I've spent the past six years teaching people that their thinking matters. I've been yelling from the rooftops "your words over their words," to "stop collecting, and start connecting," and to prioritize "your insights over more highlights."
So how can I—of all people—be okay telling you to build an AI thinking partner?
How can I be okay with letting AI think for me?
I'm not.
Let me show you what I mean—this Thursday, April 16th at 9:00 AM Pacific.
In 60 minutes, I'll walk you through principles, frameworks, and live examples behind how to build a thinking partner that thinks alongside you—not for you.
Yes, this includes a live demo showing how an AI OS works.
This weekend, you'll receive the two free emails from me—an overview and manifesto of sorts—to understand the unexpected ways AI works and how to be healthy with its use.
Can you think with AI and not lose yourself?
This fear is well-founded. AI psychosis is a real thing. People are losing themselves in AI. They're losing their You-ness, and becoming Average. Which is what AI is: the Average of all that has been poached from human knowledge. And for those who use AI incorrectly, they become just another AI: an Average Individual.
You have two camps of AI users:
- The AI Maximalists are those who don't care about privacy, security, making their own links, or thinking their own thoughts. They don't have boundaries and don't care. That's not me. That's not you (and if it is, please know that you're in the wrong spot).
- The AI Mindfuls are those of us who are curious about AI. We experiment with it. We have found useful ways to use it. But we have caution. We have boundaries. We have specific reservations. That's me. That's most of us here. And that's why this is so personal for me.
Stay with me here...Every thoughtful person I know is tiptoeing around something. "That which cannot be named." No one wants to say it out loud because it puts them at risk of being seen as less unique. I get it. It's the quiet guilt of finding AI useful when you've built your identity on thinking for yourself. (And for good reason. With all the AI slop out there, AI is not something real thinkers want to be associated with, right?)
Well I'm naming it.
I have an AI thinking partner. One that thinks alongside me, not for me.
One that knows a lot about how I think and how I work—not just the aggregate of scraped information on the internet. Because since AI has come onto the scene, I've been obsessed with finding the timeless. Not the "god prompts." Not the neat hacks. I've been obsessed to find and develop timeless principles, practices, and frameworks with AI that keep me in the driver's seat—both today and far into the future.
And I have.
I've figured out how to build an AI thinking partner without giving up the one thing I've been protecting all along: my thinking.
Can I show you how?
My latest YouTube video explores the AI OS.
Idea Updates
Obsidian Web Clipper. My next YouTube video is going to walk through the Obsidian Web Clipper. It's an interesting evolution from Evernote's ground-breaking web clipper, to solutions like Readwise Reader, to something like Obsidian Web Clipper.
Delta Day. We had 71 community members stop in for our 6-hour Delta Day on Tuesday. We conquered so much resistance and got so much accomplished on our biggest efforts it wasn't even funny. We'll probably do smaller 3-hour Delta Days later in the year.
Idea Exchange
Artemis II. It's been great checking in on the space flight that went around the moon and back. Have you seen the photos? They're inspiring. Check them out here. Note that the page has multiple galleries of photos worth checking out.
The Karpathy Method. As I shared last week, Andrej Karpathy shared a viral idea about LLM Knowledge Bases. It is a quality contribution to knowledge management, but the level of virality is perplexing. Mainly because it doesn't scale beyond ≈ 250-750 notes. Somewhere at that point, the index proposed is no longer effective. Just know, the Karpathy Method is best for small collections of research. Micro-vaults. Micro-wikis. It will definitely crumble once you get in the thousands of notes.
So what's the great breakthrough? It's the idea of having a separate wiki where you feed clipped articles and sources from online and then the AI makes atomic notes and links between (not you). Then you can query your micro-vault, get tailored answers, and deepen your comprehension on your tightly curated topic in a hurry (and grow with those insights over time). Essentially, it's using AI to build a micro environment for rapid-learning. And that's pretty cool.
I think a large part of the virality was Karpathy helping people who have never heard of PKM discover what many of us who use Obsidian have known and been excited about for years.
One of the longer term questions is how to treat a folder of notes in which you are not the main thinker. Do you just throw it into your personal vault/ideaverse/PKM system? I don't recommend it for various reasons, mainly so that you don't drown out or confuse your own writing and voice with AI's. That might not matter to hard-core AI maximalists, but it does to the knowledge management crowd.
This is why you MUST own your own AI model. Anthropic: You Can't Use OpenClaw With Claude Without Paying Extra. They essentially gave one-day's notice which is really sudden if you were entrenched in that workflow. What we teach allows you to own your own AI OS—the playbook that no company will be able to just take away.
Stay connected,
Nick
P.S... If you are into or interested in AI, don't miss my upcoming live session on how to build a future-proof thinking partner.
To respond to this newsletter, just hit reply. I love getting replies, and read all of them, but I have sadly come to the conclusion I can't realistically reply to most. Trust me, I hate this. But, well, Life puts limits on all of us. Thank you for your understanding. (And if you received this email from a friend, and would like to subscribe, please go here.)