I want to discuss something that might sting a little, and I want to say it with care. I have noticed some familiar tendencies in my professional circles, in comment sections, in hallways. Not because people are wrong to have concerns about large language models. But because I recognize the tone, I have heard it before, and I think I know where it leads.
Someone points to a study showing GPT-4 made errors. Someone else shares a screenshot of Claude getting a spreadsheet cell wrong. A third person compiles a thread of hallucinations. And the collective conclusion is relief. See? It’s not that good. We’re still needed.
I understand this impulse. There is something deeply human about wanting to find the boundaries of a new technology, to locate the places where we still matter. This is not stupidity but self-preservation.
I think it is also risky because these impressions are frozen in time, while the technology is racing ahead. It’s probably evolving faster than even the hype you are hearing.
Last Month’s Experience Is Already Outdated
The AI you encountered six months ago is not the AI that exists today. The errors documented in studies from 2025 may have already been addressed. The limitations you experienced in January may have been patched by March.
This is not how most previous technologies worked in my experience. What is happening with large language models is different. These systems are being refined around the clock by some of the best-funded research teams in history, and agents are now fixing, supervising, and evolving each other “24/7”. Building your professional strategy around their imperfections is like building a house on rapidly shifting sand.
What These Conversations Reveal
When I listen to colleagues explaining why AI will not affect their work, I notice something. The conviction often correlates more with personal investment in the status quo than with technical knowledge. And I also share these tendencies. If you have given twenty years to becoming an expert in a particular domain, you have every reason to want that expertise to remain valuable.
I remember when Who Moved My Cheese? by Spencer Johnson made the rounds in corporate circles. It was released during the run-up to the “dot com boom” and became wildly popular, and again during the 2008 economic crisis. A simple parable about mice in a maze, searching for cheese that represented whatever you valued: your job, your security, your identity. The message was equally simple: the cheese moves. It always will. And the creatures who thrive are the ones who go looking for new cheese instead of sitting in the old spot.
I think that book needs to make the rounds again.
A Kinder Way to Say This
I know how hard this is. When I eventually started my own company, it was not only because I had some grand vision. It was also because I could see that the things I had been doing were becoming obsolete, and I needed to find new ground. My background was in PMP (Project Management Professional), steeped in waterfall delivery models, and agile was taking my industry by storm. This looks very much the same; the arguments are the same I heard then, but this time it seems to be moving faster. I never renewed my PMP, and I won’t be renewing my agile certifications either.
What got us here will not move us ahead. I know this sounds harsh. But it is also the kindest thing anyone can tell you. Because the alternative to hearing it now is being blindsided later.
Consider Excel
Think about what happened when spreadsheets arrived. Accountants who had spent decades mastering ledger books faced a choice. They could resist, pointing out that Excel made calculation errors, that it could not replicate the complex judgment of an experienced bookkeeper, and that automation would put good people out of work.
Some of them did exactly that. But the accountants who thrived were the ones who recognized something different: Excel was not replacing their expertise. It was amplifying it. Suddenly, the same professional judgment that once managed one set of books could oversee dozens. The same analytical mind that used to spend hours on arithmetic could spend that time on strategy, on forecasting, on solving problems that were previously too big to tackle.
That is what tools like Claude skills offer now. Not a replacement for your domain knowledge, but a force multiplier. The twenty years you spent becoming an expert in healthcare policy, environmental engineering, or educational assessment? That knowledge becomes more valuable, not less, when you can pair it with AI capabilities that let you work at a scale you never could before.
Solving Bigger Problems
Here is what getting involved actually looks like: learning to orchestrate these plays in your own domain. Building custom tools and skills that combine your hard-won expertise with capabilities that extend your reach. Not programming yourself out of a job, but expanding what your job can accomplish.
The policy analyst who once could research three approaches now explores thirty. The educator who once could personalize instruction for a handful of students now does it for hundreds. The consultant who once delivered recommendations in quarterly reports now provides them in real time. Your expertise is still what matters. You are still the one who knows which questions to ask, which outputs to trust, and which recommendations make sense in context. The AI handles the parts that were always tedious, so you can do more of what you were always good at.
This is not surrender. You are not being replaced by AI. You are being invited to multiply what you can do with it.
The Real Question
I do not find any of this easy. When I see an AI produce something that took me hours in a matter of seconds, part of me wants to find fault with it. To reassure myself that I am still needed.
But I have learned that my ego is not a reliable guide to reality. And I have learned that “Am I still needed?” is the wrong question. The better question is “What bigger problems can I solve now?”
The cheese has moved. It is always moving. The kindest thing we can do for ourselves and each other is to start building in the new location, with better tools than we have ever had before.
