This is a comic strip I scripted. The visuals are AI generated (ChatGPT).

You have to direct AI, there's no question about it. Think of AI like a wandering mind with no direction unless you give it direction. You must have a clear goal in mind and you can use AI to help you plan towards a goal. Once you've reached a decent plan, stop. Stop the planning phase. Copy your plan to a word file, print it out, and stop the planning phase and start taking action. AI is pretty bad about ruminating and repeating things over and over. Once you've reached a peak in your discussion, end the discussion and take action. This may mean not using AI for an extended period of time in order to consistently take action on your plans. It could mean a few days, a few weeks, or even months before you check back in with AI about your progress. In my case, I'm working on some things and have a plan of action. I discussed a particular business strategy with AI. Now I will probably go a few days or maybe a week or two before I revisit. You can also let AI play the role of a coach after you've taken some action steps. So, let's say you're in unfamiliar territory. You came up with a plan using AI assistance. You took action. You sent the email and you made the follow-up call. You visited that client. You got a response and have an idea of what to do next but want a second opinion on your plan of action. YOU decide if the feedback from the AI model is helpful. Maybe you take some of what it says and ignore other things. Maybe you take all of it and add it to your next step. Maybe you ignore the feedback altogether. Or maybe what the AI model suggests is something you decide to do much later. You're in control. Let AI be an assistant for planning. But just like thinking, planning using AI will only give you a map. You have to be the one to walk out of the door, start the engine in the car, and start driving to your destination. Think of AI more like a guidance system and less like some kind of magical geanie and you'll gain a lot more from it.

True skill in any endeavor doesn't lie in trying to master too many things at the same time. Start with one, two, maybe three skills at the most and practice them over and over and over again until you become really good at them. This applies to AI or anything really. Once you've become good at the first 1-3 skills, add only one more skill at a time. Taking as much time as is needed to learn it while also using the others. Eventually what you will find is overlap. You will also inadvertently learn other small additional skills without realizing it. In learning three or four skills, you will probably unknowingly pick up an additional two or three. The only way to get good at anything is by practice and repetition. Reading about it will only give you introductory information. True skill is gained through action.

If you're not careful AI can create a disadvantage for you if misused. If you have a tendency to overanalyze without taking action, AI can amplify your tendency to overanalyze some more and to procrastinate indefinitely. So the first thing you have to do is set a limit on the amount of analysis you do and then take action. There's a place for planning and evaluating the steps you will take but at some point it will be time to create a plan of action and get going. Planning without execution is just a glorified form of daydreaming.

This is an example of an AI workflow based on a project I'm working on. I asked ChatGPT to make an image of a measuring device measuring a part. The device looked good but the part was positioned wrong even after several accurate and explicitly descriptive attempts. The part, being positioned wrong, was not in the correct position to be measured and also obstructed part of the view of the measuring device in the illustration. The illustration was photo-realistic. To correct the problem, I had to finish the job in Microsoft Paint. I had a good base that the AI model created for me but since it couldn't correct the problematic elements, I had to make the corrections myself. I had to flip part of the image and also erase parts. I then had to draw a small section in and make it look photo-realistic to blend in with the picture. It turned out good and was a bit time consuming. It took me 2 hours to do, but AI did save me time. It would have taken much longer to create a photo-realistic drawing from scratch. Probably 8 to 10 hours at least. This is just to give a practical example of what a project you are working on may look like when implementing AI. It's not a magic bullet. It has its strengths and weaknesses but can save you time. Your tasks will be done by you, doing much of the legwork, and AI will be able to assist you in certain areas. It may also enable you to do something you wouldn't normally be able to do on your own. In this example, I wouldn't be able to create a photo-realistic image from scratch but it did. I then was able to use that as a base to work off of. To create photo-realistic colors that blended in with the image, I was able to select colors from the image and airbrush them in layer by layer. It started out as looking like solid colors but as I picked different hues of the same colors and lightly airbrushed them in, the finished drawn-in section looked like a photo of something with a textured surface. The image created by AI not only gave me the object to work with, but the color pallet I could use to blend in sections that were drawn in by hand.