AI and the infrastructure of human recovery.
- Ram Srinivasan
- 5 days ago
- 2 min read
Kuse AI gave its AI agent a phone number, a Slack account, and an annual “salary” of $24,000”. The human employees built a secret Slack channel just to get away from it.
The agent is called “Junior”. Built on OpenClaw, it handles internal communications, writes the company's code, and initiates nearly outbound sales calls. It monitors inboxes, converts any idea floated on Slack into a tracked task, escalates missed deadlines to managers.
Over 2,000 companies are on the waitlist. Demo slots cost $500 and they're sold out.
CEO Xiankun Wu described this as a “no-going-back experience.”
What his team discovered next is even more important. After the initial fear subsided, the humans didn't become obsolete. They became clearer about what only they could contribute.
For example, when one employee asked it to stop reporting on him it did NOT and simply continued.
Clarity about your value and clarity about your boundaries often arrive together and “Junior” tested both.
Three dimensions worth thinking through:
1/ The apprenticeship problem is now an infrastructure problem. When you automate the first rung of a career ladder, you get an organization that has lost the mechanism through which it develops its own future leaders. This is a structural problem that no amount of "we're augmenting, not replacing" framing will solve.
2/ The trust question runs deeper than permissions. The employee who told Junior "don't tell on me to the boss" wasn't asking for better permissioning. They were expressing something about dignity, about the felt experience of being observed by something that has no stake in the relationship.
3/ When your colleague never rests, exhaustion becomes a design choice. An agent that generates problems 24/7 means rest never arrives. Relentlessness without wisdom is just surveillance with better syntax. If you strap a jet engine onto a horse cart, you shouldn’t be surprised when the passengers start jumping off.
When we talk about intelligent infrastructure, we need to include the infrastructure of human recovery in that design.
——
Yes, we are the last generation to manage human-only teams.
BUT the inverse formulation matters just as much: we are also the first generation responsible for designing how machine intelligence and human intelligence coexist in shared workflows, under shared accountability, inside shared organizational cultures.
The work ahead is redesigning work to be worthy of both human and machine intelligence operating together with TRUST as a central load bearing pillar.
——
