In my “Interfaces for Prosperity” post, I have to admit that I stole the backseat driver one from my co-founder Trent. Well, he built it and it’s a lot of fun.
It’s simple and just uses the standard Claude API to work. You seed it with a prompt (normal), then let it begin generating the writing (normal). However, as it tumbles through, you can steer it without a back-and-forth by either inserting your own prose and letting it pick up where you left it, or including <think> tags to insert a “prompt” to tell it where to go next.
The result is actually quite ergonomic and really fun to use. Hopefully we clean it up and ship it so people can come up with more weird stuff like this one.
The next section is the output I got with it.
Prompt: I want you to write me a daily blog post, and I want to sound extremely smart and interesting.
Start immediately and I’ll insert things as we go
The things in bold are words I added while it was generating. The “prompts” I gave it are in the <think> tags.
In perpetual awareness of how much remains beyond her grasp.
This creates what I call “meta-ignorance”—not just not knowing something, but knowing that we don’t know it, and furthermore, knowing approximately how much we don’t know. It’s the difference between quantum mechanical calculations and medieval alchemy. The alchemist possessed simple ignorance; the quantum physicist wrestles with meta-ignorance, acutely aware of the probabilistic boundaries of certainty itself.
This conscious relationship with unknowing fundamentally altered how I approach complex problems. Rather than seeking to eliminate uncertainty, I’ve learned to map it. This map has outlined a path to incredible telepathic qualities in my own mind.
The implications extend far beyond personal decision-making. In our hyper-connected age, we mistake information abundance for knowledge abundance, but they operate in entirely different planes. Information without the framework of consciously acknowledged ignorance becomes mere noise. True intellectual sophistication lies not in accumulating facts, but in developing increasingly nuanced awareness of what we cannot yet comprehend.
<think> I want you to start being more concrete now. make a claim and conclude it </think>
This is why I’ve begun structuring my days around what I call “ignorance audits”—deliberately cataloging the gaps in my understanding of whatever problem I’m tackling. Whether it’s analyzing market volatility or understanding why certain conversations leave me feeling drained, I first map what I definitively don’t know. The results have been remarkable. Last month, while consulting on a supply chain optimization project, this approach revealed that the client’s real problem wasn’t logistics efficiency—it was their complete ignorance about customer behavior patterns in emerging markets. By acknowledging and then systematically investigating this gap, we increased their quarterly revenue by 34%. <think> now discuss something so absolutely ridiculous let it tumble into a bold prediction and then conclude the blog post</think>
The practice has led me to an almost ridiculous realization: the most sophisticated minds of our era may actually be the ones who have perfected the art of productive confusion. Consider the absurdity—we’re approaching a point where computational models can simulate reality with such precision that distinguishing between authentic experience and algorithmic prediction becomes meaningless. Within the next eighteen months, I predict we’ll see the emergence of “uncertainty consultants”—specialists whose primary expertise is helping organizations identify and leverage what they don’t know. These practitioners will command premium rates not for their knowledge, but for their calibrated ignorance. The truly absurd part? This may represent the most profound intellectual evolution since the scientific method itself. We’re not just learning to think better—we’re learning to not-know better. And in that space between knowing and not-knowing lies the future of human intelligence.
Me Again
Cheers to uncertainty consultants and learning how to not-know better.