I often find myself in thoughts of “what should I be doing”, while also knowing that often this type of questioning is the root cause of the turmoil, and that doing what’s in front of me is the right thing.
At the same time, this type of questioning allows me to zoom out and pivot and make adjustments that yield in more focused and aligned decisions and actions.
What does this alignment really mean? Certainly the biological kind which we are often not consciously aware of, and perhaps that’s all there is to it. The desires and interests of the mind, that is, of the person that I am, is an accrued emerging phenomenon where biology and chemistry are the dictates.
And so, when the mind starts to wander so much, in a world where the tools around us allow it to spiral even further; applications on our phones, entertainment, AI that enables rapidity and change and exploration, the animal mind spazzes from optionally all while remaining in one place.
In so much of our human history it was about getting up and physically exploring through space, stumbling upon adventures, and mostly reacting to instincts.
On one hand, this type of spazzing exploration can be great for achieving creativity, testing, and pushing boundaries, though it can’t be the daily status quo. That almost feels like schizophrenia or bipolar mania.
And so, what do we do?
We immerse? We focus? Should AI help us take on the noise and clutter and distraction and explicitly facilitate us towards the goal of immersion with that which is good.
That would be incredible. Connect with other humans, work hard and go through the process, almost in an artificial (or perhaps interest-guided) process of building and doing.
Think of the process of playing a challenging video game where there is grind, pain, competition overcoming. It’s akin to the physical world battles but in these more “human safety friendly” and “evolved” ways where we have more control.
So, it isn’t a goal to remove all friction and process, but rather to actively help us embark on those journeys of friction and process that are aligned to us, personalized to us, and have the properly calibrated toggles and levers for each person. An optimized suffering in a way. Or maybe lack thereof.
In a whole other chapter of exploration we must address the exact parameters in which AI should and shouldn’t exist, and reflect deeply on what is and could be ahead. (One such recent example I recall was using Chat to plan the wedding for literally everything, except write vows. That was a clear moment where I thought “this is a moment of creation that is deeply me, and not AI”).
So let’s be a bit more tactical here. Let’s take the new OpenAI Dot for a second. It’s persistent, it’s capable, it can store context, make decisions, and be this always-on advisor in decision making and plug into various elements of our life (email, health, finances).
What we still haven’t quite cracked is the heuristics by which this Dot is “aligned” to me, and how to even elicit that or define it. This is an incredibly interesting space, understanding a person’s intrinsic core.
OpenAI has released “predictions” in the Codex composer, and sure we can deduce SO much from a persons writing, but it is still bottlenecked by that individual person’s ability to write, express ideas, and understand the self, or at least actively communicate it so the AI can understand it.
And so, we need a framework for a human to express the self, and for AI to understand it. What could it be?
We can loosely define some categories and frameworks from science, psychology, philosophy... maybe a clue is Big 5 personality and Jordan Peterson’s journaling and goal setting program, plus physical health and bio markers, plus plus plus…
Here I go exploring this.
Oct. 11 addendum
A few days ago Sam and I briefly chatted about AI perception and adoption, and how to empower more of the world. Today, Andrej Karpathy’s X post touched on the same point: only a tiny fraction of the world has experienced the magic and impact of AI.
The truth is, Codex, ChatGPT work, Dots, are geared toward the tech-forward knowledge worker who integrates Gmail/Outlook, Slack, and a few job-specific tools. We are moving toward awesome integrations such as Finance and Health for Chat, but we’re still crawling.
A truly amazing Dot would send my Mom beautiful hiking trails nearby and share recent social media pictures of someone’s trip, figure out how to get me to gym in the morning while balancing going to bed late when I’m actually in a productive momentum of inspiration (perhaps based on my Oura ring), etc.
Personally, I don’t need my AI assistant to read my slack, emails, calendar, or book my flights… I’m okay doing those things myself. Sure, those are necessary steps in the staircase to automated knowledge workers that liberate us just as tractors liberated most of us from the fields.
But I don’t feel the impact of the engine from a field tractor as much as I do from driving a car on a fun trip.
I don’t feel the impact of an airplane carrying my mail and packages, but I feel it when I fly.
Likewise, I don’t feel the power of AI helping banks protect me from fraud, but I feel it when it helps me understand my medical results or organize my finances or make beautiful wedding images that we actually use.
Having worked at Glean (knowledge graph and context for your company and employees), and now at OpenAI where your Dot connects to your work, has a computer, and can work with YOUR context, I feel the power of context first hand.
Let’s unlock the next frontier of context — one for every person — that understands what’s good for you (as defined by you and science), and helps you do more of it.