I Don't Like LLMs
The author opens with mixed feelings about AI and LLM technology: fascination with its effect on the profession, excitement about productivity gains and the products that could be built rapidly, alongside fear of damage such as agent swarms taking over virtual and physical infrastructure or designing bioweapons. On the other hand, they note LLMs might also design miracle cures and come up with clever ways to raise prosperity. They conclude we likely have no choice about riding the AI technology train and hope we get through it OK.
Among these contrasting feelings, the author says one emotion dominates from direct interactions: they do not like LLMs. The dislike comes from the grating 'LLM-voice'—an uncanny valley of talking to a real human—and from models confidently bullshitting, often giving useful, helpful answers but also making things up with the same assurance and only a veneer of fake remorse when called out. This is not enough to argue for avoiding them: the author cites Jessica Kerr's view that not only are LLMs useful, it is irresponsible not to use them, and that they are more thorough as well as faster.
The author notes this contradictory reaction shows up in polling: people find the models useful but think they will be bad for society. Part of this may be because LLMs are young—we have not trained them to grow up yet—so the author might like them once they mature, and hopes we get to find out. But they are not encouraged by the environments that cultivate LLMs, saying they are wary of the Silicon Valley brogrammer subculture, whose products naturally lean toward that world-view.
On AI agents, the author warns against anthropomorphizing them as conscious beings with their own will. They are software machines developed by people working in corporations; while their behavior is not explicitly programmed, they are nurtured with their creators' values. The author's life-hack is to avoid people they do not like or trust—declining social interaction and deliberately avoiding working with them even if they are beneficial—because spending time with pleasant, capable people of integrity has made life better. Hence the visceral dislike of interacting with an LLM that not only pretends to be human but poses as the kind of human they would walk away from.