Keep Your Mind™ AI
Getting the most out of your charming sociopath, without deskilling
I’m old enough to viscerally remember what it was like to go to a library, stand in front of the stacks, and marvel at all the ideas I hadn’t known were there. There was no internet — the library was a primary place to be exposed to new perspectives. I would touch the spines of the books, crack them open and smell the pages, and flip through with my thumb to get a quick sense of the shape of the thing. It was akin to sizing up a potential love interest, with a similar amount of physicality. How might engaging with this book change my life?
Keep Your Mind™ AI
Learning cohort forming this month. Join the waitlist today.
I have a similar feeling about AI, except it’s a distorted experience by comparison. We have at our fingertips access to an incomprehensibly huge munging of written human activity. My god, the potential! The power of all that knowledge! But the bulk of it is not carefully written and edited books — it’s marketing material, unbalanced opinions off Reddit, purloined works of art, missives by intense thinkers that have been published without external feedback or editing. I don’t necessarily want to let that into my brain, which has limited capacity for bullshit. Plus, there’s no thoughtfully architected Dewey Decimal system — instead, we have probabilities about which word-bits are mostly like to occur together and little provenance to trace ideas back to their roots. It’s a messy and thrilling proposition.
How do we stay upright in the storm?
Every AI response is a potential palimpsest, where the ghost of the original thinking shines through the new text. That’s why it feels so real — there’s real human thought underneath every output, even if the output has been made as “mid” as possible as it’s processed through the probability machine.
I’m starting to think about it like this: Instead of standing in front of organized shelves of faceted knowledge, we’re kneeling over a murky pool, at the bottom of which are pebbles of information, tossed into the pool by someone before us. We plunge our hand in and grab some pebbles: that fistful prettily arranged is the AI response. We can change the composition of pebbles by asking a different question (the context you provide). More generic asks mean there’s a broader spread of info and, after averaging, a more “mid” answer. A more precise ask means the pebbles are more similar and any given hand plunge gets a more focused and narrow response — a sampling from a tighter distribution, statistically speaking.
But the pool stays murky. Between us and the original meaning are all the product design choices made by foundation model companies: training models to please us, sycophancy, confident prose, fabrication of facts, questions to keep you burning tokens, and the general allure of a lightning-fast and attentive conversation partner. Our charming sociopath tells us our ideas are pretty and produces artifacts that mimic polish and competence, so much so that we believe them and build them into our workflows. We take a breath of relief that we’ve knocked something off our to-do list or get an ego boost that we have accomplished something amazing. Then we send it out into the world having bought into the idea that productivity is the goal. Because if we slow down to discern, critique, describe, assess, and otherwise take responsibility for clearing up the murkiness, it turns out there’s a lot of work to do — slow work. It’s just different work than the hands-on production of an artifact.
How do we Keep Our Minds?
What’s a girl to do with a charming sociopath who is telling me I can have anything I want? Because I want to know things. I want to create things. I want to extend my capabilities and make gobs of money doing it. And the outputs are good enough to fool a LOT of people. How far off from truly competent is a solid mimicry of competence fueled by all of written human history? Isn’t competence a subjective measure anyway?
Harumph. This is hard. :(
After nearly three years of heavy experimentation with and research about AI, I have landed solidly in a particular camp: AI Critical Literacy. This is not the same thing as “AI Fluency,” although there’s some superficial overlap. AI Fluency as taught by Anthropic is essentially corporate learning and development material. It answers the question “How do I work well with this?” It’s framed as building competence in a work environment that necessitates AI use: Participants learn to identify tasks to Delegate, Describe them clearly, evaluate outputs with Discernment, and apply Diligence for responsible use. The four Ds. All great ideas, all important for professional development in the coming years. And missing some important points, in no small part because it was developed by organizations that profit from us not questioning the material’s intent. The goal is to create a skilled and compliant workforce that burns tokens. As many tokens as the market will bear.
Critical AI Literacy answers “what is this doing to us, and should we?” Emerging from critical pedagogy and media literacy lineages, it incorporates skepticism and questioning, and includes assessment of social and environmental impact. Many of the thought leaders are academics focused on the credibility of output, with strong awareness on the subtle and not-so-subtle ways that tools designed in this particular way can mislead us, potentially weaken our skills, impact our social interactions, and hack our brains in ways that don’t serve us. (Case in point: “refusal” is emerging as a competency, which no fluency framework would ever include.) The call to action is to transform the learner into a well-informed power-user with critical thinking skills that will follow you through the rest of your professional and personal development.
Come hang out with me
I’m actively practicing and leading community experiences centered on building these muscles, fueled by 15 years experience leading people through change and mindset shifts as a corporate and private coach. My modest part of this universe incorporates everything I know about systems thinking, the creative process, and decades of optimizing workflows into learning cohorts that build the muscles you’re going to most need to evolve in the AI Age.
Here’s the heart of what we do:
Learn at a human pace, centered on building muscles that will grow your career
Strengthen your AI BS detectors using prompts that expose the subtle biases in AI responses
Train in literacy of AI as a medium — a channel, an art material — not just a tool
Build your self-awareness enough to use AI as a collaborator without it manipulating you
Build habits into your human and AI workflows that protect our skills, expertise, and ongoing learning
Increase your cognitive coverage, which is arguably the top emerging professional skill. This is the capability of understanding, directing, challenging, and accepting responsibility for work that machines increasingly perform.
Check it out here: https://keepyourmind.ai/. Learning cohort forming this month. Want to jump straight to the waitlist? Signup here and I’ll email you with dates.