Self-Aware Confabulation
All men are frauds. The only difference between them is that some admit it. I myself deny it. ― H. L. Mencken I think where I am not, therefore I am where I do not think. I am not whenever I am the plaything of my thought; I think of what I am where I do not think to think. ― Jacques Lacan Conscience is the inner voice that warns us somebody may be looking. ― H. L. Mencken, again The Elephant in the Brain by Robin Hanson and Kevin Simler was the piece that first introduced me to the idea. I often felt like the Elephant's takes are overly cynical, and the same goes for other pieces of Hanson's writing. That is, before I read Edward Teach's Sadly, Porn that is outright misanthropic, and still feels pretty accurate whenever I can make any sense of it. The core thesis in both of these books is
All men are frauds. The only difference between them is that some admit it. I myself deny it.
― H. L. Mencken
I think where I am not, therefore I am where I do not think. I am not whenever I am the plaything of my thought; I think of what I am where I do not think to think.
― Jacques Lacan
Conscience is the inner voice that warns us somebody may be looking.
― H. L. Mencken, again
The Elephant in the Brain by Robin Hanson and Kevin Simler was the piece that first introduced me to the idea. I often felt like the Elephant's takes are overly cynical, and the same goes for other pieces of Hanson's writing. That is, before I read Edward Teach's Sadly, Porn that is outright misanthropic, and still feels pretty accurate whenever I can make any sense of it.
The core thesis in both of these books is somewhat similar. The Elephant in the Brain says that there's a unconscious part, the Elephant, that does self-interested stuff like status-seeking. To be able to present a prosocial personality, we then have a separate layer that interprets the actions of the Elephant in a good light. It's hard to call this lying because to be believable we have to believe it ourselves first. And our brains are great at pattern matching and forgetting conflicting details.
Sadly, Porn goes the other way, or perhaps just further. You've domesticated the Elephant. It no longer dares to do the self-interested actions. It's afraid of failure. The internal narrator is repurposed from defending our selfish actions to others, into explaining our lack of actions to ourselves. We're lying to ourselves, trying to uphold our own story of having a high status. Other people are mostly required for external approval.
Reading these books was like partially breaking the 4th wall of the narrator. It became self-aware. Of course I could be just imagining a minor enlightenment instead of experiencing it. That would be such a Sadly, Porn-style mental move. Perhaps we could test this by consciously changing the actions of the Elephant? How would one interpret the results instead of retreating to another abstraction level with the lies? It seems really hard to point at something you've done and say "the Elephant did that".
Both of these models started to look a bit lacking after actually internalizing them. For a long time, I was having a really hard time identifying any motivating factors besides physical needs, hedonism, and status-seeking, thinking that anyone doing other things was lying to either themselves or others, or both. I still somewhat hold these views; I just don't think that the lying part is so absolute. People also have aesthetic preferences (read: values) that do not have obvious self-interested purpose.
But as the saying goes, "all models are wrong, some are useful". I've found both of these quite useful in modelling how others behave. And how I behave, too, albeit disregarding the narrator's explanations is tedious and squeamish work, as convenient answers look very appealing. And all this self-reflection seems to be mostly for entertainment anyway, as for actual results I use more powerful tools.
Sign in to highlight and annotate this article

Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
model
Why Your AI Agent Keeps Getting It Wrong: The Three-Layer Architecture Every Data Leader Needs to…
Why Your AI Agent Keeps Getting It Wrong: The Three-Layer Architecture Every Data Leader Needs to Know Your AI agent is not failing because the model is bad. It is failing because the architecture feeding the model is incomplete. The agent does not know what your “revenue” number means. It cannot see the CRM data it needs. It does not know that this question should be answered by the finance persona, not the sales one. The model is doing its job. The infrastructure around it is not. This is the defining challenge of enterprise AI in 2026. Everyone has deployed agents. Most of those agents produce responses that are confidently wrong, inconsistently right, or too generic to act on. The gap between a demo that impresses and an agent that actually drives business outcomes comes down to three
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Products

Why Your AI Agent Keeps Getting It Wrong: The Three-Layer Architecture Every Data Leader Needs to…
Why Your AI Agent Keeps Getting It Wrong: The Three-Layer Architecture Every Data Leader Needs to Know Your AI agent is not failing because the model is bad. It is failing because the architecture feeding the model is incomplete. The agent does not know what your “revenue” number means. It cannot see the CRM data it needs. It does not know that this question should be answered by the finance persona, not the sales one. The model is doing its job. The infrastructure around it is not. This is the defining challenge of enterprise AI in 2026. Everyone has deployed agents. Most of those agents produce responses that are confidently wrong, inconsistently right, or too generic to act on. The gap between a demo that impresses and an agent that actually drives business outcomes comes down to three

Startup Battlefield 200 applications open: A chance for VC access, TechCrunch coverage, and $100K
Nominate your startup, or one you know that deserves the spotlight, and finish the process by applying. Selected 200 have a chance at VC access, TechCrunch coverage, and $100K for Startup Battlefield 200. Applications close on May 27.






Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!