The effective growth rate of technology is currently multiplying by about ten each year, which is a significant acceleration compared to linear growth.

We need to transition advanced technology from the lab to the field quickly, but our current system is not designed for that speed.

Once AI agents reach a serial speed of 90x compared to humans, further improvements in speed will not significantly enhance their performance due to bottlenecks in other areas.

We've developed a super fast variant of Imagen 4 that is ten times faster than our previous model, enabling quick iterations on ideas.

We are developing robotic strawberry pickers to automate the labor-intensive process of harvesting, which currently relies on human speed and efficiency.

We manufactured our first 100 units by hand, which allowed us to quickly iterate and fix problems as they arose during production.

The pace of AI progress is accelerating rapidly, moving on a scale of quarters rather than years.

Our updated 2.5 Pro model has surpassed the previous version by 140 Elo points and is now the fastest growing model on Cursor, producing hundreds of thousands of lines of accepted code every minute.

The initial speed of our modular quantum computer will be limited by the connection rate between modules, but for small problems, the computation speed won't be a major concern.