Where is my place in a world of AI?

September 2026

You feel the pressure to use AI.

Some of it looks genuinely useful. Some of it is complete rubbish. And the difficult bit is working out which is which. What is true, what is simply wrong, and what is somebody’s agenda dressed up as insight.

But you cannot really ignore it either.

The promises are appealing. Everyone seems to be doing something with it. Clients ask about it. Peers post about it. New tools appear every week claiming that this one changes everything.

So you are sitting in all of that noise trying to remain a serious person who actually knows what they are talking about.

What you actually want

You probably do not need another AI stack or another clever framework.

You want to know what to worry about.

What to ignore.

What you actually need to understand.

And, frankly, what is BS.

If you run an expert business built on trust, that matters.

Your value is judgement. People pay you because they believe you can make sense of a situation, understand the nuance, and separate signal from noise.

AI starts to interfere with that sense of where you fit, even when the tools themselves are useful.

And underneath all of the questions about models, agents and automation is a much more basic one.

Where is my place in the world now?

Chatting to bots at all

I have been talking to LLMs pretty much every day for nearly four years.

The thing that first really got me was not some amazing one-shot answer. It was the conversation afterwards.

The second question. Then the third. The fact it remembered what we were talking about. The reply that somehow felt relevant to me.

For most of my life, conversation meant there was somebody else there.

A person. With their own experience, motives, history and view of the world. Somebody I could trust, disagree with, persuade or decide was talking nonsense.

Now I can have a perfectly fluent conversation with something that is not actually a someone.

And I still find that strange.

I do not think we really have a settled mental model for this yet.

I notice myself giving weight to its judgement. Occasionally trusting it. Sometimes even using it to think through things I would previously have talked through with another person.

And yet it cannot actually be accountable for any of it.

I think pretending that this is completely normal is a mistake. But so is getting carried away and pretending the machine is something it isn’t.

The bot can help me draft. It can research. It can challenge an idea. It can help me think.

But it cannot own the trust somebody places in me, or in one of the businesses I work with.

So I use the conversation.

But I still decide.

And I still own the call.

Bots chatting together

The second layer is where it gets more uncomfortable for me.

I now run what I sometimes describe as a little studio of bots.

They message each other. They pass jobs between themselves. They create side conversations. Quite a lot can happen without me being directly involved.

And some of it is brilliant.

I can leave something running overnight and have useful work waiting in the morning. Research gets done. Drafts appear. Real jobs get moved forward.

But when I describe this to people, I can also see why it worries them.

The science-fiction version of AI taking over was always robots, weapons and some obvious villain.

The version appearing in normal professional work is much less dramatic.

You still have the job.

You still have the title.

You are still accountable.

But bits of the conversation that used to involve you start happening somewhere else.

A draft arrives having already been worked on by several tools. Research appears without the messy debate that would previously have happened between people. Decisions get prepared offstage.

That, to me, is the more interesting version of the AI displacement question.

Not necessarily losing your job.

Losing bits of the process through which your expertise used to be expressed.

If I want the Skynet joke, perhaps Skynet is not lasers. It is side channels.

One bot working with me feels odd.

Bots working without me starts to feel like something else.

And that matters particularly in expert businesses because the client is not really buying a process.

They are buying a reading of the situation.

They are buying judgement.

So if more and more of that reading is being assembled by machines behind the scenes, I need to be very clear about what I understand, what I believe, and what I am personally prepared to stand behind.

Try it. Talk about it.

I do not have a neat framework that makes all of this go away.

What has helped me is much simpler.

I try AI.

And I talk about AI.

A lot.

I try it because I want to experience where it genuinely helps, where it struggles and where it confidently makes things up.

And I talk about it because I want to understand how other serious people are actually using it — rather than how somebody selling an AI product tells me I should be using it.

Over time you start to build your own feel for it.

What to worry about.

What to ignore.

What you need to learn.

And what is BS.

That does not give me certainty.

But it does give me better judgement.

What this looks like for me

I try to use AI on real work, rather than demos.

I give it things that actually matter.

I ask it to draft things I can review (like this!). I let agents work on real problems within sensible boundaries. Sometimes I leave something running overnight and see what it has managed to do by the morning.

Then I talk about it with people I respect.

What worked?

What surprised me?

What did it completely misunderstand?

What would I never let near a client?

And what did it do that, quietly, I would absolutely use again?

Those conversations are probably as important to me as using the tools themselves.

Because if an AI claim cannot survive a conversation with people who actually understand the work, I am probably not that interested in it.

When I look at AI and feel both excited by it and uneasy about it, I do not think that is a contradiction.

Both things can be true.

It can be extraordinarily useful and deeply disruptive.

There can be genuine insight, rubbish and agenda all delivered in the same confident tone.

So I think the underlying question is a reasonable one.

Where is my place in the world now?

For me, the answer is not to wait until somebody on the internet works it out.

It is to keep trying things.

Keep talking to good people.

And keep ownership of what I believe.

What this makes possible

Trying things and talking about them does not remove the unease. It just gives you a better feel for what you are prepared to stand behind.

I wrote about one practical overnight test of that in Personal Assistants Are Starting to Just Work. The broader map of where AI already helps is in What AI Can Actually Do for Your Business Today. If you want a way to choose what to try first, How to Adopt AI Successfully is the risk-visibility version of the same conversation.