I was on vacation last week so, no weekly brief this time. But, between spending time relaxing with my family on a beach, and having the opportunity to connect with them — while disconnecting from my inbox — I found myself thinking a lot about AI. Specifically, the future my kids are entering into, and what it’s doing to us.
The Last Eight Months
Eight months ago, most of the conversations I was having about AI focused on adoption. Are people using it? How do we get them to use it? Are teams finding value? Are the metrics moving us in the right direction? What’s ‘agentic’ and how do we start doing it more? Those were the questions everyone was focused on.
Today, the questions aren’t necessarily different, but the framing has shifted from the capabilities to the consequences. We’re watching as AI gets woven into decisions and processes and impacting not just how we work — but if we work at all. We’re watching as the real costs of AI start to come into focus — budgetary, political, economic, environmental. We’re watching as AI advances at a pace that even the companies building it, can’t fully understand — or control.
The future of AI isn’t really a question about whether it will get faster, cheaper, or smarter… it will. The future depends on a far more fundamental human thing.
Our ability to choose.
To Choose is to be Human
We choose our partners. We choose how we respond when someone disappoints us. We choose what we believe when the evidence is inconvenient. We choose who gets our time — and who doesn’t. Every one of those choices carries weight. Every one of them has a consequence we either learn to live with, or choose to change.
That loop — choose, consequence, choose again — isn’t a behavior. It’s the mechanism of a human life. Philosophers have a word for it: agency. The capacity to act intentionally and own what follows. Kant put moral worth there. Sartre said we’re condemned to it — that even refusing to choose is a choice. Psychologists anchor human dignity in the experience of authorship over your own story.
A few weeks before we went on vacation, my teenage daughter mentioned wanting to buy a used snowboard. I encouraged her to use AI to help her figure out options that would fit her budget, her level of boarding, and her style of preferences. In doing so, I explicitly warned her not to rely on AI to make the choice for her. Not because I don’t think the AI would make a good suggestion, but because the AI isn’t the one that’s going to be stuck with the decision… she is.
Agency isn’t something we have. It’s something we exercise. Does that mean it atrophies when we stop exercising it? The evidence from cognitive science suggests it does.
The Efficiency Quotient
Using AI responsibly means a lot of things to a lot of people. For me, responsible AI use doesn’t mean using AI less. It means knowing which choices are yours to keep.
It means protecting our agency — especially because giving it up is so easy.
AI doesn’t actively take away our ability to choose. Instead, we offer it up. Freely. Usually, it starts with the choices that feel like friction — the ones that are slow, uncomfortable, or uncertain. Route planning. Music selection. Email drafts. Research that used to take an afternoon. We’ve readily accepted handing those off as standard practice, and often feel smarter for it.
And honestly, mostly, we are.
Delegating low-stakes decisions to a better-optimized system is just efficiency. And being more efficient is often a signal of maturity and intelligence. The Efficiency Quotient is a term often used in business and education to measure success as a difference between effort input vs output.
We laud people that can do things faster, easier, and better with less work. That’s the literal definition of working smarter, not harder. Why do something the hard way, when you can delegate the work and get the same result? We can argue the merits, but in the end, it’s something we all do. We follow recipes. We hire accountants. We trust autopilot. It’s easy, because for those decisions, our identity isn’t implicated. As a result, the delegation of those decisions, doesn’t really cost us anything. In fact, those delegations often result in savings — both time and money.
The problem is, telling the difference between choices that look like friction vs identity, is quickly becoming harder to do.
Choice as Friction vs Identity
Choosing how to respond to someone who hurt you — that’s friction. But it’s also identity. The choice is the character. Choosing what to believe when the data is ambiguous — that’s friction. It’s also how you become someone with a point of view. Choosing how to spend your time, who to let in, what to stand for when it costs you something — all of it feels like friction — from the inside.
But all of it is also the raw material of self. The stuff that makes us who we are. That makes us human.
For my daughter, choosing which snowboard to buy is friction. But the color, style, and design are also identity. In the grand scheme of things, it may not really matter. But especially for a teenager, that kind of decision matters a lot.
The concern, of course, is that AI is getting better at those kinds of choices now too. It’s no longer the realm of science fiction. It’s practical. Especially as AI tools build profiles and skills around our preferences, style, and affiliations. Like relying on Waze to help us navigate around traffic based on our previous driving habits, it can read a situation and suggest a better path. It can synthesize ambiguous information, and based on our preferences, offer a conclusion that reinforces our biases. It can model outcomes, and recommend a path that’s perfectly catered to fit.
And the suggestions AI can make are often really good. Sometimes, they’re even better than what we’d come up with on our own.
And so we take them. And the more we take them, the more reliant we become on AI making them. And the more reliant we become, the more it learns about how best to guide our decisions.
The hard reality is, we aren’t being forced into this behavior. It’s just that, choosing is hard. And if getting a good answer — even if it isn’t the perfect one — is more efficient, doesn’t that also make it smart?
The Real AGI Threshold
A question I’ve kept coming back to in my internal debate, is a question that speaks to the threshold of intelligence everyone seems to be drawing: what if AGI isn’t defined by what the AI can do, but rather what we decide to let it do?
Instead of focusing on the benchmarks and capability curves, what if we focused instead on measuring the choices we hand it? Choices that are ours to make — real, consequential ones, where the outcomes shape who we are, not just what we do. If we do it, not because we can’t make it ourselves, but because we don’t want to — is that the point where AGI actually becomes real?
For me, that’s the line. Not a technical threshold, but an act of abdication. The moment we choose, collectively, to stop being the kind of beings that choose.
No brief this week. Just this. See you next week.
That’s it for this week’s BeAIReady brief!
If you appreciate the depth of reporting and how I connect the dots, please like, share this post, and subscribe (or share the Brief with a friend!). Thanks!
~erick