The delegation of critical thinking

Are we headed towards idiocracy

While I wanted to start this substack earlier, you fall into a trap, a trap that what you write has to be good, otherwise why write it, and then nothing gets done. As my friend Churchill said, Perfection is the enemy of progress, and this post is by no means perfect, but its progress. So, in this braindump I call progress, I write to share a disease that is starting to affect us all.

These are dangerous times we are living in, no matter what side of the fence you are, and we have more misinformation out there than we do (accurate)information. Though if you look throughout history, you’ll realize that if you choose any decade, those were dangerous times too. We are always in perpetually dangerous times and we somehow figure a way out of them, but this time around we have a monkey on our back, a gremlin on the airplane wing - A.I - and for all its wonders and gifts it’s given society, it’s also doing something to all of us which is dangerous, it’s taking away our critical thinking.

Be honest with yourself, how many times have you gone to ChatGPT or Claude, searched up a question and taken what it said at face value? You might even think you are clever, and ask “Are you sure?” and rely on the answer it spits out. This is not to say that the LLM’s, especially these days, is not wrong, but it doesn’t change the fact that we are delegating the most important thing to us, critical thinking, to a machine, more than we ever did before.

You might say we’ve been doing that since the advent of Google, but I disagree, with Google you have the option of a different opinion, not an amalgamated opinion of a thousand scraped websites synthesized to give you a “factual” answer.

This also isn’t the same argument as “Calculators” make us bad at math, A calculator is a tool to take my input and produce an deterministic output. Just because I can’t tell you the square root of 90210 from memory, doesn’t mean I don’t understand how to get there by doing it the hard way. It still requires me to do the critical thinking while putting numbers in, and critical thinking processing the output.

With LLM’s, I may have some critical thinking on the input, depending on the task, but I strongly believe a majority of people have no critical thinking on the output and are taking what this algorithm is spitting out as gospel and moving on, which is causing our critical thinking muscle to atrophy, and when that eventually atrophies enough, then we are all President Camacho expecting gatorade to supercharge our plants.

With the explosion of the deep research trick that the LLM’s are playing, it makes things even worse, you now have essays of “deep research” done by LLM’s where no critical thinking is required anymore, you have a so called expert doing research, and you just have to consume.

I agree that there are things that don’t need critical thinking, for example when you get a long email that could just be 2 sentences using Gemini, but that ease of use can create “laziness” and that can permeate through you, and things that do need critical thinking become jobs you end up delegating to ChatGPT.

This also applies to software development, which I know a bit about, I see many developers delegating the critical thinking part of software development to the machine, rather than delegating the boilerplate parts and I foresee a lot of bad times coming for those developers, but this post is about all of us, not just the “prompt engineers”.

So what’s the solution to this? Most humans take the path to least resistance, and being told by Claude what the answer is, is far better than doing the work yourself, but I challenge you to think twice, or three times about why the answer is - the way it is. Do the hard task of building those mental models, before the model leaves you mental.

As a side note, I avoided using LLM’s at all to write this piece, and I will continue to do so, there is no need to pretend to be a great writer, when I just want to share my opinion on things that matter to me.