When intelligence becomes abundant, everything depends on what cannot be generated.
Abundance is not the story. The story is the inventory of what stays scarce.
Walk into Node in Devonport and several versions of the future are usually sitting in the same room. A Bitcoin miner hums in the background. Live network data moves across a screen. A teenager is working something out at a computer. An adult is asking what an AI model can actually be trusted to do. On another table there may be a half-built project, a cable, a tool, or a problem that refuses to fit neatly into a slide deck.
That mixture is deliberate. Node is physical because I do not believe understanding arrives through explanation alone. People need contact with the thing itself: the machine, the model, the network, the other person who knows something they do not. They need the chance to try, get it wrong, notice why, and try again.
In that room, something becomes obvious. We are entering an age of extraordinary abundance. More software, words, images, analysis and competent-looking output can now be produced by more people at lower cost than at any point in history. Yet the more output floods the room, the clearer its limits become.
Most discussion about AI begins with what it makes abundant. Intelligence on demand. A tutor for everyone. Software built from a sentence. Expertise compressed into an interface. These claims are not entirely wrong. They are simply incomplete.
AI is a multiplier. It can expand a strong idea, accelerate an experienced operator, and make knowledge executable. It can also multiply confusion, conceal weak foundations behind polished language, and allow people to produce work they cannot properly evaluate. The output may look the same right up until the moment consequences arrive.
So the more useful question is not what becomes abundant. It is what remains scarce after abundance arrives. Where does value move when plausible output is nearly free? What can a model imitate but not acquire? What must still be earned, lived, tested or protected?
Every Node Labs paper is an entry in that ledger. Together they point to three forms of scarcity that matter more as machine capability rises: earned judgment, lived domain truth, and neutral rules.
Models can reproduce the visible residue of expertise. They cannot reproduce the years of feedback that taught an expert what to ignore, when to doubt the obvious answer, or how to recognise a failure before it becomes legible to everyone else. Judgment is built through consequence. If we automate the junior work without rebuilding the learning loop, we do not eliminate apprenticeship. We eliminate the conditions that create the next generation of experts.
The person inside a problem often knows more than the systems built around them acknowledge. A tradie, teacher, operator, caregiver or small-business owner carries knowledge in routines, exceptions, relationships and hands. AI lowers the wall between knowing and building. That does not make domain knowledge obsolete; it finally gives it leverage. The person who genuinely understands the problem can now shape the tool more directly.
Digital scarcity is not new. Games taught millions of young people to understand supply, ownership, markets and rule design. They also taught the limitation: an administrator can patch the economy, spawn the item, freeze the account or close the world. Bitcoin matters because it offers scarcity without that master account. Its rules are not neutral because people are perfect, but because no person gets unilateral control over them.
These scarcities share a structure. Each is produced through contact with reality.
Judgment comes from decisions that can be wrong. Domain truth comes from living close enough to a problem to encounter its awkward exceptions. Trustworthy scarcity comes from rules that remain binding when someone powerful would prefer an override. None of these can be secured by making an output more fluent.
This is why the future of education cannot be reduced to teaching people how to prompt. Prompting matters, but it sits downstream of something more important: the ability to form a useful question, inspect an answer, test it against the world and accept responsibility for what happens next. A young person who can generate anything but judge nothing has not been empowered. They have been given acceleration without steering.
It is also why the future of work will not belong automatically to the most technical person or the person with the newest subscription. It will favour those who can combine a real body of knowledge with the leverage of the machine. AI removes translation layers. It lets the person nearest the problem participate more directly in the solution. That is a genuine redistribution of capability—if we recognise what the machine is multiplying.
And it is why Bitcoin belongs in the same body of work. Intelligence abundance raises the value of systems whose constraints can be verified. In a world where images, identities, arguments and outputs can all be generated, rules that cannot be casually rewritten become more—not less—important.
Node Labs thinks in public, but Node itself exists in a room. The research and the physical space are two parts of the same experiment.
If judgment is scarce, we need environments where people can build it. If domain truth is scarce, we need people with different kinds of knowledge working beside one another. If neutral rules are scarce, we need ways to make systems visible enough to inspect rather than merely trust.
That is why Node is not simply a website, an AI course or a Bitcoin club. It is a workshop for agency: a place to meet real tools, real constraints and other people. The screens and miners matter. So do the conversations, the projects that fail, and the moment someone moves from consuming a system to understanding it.
The purpose is not to predict the future perfectly. It is to help more people remain capable inside it.
The papers can be read in any order. They are different doors into the same argument.
There will be more entries. AI will become more capable. Output will become cheaper. The temptation will be to treat everything newly abundant as newly valuable.
The work of Node Labs is to keep the other ledger: the things that cannot be summoned from a text box, the capabilities we still have to cultivate, and the systems whose rules deserve to survive contact with power.
— Jason Langis
Founder, node.org.nz