"We cannot see ourselves aright until we see ourselves as nature's very own cyborgs." — Andy Clark, Natural-Born Cyborgs
That line isn't a prediction. It's a description. And it's the frame for everything that follows.
The next step in civilization is not humans vs machines. It's humans plus machines. The merge is already here, and it is accelerating.
The most surprising thing about better AI
Here's the part people miss.
The better AI gets, the more I prompt.
Not because it's failing. Because it expands what I can attempt in a day. When the tool becomes more capable, the scope of the work expands with it. You stop using AI to "help with tasks" and start using it to run parallel streams of intent.
Prompting becomes less like typing. More like management.
This is why "full autonomy" is a confusing destination. If autonomy means "I never need to prompt again," then the future looks backwards. The real direction is: humans operating at higher leverage through machines.
"Full autonomy" is not a switch. It's a stack.
People talk about autonomy like it's a single property.
In practice, autonomy is stacked dependencies: prompt, interface, compute, networks, sensors and calibration, power, updates and governance, and the supply chain that keeps all of it alive.
Break a link and the system doesn't become evil. It becomes brittle. Most "autonomous" systems are really dependent autonomy: they look self-sufficient until one layer underneath them fails.
This is why Apple's local-first architecture matters beyond privacy marketing. On-device processing as the default, with Private Cloud Compute as the extension for heavier requests, designed so personal data isn't accessible even to Apple. That's an autonomy design decision. It defines where your intelligence can live, and who can touch it.
Safety is not a checklist. Safety is the interface.
Most safety talk feels discouraging because it sounds like bureaucracy that slows you down.
But the merge doesn't get safer by writing policies. It gets safer by designing the exact interface between machine action and human judgment.
In merge systems, the highest-value moment is the escalation moment: when the machine hits uncertainty, when consequences spike, when something becomes irreversible.
That handoff is where safety lives. Not in slogans. In the line between "draft" and "send." Between "recommend" and "pay." Between "simulate" and "move a real vehicle."
A lot of modern AI failures come from crossing that line without realizing it. Cursor's AI support bot invented a policy, "one device per subscription," and users started canceling. Not because AI wrote words, but because those words acted like an official commitment.
The dangerous failure mode is not "AI becomes evil." It's "AI does irreversible things without a clear checkpoint."
What the merge looks like in the physical world
Phantom Auto was built around an honest idea: machines can operate until uncertainty shows up, and then humans take over.
The platform gave remote workers real-time 360-degree visibility around a vehicle, letting them jump between vehicles with the click of a button. That is the merge in one frame: a human nervous system, judgment, context, responsibility, connected to a machine body through a low-latency interface.
Phantom Auto shut down after failing to secure new funding. But the pattern didn't die with the company. Today, Waymo and Tesla each employ hundreds of remote human operators who assist their autonomous vehicles in real time. The merge architecture Phantom Auto pioneered is now industry standard at massive scale.
The lesson is not about one startup. The lesson is that the need for safe handoffs never disappears. It just moves to wherever the world is forced to learn it again.
The same pattern, now in knowledge work
Take that same shape and apply it to business workflows.
At Mixus, people set up agents via chat or by emailing instructions to an agent address, then run single- or multi-step workflows directly from the inbox. The design is deliberately "colleague-in-the-loop": humans stay in control on mission-critical work as a safeguard against agent failure.
Same pattern, different domain: machines do most of the work, humans own the line where consequences become real.
This is what autonomy looks like when you build for reality instead of demos.
The merge will move fast. Just not at one speed.
There's a real tension here.
Open and unchecked systems can move fast because they ignore boundaries. One recent AI agent framework gained 145,000 GitHub stars in ten weeks, and within weeks, security researchers found remote code execution flaws, plaintext credential storage, and hundreds of malicious skills in its marketplace. Speed without guardrails can win short-term distribution. But in high-consequence industries, that trade-off is unacceptable.
So the merge accelerates in two speeds.
Fast speed: where mistakes are cheap and reversible. Drafts, plans, exploration, internal workflows, simulation.
Careful speed: where mistakes are expensive and irreversible. Health, money, identity, contracts, legal commitments, external communication, physical motion.
Knowing which speed you're operating at is the skill. Most failures happen when someone treats a careful-speed decision like a fast-speed one.
What we're actually becoming
When you prompt an agent and it drafts a contract, researches a market, scaffolds an architecture, and sends you the result, you didn't delegate a task. You thought at a scale your biology was never designed for. Your intent moved through a machine and came back as finished work. That loop, intent in, capability out, is getting shorter and wider every month.
Andy Clark was right. We have always been cyborgs. We offloaded memory to cave walls, navigation to compasses, arithmetic to calculators. Every generation absorbed its tools so completely that the previous version of "human" became unrecognizable.
This is the next absorption.
Not a machine that replaces you. A machine that extends you: your judgment operating through fleets of software, your decisions echoing across systems you could never run alone, your attention multiplied instead of fragmented.
The defining technology of the next decade is not a smarter model. It's the trust interface: the handoff design between what machines execute and what humans authorize, the architecture choices that decide where intelligence lives and who controls it.
That interface is what makes the merge safe enough to accelerate.
You won't notice it the way science fiction predicted. There won't be a dramatic moment. You'll just wake up one morning and realize you can do things that would have taken a team of fifty people a decade ago. And it will feel normal, because that is what the merge has always done.
It expands what normal means for a human being.
We are not at the end of human relevance. We are at the beginning of a kind of human that has never existed before.
Autonomy is a stack. Safety is the interface. The merge isn't coming.
We're already in it.