It only has the text.

So if you're vague, it exposes your vagueness immediately.

If you can explain your concept clearly enough for a machine to replicate it.

Then you truly understand your concept.

You've moved from map to territory.

And if you can't, well, then the mirror reveals the cracks in your own.

Thinking that actually flips the whole narrative.

It's not about how do I survive AI, it's how does AI help me think better?

Correct.

The tech industry narrative is adapt or die.

It's fear based.

The NLP narrative is clarify and empower.

You're not trying to survive the tool, you're trying to shape it.

So we usually end these with a specific takeaway, but the source actually ends with an invitation.

Yes.

The invitation is simply notice.

Just notice.

That sounds almost too simple.

It's deceptive.

The invitation is don't try to fix everything at once.

Just notice the dynamic.

Notice when the AI behaves like a confused client.

Notice when it pushes you toward clarity.

And notice when it just accepts your vague thinking and hands you back fluff.

It's about building awareness of that loop.

As the source says, the noticing is where power lives.

Once you see the pattern, you can't unsee it.

You'll stop seeing a magic black box and start seeing a linguistic structure.

And that is when you stop learning AI and start recognizing it.

And once you recognize it, you can master it.

Exactly.

You realize the call is coming from inside the house.

I have to say, this has completely changed how I'm going to look at that blinking creaser next time.

It's not a blank page.

It's a mirror waiting for a clear reflection.

And it's up to you to provide it.

That is a wrap for this deep dive.

We hope this gives you a little less anxiety and a little more agency in your AI Journey.

As always, thanks for listening.

Take care of.