AI is scaling fast, but without a system to preserve context, accountability breaks, trust erodes, and value concentrates in the hands of those who control memory.
Can't resonate with the article more. You can't control something when you don't even shape its finite state. Remember the good old mantra "People, Process, Tool" to subject computerisation under boundary of tool (aka automation) in order to run organisation well? Where do AI stands in the equation now? The orderly structure will arrive when we have a conclusion over it. Is still rest on biological intelligence that is dominion because AI cannot audit itself.
Hi Jimmy — thank you, that's a really thoughtful take.
The “people, process, tool” framing is exactly where the tension is showing up.
AI doesn’t sit neatly in the “tool” category anymore. It’s starting to behave more like a participant in the system, which breaks the old boundaries that kept things governable.
That’s why your point about finite state and control matters. If we can’t clearly define what happened, in what sequence, and why, then we’re not really managing the system. We’re reacting to it.
I’d slightly push on one point though.
It’s not that AI can’t audit itself. It’s that we haven’t built the infrastructure that allows it to be audited in a trustworthy way.
Audit requires memory.
Memory requires structure.
And structure has to be designed upfront, not reconstructed after the fact.
Right now, most AI systems operate without that layer.
So I agree with you that biological intelligence is still the backstop. But the real shift will happen when we can extend that oversight into the system itself through verifiable records of interaction.
That’s when “people, process, tool” gets redefined into something more stable again.
Can't resonate with the article more. You can't control something when you don't even shape its finite state. Remember the good old mantra "People, Process, Tool" to subject computerisation under boundary of tool (aka automation) in order to run organisation well? Where do AI stands in the equation now? The orderly structure will arrive when we have a conclusion over it. Is still rest on biological intelligence that is dominion because AI cannot audit itself.
Hi Jimmy — thank you, that's a really thoughtful take.
The “people, process, tool” framing is exactly where the tension is showing up.
AI doesn’t sit neatly in the “tool” category anymore. It’s starting to behave more like a participant in the system, which breaks the old boundaries that kept things governable.
That’s why your point about finite state and control matters. If we can’t clearly define what happened, in what sequence, and why, then we’re not really managing the system. We’re reacting to it.
I’d slightly push on one point though.
It’s not that AI can’t audit itself. It’s that we haven’t built the infrastructure that allows it to be audited in a trustworthy way.
Audit requires memory.
Memory requires structure.
And structure has to be designed upfront, not reconstructed after the fact.
Right now, most AI systems operate without that layer.
So I agree with you that biological intelligence is still the backstop. But the real shift will happen when we can extend that oversight into the system itself through verifiable records of interaction.
That’s when “people, process, tool” gets redefined into something more stable again.