pull down to refresh
This is perhaps an error of my framing: Block's edges are still occupied by bots (writing code) with humans communicating, coordinating, and integrating at the scale humans are good at. To the extent human employees existed to communicate, coordinate, and integrate the work of 5000 humans, it sounds like their theory is that bots (with human operators) are better at that.
reply
It depends what the middle area is for. If the issue is one of managerial bandwidth, then there are lots of upper management largely because there’s lots of lower management.
If AI enhances managerial efficiency, then the org chart will get flattened.
reply
Who am I to question Dorsey, but this is almost the opposite of my instincts.
I would think that the bots would operate best at the edges of a process, where the tasks are often well defined and scoped. It's in the hazy middle where you need multimodal communication and coordination between tasks of various types, data models, and evaluation metrics. That hazy middle seems much harder for bots than for humans.
I can see how you need a human at the edge if it's genuinely unexplored space, but that's rare and not the bulk of the activity of most companies, I'd think.
But yeah, this surprises me because my first instinct is that where humans are needed most is the communication, coordination, and integration layers.