AI Is Here. I Have to Bring the Rest.

This has been one of the hardest years of my life. I have a long work history, clients I care about who have been happy with what I delivered, and work I am proud of. I am still struggling to make it month to month. I do not know what the future holds.

That is why I cannot afford to pretend AI is temporary, or that it only matters to somebody else. I am not writing this from a position of security. I am learning to adapt because I have to.

That does not mean every AI output deserves to exist, or that people worried about it are stupid. There are real concerns about rights, jobs, energy, surveillance, bad information and the way companies will use any new system to cut costs before they understand the consequences.

But I believe AI is here. Treating it as a temporary annoyance will not make it less present in the work, services and decisions around us. It is becoming a service, and I need to learn how to use it properly.

A search engine did not make everyone a researcher. It put a search box between us and an enormous amount of information. The hard part remained: asking a question worth asking, deciding which sources deserved trust, and knowing what answer should change a real decision.

AI goes further. It does not only retrieve or list information. It can assemble, draft and answer. That makes it more useful, and it makes my standards more important, not less.

AI is a generalist. It can move across subjects, formats and patterns quickly, but it does not begin with the specific knowledge, point of view or framing that makes somebody good at their work. It does not understand who I am: my accumulated reference frame, my experience, judgement, taste, decisions and working history. I have to give it that context.

Until I do, it does not know why a particular client has rejected three apparently sensible options before. It does not know which part of a brand is essential and which part is inherited clutter. It does not know what a previous handover promised, which decision was provisional, where the evidence is thin, or what consequence a bad recommendation will create for the person receiving it.

This is where I think AI slop begins. If I give a generalist system a vague request, I get an answer that could belong to anybody. It may look polished. It may arrive quickly. It may even feel clever because I was there when it appeared. But if the thinking, source material and standard are generic, the result will be generic too. That is not a mystery. It is the expected result.

When I say I need to develop my use of AI, I do not mean training it on my life until it becomes a private expert. I mean getting better at the work around it: keeping the relevant context, asking better questions, stating the boundaries, checking the answer against evidence and knowing when to reject it.

I do not only need to learn how to use a new tool. I need to get better at making sense of what is in front of me: seeing the connection between a client history, a loose brief, a piece of research, a visual reference and a consequence that will not be visible in the first answer. AI can produce fragments quickly. The work is still in deciding what belongs together, what matters, and what should be left out.

My Second Brain matters here because it gives my work somewhere to keep the things a blank prompt cannot hold: the history behind an instruction, the boundary between a draft and an approved claim, the method that worked last time, the exception that must not be repeated, the voice a piece of writing needs to carry, and the reasons a seemingly sensible answer may still be wrong.

It does not make AI think like me. It lets me bring my own material, standards and judgement to the work, then use AI to find connections, recover decisions, challenge an outline, make a rough draft or see what I have missed. The model stays broad. It amplifies what I put into it: my context, my habits, my taste, my blind spots and my level of care. If I bring it nothing but a vague request, it will amplify vagueness. If I bring it a real body of work and keep checking what comes back, it can help me use that body of work more deliberately.

There is a blunt part to this. If I am not willing to learn, change and grow, including becoming better at what I do and clearer about who I am, I will be left behind by people who are. Not because AI is magic. Because they will learn how to use a new service in ways that make their work more useful, more specific and more available.

I could spend the next few years saying AI took the work away. Sometimes it may take work away; sometimes it will change work until it no longer looks familiar. People may be treated badly while that happens. Those are real problems. But resentment is not a plan.

Learning to use AI is not a financial guarantee. It will not solve anyone’s finances, and it cannot replace the need for fair pay, good work or accountability from the people building and buying these systems. I am not adapting because I think it is easy or because I feel certain about the outcome. I am adapting because standing still will not carry me through this.

I cannot expect old proof of good work to explain the new possibilities for me. I have to keep developing my judgement, my skills, my systems and my ability to show how I think. AI is part of that work now.

Use it as a service. Give it a real job. Give it the relevant context. Set a standard it has to meet. Check its answers. Reject the polished nonsense. Keep the responsibility for the result.

Do not ask it to replace who you are. Make it answer to the work you are still becoming capable of doing. Do not ask it to replace who you are. Make it answer to the work you are still becoming capable of doing.