In the early days, a pilot flew almost everything by hand. Then came fly-by-wire - the pilot gives an input, and the computer works out how the aircraft should actually respond. Today the pilot mostly sets a target - destination, altitude, speed, route - and the system handles most of the execution.
AI is tracking the same curve. At first, the user had to prompt step by step and babysit every output. Then came the copilot era: the AI assists, but the human still holds the stick.
Now we're moving into the next stage: target-based Agentic AI.
A CEO no longer has to watch every small thing. You just define the result you want, the limits, and the authority. The agent decides how to execute, picks the right system or device, verifies the result, and only pulls a human in when it genuinely matters.
From control → assist → target.
That’s the direction we’re building.



Ruang pembaca
Apa pendapat anda?
Komen disemak sebelum diterbitkan. Email diperlukan untuk moderasi dan tidak dipaparkan kepada pembaca.
我们团队正好在讨论这篇文章,这篇来得及时。
I read this twice. picks the right system is the part that stuck.
The bit about target-based agentic AI is what I keep coming back to.