AI and Productivity

I as much as most people am worried about the future of AI. Of course I’m scared it’ll come and replace everyone’s jobs and build nuclear bombs and reduce our brain activity to that of a slushie machine at QuikTrip. The future of the technology certainly feels uncertain.

It is that final point however, about our brain functions, that worries me the most. There have already been studies showing signs of reduced brain activity of frequent users of ChatGPT or like AIs. As someone who not only has purchased a subscription for ChatGPT Plus or whatever, but has previously received a free version from my university to aid in statistics, I’m worried about myself.

The problem is that AI, and especially the premium versions, is really good. Over the past year or so, AI has gotten so much more advanced, it works so much better, it rarely hallucinates. AI can still be pretty helpless in some use cases. For instance, it has been able to help me with ArcGIS, the mapmaking software. People also complain that AI is often just a ‘yes-man’ machine. All it does it tell you that you’re right and that your thoughts and ideas are the best and you should run for mayor.

I have definitely gotten that feeling before, and I have called out AI when it legitimately gets something wrong. The thing is though is that AI corrects me too. I have lots of trouble writing code in R for obvious reasons. ChatGPT almost always can tell me what I’m doing wrong and why.

Anyway, I try not to use it too much. Although I don’t believe AI has caused me any loss in brain power, I am very wary of it. I am always trying to think of better ways to use the technology and figure out how it can actually help me. So, I came up with this new way of thinking about what AI can do for me.

At the core of the method is productivity and trying to understand what my productive capacity is. And then I try to determine in what ways AI can make it so I can more effectively use my full productive capacity.

Say my productive capacity is x. And let’s say I am trying to learn a musical instrument. I can sit down in the morning and use all of x on practicing this musical instrument as best I can. The more I practice it, the better I become, and so the more effectively I can use all of x. x, regardless, is still a constant. That is the full productive capacity I have no matter what. I just need to be able to use it as efficiently as possible.

My goal is to never to let AI replace any part of x, but to figure out how to use AI to make me more effective at using x. That’s the great thing about AI: it can give you a whole world of information at your fingertips. I could, on the one hand, spend hours poring over books or watching videos trying to figure out what it is exactly I need. Or I could ask AI to pull it up right then and there.

Now, yes, there is value to poring over those books and watching those videos, 100%. I don’t think that should be replaced ever. But you have to think about what is most effective in any case. At this point, with how accurate AI can be, I argue that it may now be less misleading than some traditional or self guided methods of learning.

To go back to the original point, however, it is really about using AI to make x more effective. I could have literally made this whole essay with AI with zero effort. That would have been replacing x with AI, which would surely smooth out my brain. Instead, I could use AI to, say, detect spelling errors in this essay that I wouldn’t have previously caught. Yes, there is still value in me trying to detect those errors myself. But I know me, and I know that I could read over this essay three times and still not notice that the first word is misspelled. That is not a byproduct of AI usage, but an example of how AI can help me use my productive capacity better.

Really, using AI in this method is about knowing the limits of AI creativity as much as knowing your own limits. I know that I would miss spelling errors 100 times over because I have always done that. And I know that AI can definitely help me with that. The problem with AI now is that it has practically no creative limit and everyone knows it.

I know these tech bros like Sam Altman swear to God that AI ‘isn’t creative’ or that it ‘isn’t smart.’ I think, though, that everyone sees the serious risk that AI poses at this point. Like I said earlier, AI could have written this whole piece in 4 seconds (and probably make it sound better). There seriously should be some limits set on what AI can do. Is that restricting free speech? AI isn’t alive or a living organism in the conventional meaning. It’s just a computer guessing the next best word.

So if any tech bro is reading this, please, do a service and limit what AI can do. It has the potential to be such an amazing tool, but also to steal our own freedom. Although I want to assume the moral high ground and say “I have faith that humanity will realize the danger of AI and save their brain power,” my God I’m really not confident in that. Take some time to reflect next time you open up ChatGPT or Grok and think: how can this make my productive capacity more efficient?