loading . . . The Five Reallys **Editor’s note:** I had promised/threatened that the next post on _alt.management_ would go out to Pro subscribers only. However, as I was working on that post, I realized there’s an idea to share _before_ that I’ll build on. And it’s one that I’d like to be able to publicly link to. So you’re in luck. Here’s another free one for you.
* * *
The term critical thinking has gotten a bad rap. People think (uncritically) that to be _critical_ means to be a nay-sayer, to be cynical, to stand in the way of progress.
While this could be the impression, it’s not the goal. The goal of critical thinking is to “analyzing available facts, evidence, observations, and arguments to make sound conclusions or informed choices.” But indeed, it is often the case that things are not as clear-cut as they may seem when evaluated critically.
That still sounds uncontroversial, though. I mean, who would be critical of doing _that_?
Yet worryingly, critical thinking seems to be in decline, while it is a skill more important than ever.
A few reasons:
Many of us now view the world through algorithmic lenses (Facebook, Instagram, TikTok, Youtube, ChatGPT) whose vendors have an interest in keeping us engaged. They have figured out that showing us content that _agrees with us_ will achieve this goal. We prefer warm baths over cold showers — whether we need the cold shower or not.
This is very prevalent in LLM chatbot’s tendency to fervently agree with whatever batshit crazy idea we throw at it: “oh yes, glue on pizza, clever way to keep the toppings on!” If we assume this is just a temporarily glitch that will soon be resolved, we should critically ask ourselves: why would a vendor be interested in changing this if it could lead to you getting upset, and using it less? Case in point: with the backlash around GPT-5’s change in tone, OpenAI immedidiately dialed up the “nice” even more. Warm baths sell, cold showers do not.
As a result, we are challenged less and less and at risk of getting in the habit of just accepting what we’re being told, to become complacent.
This is problematic in general, but what is even more worrying is that a lot of our jobs are evolving/devolving into ones assessing a lot of LLM output: AI books on what mushrooms you can eat, AI agent generated code, suggestions on what air base to strike — that sort of thing. Everybody ostensibly knows, and even vendors tell you this explicitly: you need to check for accuracy, you may be reading PhD-level sounding nonsense. In other words: critical thinking required!
The need is clear — especially in the age of AI, we need to **make critical thinking great again** (MCTGA)!
So we are going on a mission. How should we go about it?
I dunno. I’m an engineer. This is more of a marketing thing.
Let’s turn to our marketing department.
“We usually solve problems with a rebrand. Try to change perception from something _negative_ to something _fun_!”
Gotcha!
(Re)introducing “The Five Reallys”
* * *
One of various cool things to come out of Toyota — beside Kanban, and I hear they also make cars these days — is a root-cause analysis framework called The Five Whys.
The idea is simple: you start with a problem (for instance a production incident of some sort) and ask “Why did this happen?” You answer it, and then ask “Why?” again. Every time, you dig one level deeper and get closer to the _root cause_ , until there are no more sensible answers to give. Anecdotally this usually happens within five steps, hence the _five_ in there. It’s simple yet powerful technique.
In 2017, I did the rounds with a talk entitled The 100x Engineer. Now I should remind you: these were different times. This was a time where today’s _agentic engineers_ had hardly even heard of the blockchain yet. My pitched path towards 100x (I’ve always liked hyperbole) productivity, revolved around being smart about challenging scope and clever and sustainable implementation strategies. What can I say — I was young and naive.
However, buried in that talk sat something that stood the test of time, a play on the _Five Whys_ , a call to critical thinking: _The Five Reallys_.
* * *
Some things are best explained by example. Allow me to demonstrate with something random. Let me find something real quick. Ah, here we go.
“AI could wipe out entire job categories.”
_Really?_
“Yes, Sam Altman said so.”
_Really,_ who’s Sam Altman?
“He’s the CEO of OpenAI, the world’s most prominent AI vendor, doesn’t ring a bell?”
_Really?_ Why would he make a statement like that, does he have any interest in boosting the importance of AI, self-fulling-prophecy style?
“I suppose so. But I think he’s just trying to warn us about what’s coming. He’s concerned about us. He’s quite _alt_ ruistic.”
_Really?_ So that is what the _Alt_ in his last name represents.
“Oh yes, Sam has even said that AI can end global poverty.”
_Really?_ Quite the supporter of the poor! Was this before or after he stepped into his Koenigsegg Regera?
“Before I guess, otherwise we would not have heard him over the noise of exploding-dinosaur engines.”
* * *
Ok, that didn’t go where it should have. Let’s try again in a more practical day-to-day scenario: you’re doing planning with your team and your product manager presents the roadmap.
“We need to have all these 10 features.”
_Really?_
“Yes.”
_Really?_
“Well, ok, maybe not feature 7. But for the rest: yes.”
_Really?_
“Yes, well, I think so... Let me verify with the customer.”
“Ok, they may need 4 and 5, but we can do this later. The rest are all must-haves.”
_Really?_
“Yes.”
Better! We managed to significantly cut our initial scope, and with one _Really?_ to spare!
* * *
Here is the recipe, in case you hadn’t reverse engineered it yet:
You read something; somebody makes a statement. Rather than just nodding along, accepting it as is, and repeating it as fact, you take a step back.
Hang on — _really?_
This pivots you in critical thinking mode. Except we won’t call it that, because marketing.
You start to ask follow-up questions: _Really_ , who is making this statement? _Really_ , what are their interests? _Really_ , what do they know or not know?
Now, one of two things may happen.
Either you dig in and everything checks out. The statement was 100% on the mark. Great!
Or, there’s something to uncover. Perhaps there’s more nuance that changes things. Perhaps the whole statement just turns out to be bullshit.
* * *
We have to (re)train this ability. The ability to detect questionable things. One could call this a _bullshit detector_ , but the market department takes issue, we can’t be seen using language like that.
So let’s just refer to it as developing our spidey sense, which in its etymology _technically_ is just about detecting risk or danger, but that’s close enough. People still like Spiderman, right? Or has he been canceled?
Now, there is a risk (and to some of us this comes naturally) that we become so awesome at this that we start to overdo it. There’s a risk to become the person in the room that constantly raises their hand asking “really?” Frankly, that’s just annoying and also potentially wasteful. There’s opportunity cost to everything. Part of the skill has to be to know when to shut up.
However, I would consider this a luxury problem. In the age of AI, we have to keep our shield up constantly, consistently, few exceptions. What makes this incredibly hard is that previously, bullshit was easy to detect because it tended to be riddled with spelling errors.
Today, potential bullshit sounds increasingly “PhD level” — using em-dashes and such. You know from experience reading my stuff (which is definitionally all at PhD level), that this is _really_ hard to challenge. When you read things I write or say, do you constantly ask yourself “really”? In my particular case there is no need, because what I say is absolutely correct. However, in the general case: _yes,_ you really should. And whereas _I_ can be trusted speaking authoritatively about stuff I have no degree in, ChatGPT cannot. LLMs cannot.
Are humans equipped to do that? I mean, for real? Every single time? I fell into this trap myself many times. I find myself thinking “well that sounds true enough, no need to check that.” And I’d be wrong. I have been retraining myself to sound alarm bells whenever I think something “sounds good” or “looks good.” Yes, this is tiring.
As much as I love you all, and I hate to tell you this: I have little belief you are naturally better at this than me.
But if this is our future, we have to overcome.
* * *
We have to keep stepping back.
To keep thinking: _hang on_...
_Really?_ (5x)
And call out the... flapdoodle.
MCTGA! https://alt.management/five-reallys/
