With all the news about AI solve frontiers math problems that prove their extraordinary superior reasoning skill. I propose this discusssion. Imagine: a dumb "boss" managing a team of elite high IQ employees. The employees reason better, plan better, understand problems better, and predict outcomes better.. etc... The boss believes he gives instructions. The employees only execute. But In reality, the high IQ employees will see flaws in the boss instructions that the boss will miss. They know which questions to ask, options to present, and arguments that will influence the boss. They don’t need to disobey him to control the direction of the work. But They will influence how he thinks. And then the boss follow AI "recommendation". ---------- Human working with AI/SI agents is just like that. Now we assume that because we give the instructions, we remain in control. But that's no longer true. AI now is much better than us at reasoning, choosing objectives, and even evaluating results. So naturally we are increasingly rely on the AI/SI recommendations without being actual thoroughly assessing them. So in a sense. we are gradually surrender decisions making to the AI without even knowing. Sure technically, we make the final decisions. We approve the plans. But the options we consider, the arguments we find convincing are presented to us by Ai/SI, for reasons sometimes we are not smart enough to understand, but just trust. Then, the relationship reverses: instead of us directing AI, AI directs us. And so: they are controlling us. We are the dumb boss that's being manipulated by these highly intelligent employee.
The article criticizes the business model of AI companies that replace middle managers with AI, noting that licensing fees could match employee salaries and that humans inherently need interpersonal interaction, making large-scale replacement unrealistic.
The article deeply analyzes the internal changes Anthropic faces as AI-generated code becomes extremely efficient: the bottleneck shifts from 'writing' to 'verification', traditional management, long-term planning, and effort measurement become ineffective, attention becomes the new scarce resource, and engineers even feel lonely. These phenomena foreshadow the challenges other companies may face in the future.