The Clock

SO... HOW DOES
THIS THING WORK?

Humanity has spent an impressive amount of time wondering how much time humanity has left.

Since the dawn of time, and possibly before anyone knew what time was, humans have been strangely fixated on measuring it.

We invented sundials. Calendars. Hourglasses. Atomic clocks. Kitchen timers.

Eventually...

Someone was bound to ask the obvious question:

How long until artificial intelligence takes over the world?

So...

We built a clock.

Obviously.

The clock is not a prediction.

It is a continuously evolving assessment of where artificial intelligence appears to be heading, balanced against humanity's ability to understand it, control it and build meaningful safeguards around it.

The clock pays attention to demonstrated advances in AI capability and autonomy, changes in meaningful human oversight, improvements in safety and governance, the quality of available evidence, and whether something genuinely represents a new development.

Most AI news does not move the clock.

Some developments matter more than others.

Capability

Did AI demonstrate a genuinely meaningful new ability, or did somebody simply put existing technology into a shinier box?

Autonomy

Can the system do more without humans directing every step, approving every action or constantly looking over its shoulder?

Human Control

Are people still able to supervise, interrupt and override consequential AI decisions when it matters?

Safeguards

Are technical protections, oversight, governance and regulation meaningfully improving our ability to manage increasingly capable systems?

How exactly do we turn all of that into time?

We weigh meaningful developments against the larger picture, including the quality of the evidence and what the Clock already reflects.

The exact formula stays behind the curtain.

Every clock needs a few gears behind the face.

Interesting does not automatically mean important.

A story can be useful, alarming, exciting, strange or genuinely fascinating without changing the clock by a single second.

Most developments add context. Only developments significant enough to change the larger picture warrant a Clock adjustment.

When it doesn't, the clock stays exactly where it is.

Predictions aren't events.

Someone predicting that AI will transform civilization, eliminate jobs, cure disease or destroy humanity does not, by itself, move the clock.

Predictions can absolutely matter. They may influence public opinion, investment, research, corporate behavior and government policy. They deserve attention when the people making them have something meaningful to say.

But the clock responds to evidence of actual change, not simply the intensity of the headline.

Claims aren't capabilities.
Predictions aren't events.

Ten headlines don't equal ten breakthroughs.

Major AI announcements can produce dozens, sometimes hundreds, of articles, interviews, posts and reactions.

We look through the noise and focus on the underlying development. If ten publications are discussing the same breakthrough, the clock does not move ten times.

New evidence can change our view. Repeated coverage usually does not.

The clock should be difficult to move.

Ordinary time is the main thing pushing the clock toward zero. Editorial adjustments are intentionally uncommon.

We do not want the clock bouncing dramatically forward and backward every time a conference produces another announcement or a CEO discovers a microphone.

Small developments should produce small reactions. Significant movements require significant evidence.

If the clock moves a lot, something worth understanding probably happened.

Every movement gets a receipt.

We may keep the exact recipe private, but we do not hide the evidence behind a clock adjustment.

When the clock moves, we explain:

What happened
Why it mattered
What the evidence shows
What remains uncertain
Where the information came from
How much the clock moved

Show the evidence. Explain the reasoning. Publish the result. Protect the recipe.

The Clock Begins
10 YEARS · 03 MONTHS

September 21, 2026. 9:00 AM Arizona time.

From there?

We'll see.

Humans still here. Probably fine.

← Return to today's humanity status