One of the most underappreciated effects of capable AI is what it does for very small teams and solo operators. A single person can now produce research, drafts, analysis, and materials that not long ago would have required a small staff. The leverage is real, and it is genuinely empowering. But I think it sets a trap that many will walk straight into, and the way out of that trap is a fundamental rethink of what you are actually selling.
The trap is this. When AI lets you produce far more output far faster, the instinct is to compete on output — to do more, cheaper, quicker. But if AI gives that same superpower to everyone, then output itself becomes abundant and therefore cheap. Competing on volume in a world where volume is nearly free is a race to the bottom that you cannot win, because there is always someone willing to produce the same commodity output for less. The very capability that empowers you also commoditises what you produce.
The way out is to stop selling the output and start selling the judgment behind it. AI can generate a draft, but it cannot reliably know which draft is worth producing, which problem is worth solving, or whether the result is actually right for the specific situation in front of a real client with real constraints. That discernment — knowing what matters, what to ignore, what is good enough, and what is dangerously wrong — is the thing that does not get commoditised, because it is precisely the thing AI does not supply on its own.
I think this is the central career and business lesson of the moment. The value is migrating away from execution, which AI increasingly handles, and toward judgment, which it does not. The keystroke is becoming worthless. The decision about which keystrokes are worth making, and whether the result can be trusted, is becoming more valuable. If you are charging for the former, your prices will fall. If you are charging for the latter, you have something defensible.
What this looks like in practice is a shift in how you position what you do. Instead of "I will produce this for you," it becomes "I will make sure the right thing gets produced, that it is correct, and that it fits your situation." The AI does the heavy lifting of generation; you provide the taste, the accountability, and the contextual wisdom that turns raw output into something worth paying for. The client is not paying for the words. They are paying for the confidence that the words are the right ones.
There is a discipline required here that I think trips people up. The temptation, once you have AI leverage, is to act like a miniature version of a large operation — to chase scale, to take on everything, to compete on breadth. I think that is a mistake for small operators. The advantage of being small is not that you can now do what large organisations do; it is that you can apply sharp, specific judgment to problems where that judgment is scarce, at a speed and cost large organisations cannot match. Default to the lightest possible tooling, build one repeatable thing that works, and charge for the judgment, not the labour.
This also has implications for how we should think about education and skill. If execution is being automated, then training people primarily to execute is preparing them for the part of the work that is disappearing. The durable skills are the ones AI does not replace: discernment, contextual reasoning, the ability to know when something is wrong, the responsibility for an outcome. Those have always been the harder things to teach, and they are now the more valuable ones to have.
So the question I would put to anyone building a small business or a career in this environment is direct: if a capable AI can now produce what you produce, what exactly is the client paying you for — and have you made that the centre of what you sell, or are you still quietly competing on the one thing that just became free?
The above reflects my personal views only and is intended for informational and discussion purposes. It does not represent the position of any employer or organisation.