AI makes novices (not experts) more efficient (than other novices) at doing non-innovative work. Experts are already good at that and see little benefit in using AI.
But AI makes experts (not novices) more effective (than other experts) at true innovation. Novices will have to become experts first to push AI into unfamiliar terrain.
That's why, even for experts, it's worth understanding the capabilities of AI, even though the payoff seems negligible at the start.
💯 thinking has different meanings and AI is forcing us to climb the 🪜 or, if we don’t, we risk to be surpassed by its lower costs and higher speed. IMO nothing new in the technology history 😊:
Earlier this afternoon I was walking with my cousin and this was the exact topic we briefly discussed. The more expert u are in a field, less likely u will use AI because experts have their ego build up on their expertise. When a machine could replicate their expertise, some part them do not want to believe it, hence they resist to use it. Instead of going deep with articulating what they want by prompting it, they just try as minimum as possible and blame AI for not being able to replicate what they actually want. Turns out extracting system out of your brain is much harder than we thought. If only experts can be a little bit open minded, i’m sure they will be the one who lead with AI by leveraging more things out of AI compare to most people.
Totally with you, Wyndo. Another variable here, alongside ego, is that resistance can also be a defense mechanism sometimes out of fear, whether conscious or not. The more your identity is built around “knowing” and being good at doing that stuff yourself, the harder it is to trust the system, and the more you might feel less unique or valuable when the system can prove that it can "replicate you". It’s probably not always the case, but I’m sure time will help skeptical people see there’s even more value in how strategically they apply their knowledge than in the knowledge itself. Hope it will happen sooner rather than later.
This reminds me of the early days when true experts refused to use AI, not always because of identity threats, but because they could often achieve better results faster on their own. In a way, it was the curse of knowledge and a disbelief in AI’s ability to iterate meaningfully.
That said, their expertise is still incredibly valuable. Once they do leverage AI, their sharp thinking and unique taste can reach entirely new levels.
Personally, I feel lucky to be experimenting with AI, it's helped me unlock different angles within myself. I’m sure you and many other power users feel the same. AI still rewards those who are willing to get their hands dirty and bring deep domain expertise to the table.
The effort you need to put in at the beginning is the hardest step. I see this as it is with products with poor onboarding: if it’s too complex before you experience the value, you give up. But there’s no shortcut for this, nor should there be, because skipping the hard part would diminish the actual value of the result.
I think this issue is even deeper for knowledge workers and employees (that’s just my observation), but it seems freelancers and entrepreneurs are often more motivated to use new technology because they want to move faster, get more clients, or sell their expertise. Professionals in companies, on the other hand, don’t always have the same drive, and most organizations still lack proper training or processes to support AI adoption.
But yes, experts have the biggest advantage, and not just to improve their own work, but also to lead the way for others. As AI continues to democratize access and help people achieve better results without deep expertise, I think the pressure on experts to adapt will only increase, as they're going to lose a lot of ground if they don't.
Thanks for sharing this, Daria, it really resonated.
Working with more people outside my day job made me realize how fast things move when you’re out of the corporate bubble. It’s not even 10x, it’s just a totally different mindset. The pressure to move fast and cut costs drives people to adopt and iterate way more aggressively.
And yeah, experts often default to judging what AI can’t do, instead of asking how to make use of what it can. That mindset can really hold people back.
I also see the pressure to adapt is only going to grow. Some folks will be forced into it, and the skeptics might get hit the hardest. But I do think most people are more adaptable than they give themselves credit for, especially when the stakes are real.
Thanks for posting this article! I've been instinctively doing these things without guidance from other people, creating AI agents through vibe coding for my recurring tasks and helpful prompts. It's good to know I'm heading in the right direction!
This reminds me of a quote by Richard Feynman: "If you can't explain it simply, you don't understand it well enough."
In product management we often think we know something until we try to teach it to someone else (or in this case, an AI). The process of creating these systematic prompts and workflows forces us to articulate the expertise we carry around. It's like reverse engineering our own brain, which makes both us and the AI better at the work.
I really like the metaphor you used, that building these systems is like reverse engineering our own brains. What’s fascinating is that as we break down our thinking for AI we also become more self-aware of how we think, our capabilities, strengths, ways of doing things etc.
Amazing! The process you outline is very reminiscent of the Authentic AI process I created for my audience -- preloading a custom GPT with your expertise and nuances. Though in my case it has more to do with writing and writer values, thoughts, and opinions.
I guess using genAI effectively all really boils down to: how you can shape AI to your own hyper-specific needs.
Giving up on the tool because it doesn't yet fit you is like giving up on walking because you can't find the right pair of shoes. It's a process of optimization, and the more you dig into it, the more beneficts you get. This is an excellent break down of the mentality.
Thank you, James! this process really is universal and can be applied across any industry or context.
And yes, the hard part has nothing to with AI actually, it’s actually sitting down to assess your own abilities and figuring out the best way to transmit them. Once you do that, AI becomes a much more effective tool.
Daria, bravo. This is precisely what the AI conversation needs more of...systems thinking instead of magic thinking. Too many "experts" are treating AI like it will just "do the work for you." You've laid out the reality: AI amplifies your process, but you still need a process worth amplifying.
The examples and citations here should be required reading for anyone who thinks prompt engineering is the same as strategic thinking. Well done.
What I notice with experts over and over again is that when they try using AI for something they know deeply and see it underdeliver, they quickly conclude it’s not good enough. But in reality, the missing piece is that we haven’t taught the AI how to think like us yet.
Secretely hoping this brings some light and changes some skeptical minds.
I feel this way too. Up until about 3 days ago I realized one of my biggest oversights with AI was pushing it to do several tasks at once. In cases like that it change my words, inferred meaning I didn’t pass across and made me frustrated all at the same time.
Big takeaway here for me: many know context is king with AI, but clearly defining what “good” looks like inside our knowledge packets or context banks is what actually makes it work.
Or work well anyway.
You explained that so well and backed it up with real examples. Nicely done.
Expertise indeed intensifies the challenge of collaborating with AI, not because it makes tasks simpler, but because it clearly reveals where AI lacks subtlety and nuance. This process align with what I call AR (Amplified Reflection), using AI to amplify human judgment, clarity, and refined taste. Your structured framework for embedding expertise into AI is precisely the kind of intentional approach I advocate with Coherence Engineering, which ensures human authenticity and coherence remain central to AI collaboration.
"When experts choose to engage in this more demanding form of AI collaboration, we're doing something larger than improving our individual productivity. We're creating a new form of intelligence, not artificial intelligence replacing human intelligence, but hybrid intelligence that combines the best of both." This is still a new technology and if given some time (mid 2026) and leeway it will stop churning out the occasional miss fires and become your best tool in the garage.
I agree, I think the same & it’s not just about waiting for the technology to improve, but also about developing the skills and mindsets to work alongside it effectively
This analogy is spot on. Having AI but not learning how to use it well really is like owning a car and walking everywhere. I’ve also realized that the more I understand a topic, the more effectively I can partner with AI not just accept answers, but refine and question them. Expertise doesn’t make AI unnecessary; it makes it more powerful. This really shifted how I see the tool. Thank you for this insight.
I know what you're saying, I actually sat with these same thoughts and frustrations for a long time before I realized what was really going on. Glad it resonated!
AI makes novices (not experts) more efficient (than other novices) at doing non-innovative work. Experts are already good at that and see little benefit in using AI.
But AI makes experts (not novices) more effective (than other experts) at true innovation. Novices will have to become experts first to push AI into unfamiliar terrain.
That's why, even for experts, it's worth understanding the capabilities of AI, even though the payoff seems negligible at the start.
So well put. Restacking this.
And indeed, it can feel counterintuitive for experts to invest time in learning a tool that initially seems to offer little advantage.
💯 thinking has different meanings and AI is forcing us to climb the 🪜 or, if we don’t, we risk to be surpassed by its lower costs and higher speed. IMO nothing new in the technology history 😊:
https://antoninorau.substack.com/p/from-eikasia-to-noesis-what-plato
Earlier this afternoon I was walking with my cousin and this was the exact topic we briefly discussed. The more expert u are in a field, less likely u will use AI because experts have their ego build up on their expertise. When a machine could replicate their expertise, some part them do not want to believe it, hence they resist to use it. Instead of going deep with articulating what they want by prompting it, they just try as minimum as possible and blame AI for not being able to replicate what they actually want. Turns out extracting system out of your brain is much harder than we thought. If only experts can be a little bit open minded, i’m sure they will be the one who lead with AI by leveraging more things out of AI compare to most people.
Totally with you, Wyndo. Another variable here, alongside ego, is that resistance can also be a defense mechanism sometimes out of fear, whether conscious or not. The more your identity is built around “knowing” and being good at doing that stuff yourself, the harder it is to trust the system, and the more you might feel less unique or valuable when the system can prove that it can "replicate you". It’s probably not always the case, but I’m sure time will help skeptical people see there’s even more value in how strategically they apply their knowledge than in the knowledge itself. Hope it will happen sooner rather than later.
This reminds me of the early days when true experts refused to use AI, not always because of identity threats, but because they could often achieve better results faster on their own. In a way, it was the curse of knowledge and a disbelief in AI’s ability to iterate meaningfully.
That said, their expertise is still incredibly valuable. Once they do leverage AI, their sharp thinking and unique taste can reach entirely new levels.
Personally, I feel lucky to be experimenting with AI, it's helped me unlock different angles within myself. I’m sure you and many other power users feel the same. AI still rewards those who are willing to get their hands dirty and bring deep domain expertise to the table.
The effort you need to put in at the beginning is the hardest step. I see this as it is with products with poor onboarding: if it’s too complex before you experience the value, you give up. But there’s no shortcut for this, nor should there be, because skipping the hard part would diminish the actual value of the result.
I think this issue is even deeper for knowledge workers and employees (that’s just my observation), but it seems freelancers and entrepreneurs are often more motivated to use new technology because they want to move faster, get more clients, or sell their expertise. Professionals in companies, on the other hand, don’t always have the same drive, and most organizations still lack proper training or processes to support AI adoption.
But yes, experts have the biggest advantage, and not just to improve their own work, but also to lead the way for others. As AI continues to democratize access and help people achieve better results without deep expertise, I think the pressure on experts to adapt will only increase, as they're going to lose a lot of ground if they don't.
Thanks for sharing this, Daria, it really resonated.
Working with more people outside my day job made me realize how fast things move when you’re out of the corporate bubble. It’s not even 10x, it’s just a totally different mindset. The pressure to move fast and cut costs drives people to adopt and iterate way more aggressively.
And yeah, experts often default to judging what AI can’t do, instead of asking how to make use of what it can. That mindset can really hold people back.
I also see the pressure to adapt is only going to grow. Some folks will be forced into it, and the skeptics might get hit the hardest. But I do think most people are more adaptable than they give themselves credit for, especially when the stakes are real.
Thanks for posting this article! I've been instinctively doing these things without guidance from other people, creating AI agents through vibe coding for my recurring tasks and helpful prompts. It's good to know I'm heading in the right direction!
This reminds me of a quote by Richard Feynman: "If you can't explain it simply, you don't understand it well enough."
In product management we often think we know something until we try to teach it to someone else (or in this case, an AI). The process of creating these systematic prompts and workflows forces us to articulate the expertise we carry around. It's like reverse engineering our own brain, which makes both us and the AI better at the work.
Thank you, Dutch!
I really like the metaphor you used, that building these systems is like reverse engineering our own brains. What’s fascinating is that as we break down our thinking for AI we also become more self-aware of how we think, our capabilities, strengths, ways of doing things etc.
And on Feynman’s philosophy, I actually created a prompt inspired by it for accelerating learning: https://aiblewmymind.substack.com/p/the-one-thing-that-changed-how-i
Could be a greaaat way to teach AI your expertise! It didn't cross my mind before you said it.
Amazing! The process you outline is very reminiscent of the Authentic AI process I created for my audience -- preloading a custom GPT with your expertise and nuances. Though in my case it has more to do with writing and writer values, thoughts, and opinions.
I guess using genAI effectively all really boils down to: how you can shape AI to your own hyper-specific needs.
Giving up on the tool because it doesn't yet fit you is like giving up on walking because you can't find the right pair of shoes. It's a process of optimization, and the more you dig into it, the more beneficts you get. This is an excellent break down of the mentality.
Thank you, James! this process really is universal and can be applied across any industry or context.
And yes, the hard part has nothing to with AI actually, it’s actually sitting down to assess your own abilities and figuring out the best way to transmit them. Once you do that, AI becomes a much more effective tool.
Yeah. Across all tools, it’s always the thinking that’s most difficult.
Golf clubs make it easier to hit a golf ball, but, as golfers know, that's only the beginning!
Daria, bravo. This is precisely what the AI conversation needs more of...systems thinking instead of magic thinking. Too many "experts" are treating AI like it will just "do the work for you." You've laid out the reality: AI amplifies your process, but you still need a process worth amplifying.
The examples and citations here should be required reading for anyone who thinks prompt engineering is the same as strategic thinking. Well done.
Thank you, Rich, I really appreciate your words.
What I notice with experts over and over again is that when they try using AI for something they know deeply and see it underdeliver, they quickly conclude it’s not good enough. But in reality, the missing piece is that we haven’t taught the AI how to think like us yet.
Secretely hoping this brings some light and changes some skeptical minds.
I’m going to send you a DM, I’ve an idea
Send it over! my internal archives are never too full of ideas, haha
Sent
I feel this way too. Up until about 3 days ago I realized one of my biggest oversights with AI was pushing it to do several tasks at once. In cases like that it change my words, inferred meaning I didn’t pass across and made me frustrated all at the same time.
Interesting - in different conversations?
Some conversations are smooth, especially if it’s a single task but the frustrating ones are ones that have to do with my write ups.
When it’s refining your writing?
Big takeaway here for me: many know context is king with AI, but clearly defining what “good” looks like inside our knowledge packets or context banks is what actually makes it work.
Or work well anyway.
You explained that so well and backed it up with real examples. Nicely done.
Indeed, context is king. Thank you, Tam!
Good insight 😃. Can i translate part of this article into Spanish with links to you and a description of your newsletter?
Thank you, David! Yes, for sure. Appreciate you helping to share these ideas with a wider audience!
Many thanks, Daria.
Expertise indeed intensifies the challenge of collaborating with AI, not because it makes tasks simpler, but because it clearly reveals where AI lacks subtlety and nuance. This process align with what I call AR (Amplified Reflection), using AI to amplify human judgment, clarity, and refined taste. Your structured framework for embedding expertise into AI is precisely the kind of intentional approach I advocate with Coherence Engineering, which ensures human authenticity and coherence remain central to AI collaboration.
"When experts choose to engage in this more demanding form of AI collaboration, we're doing something larger than improving our individual productivity. We're creating a new form of intelligence, not artificial intelligence replacing human intelligence, but hybrid intelligence that combines the best of both." This is still a new technology and if given some time (mid 2026) and leeway it will stop churning out the occasional miss fires and become your best tool in the garage.
I agree, I think the same & it’s not just about waiting for the technology to improve, but also about developing the skills and mindsets to work alongside it effectively
This analogy is spot on. Having AI but not learning how to use it well really is like owning a car and walking everywhere. I’ve also realized that the more I understand a topic, the more effectively I can partner with AI not just accept answers, but refine and question them. Expertise doesn’t make AI unnecessary; it makes it more powerful. This really shifted how I see the tool. Thank you for this insight.
Thank you, Sae! It’s great to hear your own experience and how your own understanding has changed the way you work with these tools.
Interesting take. Had not thought of this before.
I know what you're saying, I actually sat with these same thoughts and frustrations for a long time before I realized what was really going on. Glad it resonated!