As digital technology has been playing an ever
larger role in our lives,
we have slowly been teaching our computers to fit our needs,
...just one feature or app at a time.
This has now reached a point where the simplest of tasks
seem impossible.
...and every website thinks it's the only website in the world.
So we now have 141 apps, 200 websites and 332 passwords to run our lives.
And 9 ways to talk to each other on platforms that don't talk to each other.
While we think it is our universal right to keep making more.
This is the world our agents are inheriting.
We built the most powerful reasoning machine in human history.
We squeezed it into a browser tab...
and basically every other place we had access to.
Now, three years later, we have run out of places to squeeze it into.
Turns out AI is a lot more useful with access to the right context.
⌘⇧4 ⌘C + ⌘V
But the way we currently use computers was never designed
for sharing context with any other intelligence than our own.
So how could it be different?
Turns out there's a much smarter idea,
instead of squeezing 141 different intelligences into different platforms,
we squeeze one intelligence into all of them.
But when we actually built it,
we kept finding that most of the code is wasted on translating messages between platforms.
The intelligence is no longer the bottleneck.
Our computers are.
So here's a wild thought...
What if instead of teaching your agent to use your computer,
we built our computers for our agents.