I see a similar pattern in interview coaching. Candidates are getting increasingly good at using AI to structure their answers, but that doesn't necessarily mean they understand their own decisions any better. The moment someone asks a difficult follow-up question, the difference becomes obvious. Tools can improve how you communicate experience, but they can't replace having it.
Agree on the order. What I see less often is someone saying which old skill the AI work actually leans on. Writing clearly seems to do more work than any prompt trick.
It's so easy for a leader to chase the tool and forget the judgment underneath it. I've found AI is useful when it sits on top of experience you already have, not when you pretend the tool replaces that. It can help an expert be more effective and efficient at their craft. Your position comes with real influence and responsibility for what makes sense and what doesn't. Thank you for writing about this, Joel.
The order is the whole point for anyone mid-career. Most people I talk to after a layoff assume they have to start from zero, when the AI skill only pays off on top of the judgment they already spent years building. I see it in my own media and ad tech practice too, where the tools are only as good as the person deciding what to do with them.
I really like this, Joel. Understand what you're doing wrong, understand your strengths, understand your weaknesses, and then use AI on top of that. Most people do it backward, which is why so many AI skills don't stick.
I look at AI as a booster to one's expertise, provided *they have expertise*.
Why would any company start an employee from zero?
Yes, we do need to evaluate how that expertise can be applied in the era of AI, but without it (expertise or domain knowledge), AI skills don't have the same compounding effect on a business.
A person who just knows how to use Claude is not nearly as useful as someone who is an expert in advertising and is just learning AI skills.
The latter can more easily define problems (and use AI for solutions) in their domain.
Best explanation of why you need skills I've read in a while! It's also why skills without lived experience are worth a fraction of those imbued with human expertise.
Same in sales. AI can help a rep work out what the buyer's problem costs if nothing changes, but only if the rep already knows how to find that in discovery and has done it without AI. It's also what I've been trying to teach my kids - how to think critically without AI.
This is so well said and exactly what I’ve experienced in corporate communications. I’ve spent hours and hours in Claude Code, and the expertise and judgment built throughout my career have never been more important while using it. It informs what to build, inputs and outputs, what good looks like, where a human in the loop is critical, etc. Some days it feels like I’ve never used my expertise more than I have with AI as a partner.
AI is a powerful force multiplier, but it only amplifies what is already there... without deep domain expertise and strategic judgment beneath it, even the sharpest tools lack direction.
Joel, the order is the whole thing. I teach high school chemistry and physics, and your list explains why adults with twenty years of experience can bolt AI on and win. The foundation is already under them.
My students do not have that foundation yet. They get the tool before they get the experience, so there is nothing for the AI to stand on. Your closing question has a hard answer for them. They have not built anything yet to put underneath it.
That is where schools have to be careful. The entry-level ladder Rob Lee writes about starts in a classroom. If students skip the climb, they reach the first rung already outsourced.
Syd, that’s such a great point! I have two young kids and it’s the reason I am pushing them to learn, to develop areas of expertise, to write, to play music, to learn basic robotics. Ai needs something to be bolted into.
I do dream of a future where this kind of thing is taught in schools in an AI literacy course of some sort of even workshops led by experts. I’ve done a couple but wish it was more common
Your list is the right list. Writing, music, and robotics all make a kid wrestle with something real before anything can do it for them. My robotics students reached the state championships twice, and every bit of that was foundation work.
That AI literacy course is what I am building. I run Thinking Labs by Temple Academy, a six-week program where students learn to own their reasoning before they extend it with AI. The tool takes a day to learn. The order takes six weeks.
If you ever want a classroom view for a post or a workshop, I would be glad to compare notes.
I see a similar pattern in interview coaching. Candidates are getting increasingly good at using AI to structure their answers, but that doesn't necessarily mean they understand their own decisions any better. The moment someone asks a difficult follow-up question, the difference becomes obvious. Tools can improve how you communicate experience, but they can't replace having it.
Yes! That s a terrific example! I will be using it in the future :)
Love it man.
Every AI agent I run for WriteStack started as a job I'd already done by hand so many times..
Having an agent do it for you is underused these days.
That’s awesome! That’s successful leadership in the age of ai, and so hard to do! Well done
Agree on the order. What I see less often is someone saying which old skill the AI work actually leans on. Writing clearly seems to do more work than any prompt trick.
Good point. I do think strong soft skills will always translate well
This reminds me of teaching a new hire the software before they know what a good invoice looks like.
Yeah! Hey does it notify you that I mentioned your article in the post?
Yes it did. Thanks Joel!
It's so easy for a leader to chase the tool and forget the judgment underneath it. I've found AI is useful when it sits on top of experience you already have, not when you pretend the tool replaces that. It can help an expert be more effective and efficient at their craft. Your position comes with real influence and responsibility for what makes sense and what doesn't. Thank you for writing about this, Joel.
Jason, exactly! Thanks for sharing, too many think expertise no longer matters in this time
The order is the whole point for anyone mid-career. Most people I talk to after a layoff assume they have to start from zero, when the AI skill only pays off on top of the judgment they already spent years building. I see it in my own media and ad tech practice too, where the tools are only as good as the person deciding what to do with them.
Exactly! I like the point on layoffs, we build on what we have, the foundation that has been set.
I’ve found the same. AI gets a lot more useful when I already know the work well enough to spot where it’s gone wrong.
Exactly, Daniel!
I really like this, Joel. Understand what you're doing wrong, understand your strengths, understand your weaknesses, and then use AI on top of that. Most people do it backward, which is why so many AI skills don't stick.
That’s exactly the reason! Both for people and companies
I look at AI as a booster to one's expertise, provided *they have expertise*.
Why would any company start an employee from zero?
Yes, we do need to evaluate how that expertise can be applied in the era of AI, but without it (expertise or domain knowledge), AI skills don't have the same compounding effect on a business.
A person who just knows how to use Claude is not nearly as useful as someone who is an expert in advertising and is just learning AI skills.
The latter can more easily define problems (and use AI for solutions) in their domain.
Exactly! And that last point you made is so crucial for students in the midst of getting education and preparing to go into the workforce
Best explanation of why you need skills I've read in a while! It's also why skills without lived experience are worth a fraction of those imbued with human expertise.
Same in sales. AI can help a rep work out what the buyer's problem costs if nothing changes, but only if the rep already knows how to find that in discovery and has done it without AI. It's also what I've been trying to teach my kids - how to think critically without AI.
This is so well said and exactly what I’ve experienced in corporate communications. I’ve spent hours and hours in Claude Code, and the expertise and judgment built throughout my career have never been more important while using it. It informs what to build, inputs and outputs, what good looks like, where a human in the loop is critical, etc. Some days it feels like I’ve never used my expertise more than I have with AI as a partner.
AI is a powerful force multiplier, but it only amplifies what is already there... without deep domain expertise and strategic judgment beneath it, even the sharpest tools lack direction.
Tools make good people faster, but they can't replace knowing the business.
Joel, the order is the whole thing. I teach high school chemistry and physics, and your list explains why adults with twenty years of experience can bolt AI on and win. The foundation is already under them.
My students do not have that foundation yet. They get the tool before they get the experience, so there is nothing for the AI to stand on. Your closing question has a hard answer for them. They have not built anything yet to put underneath it.
That is where schools have to be careful. The entry-level ladder Rob Lee writes about starts in a classroom. If students skip the climb, they reach the first rung already outsourced.
Build the person first. Then hand them the tool.
Syd, that’s such a great point! I have two young kids and it’s the reason I am pushing them to learn, to develop areas of expertise, to write, to play music, to learn basic robotics. Ai needs something to be bolted into.
I do dream of a future where this kind of thing is taught in schools in an AI literacy course of some sort of even workshops led by experts. I’ve done a couple but wish it was more common
Your list is the right list. Writing, music, and robotics all make a kid wrestle with something real before anything can do it for them. My robotics students reached the state championships twice, and every bit of that was foundation work.
That AI literacy course is what I am building. I run Thinking Labs by Temple Academy, a six-week program where students learn to own their reasoning before they extend it with AI. The tool takes a day to learn. The order takes six weeks.
If you ever want a classroom view for a post or a workshop, I would be glad to compare notes.