Martin Fowler: I don't like LLMs — not because they're useless, but because they act like people I'd walk away from

I Don't Like LLMs

Martin Fowler's Sep 17, 2026 post says his dominant feeling toward LLMs isn't fear or excitement — it's dislike. He acknowledges they're useful, citing Jessica Kerr's line that not using them is irresponsible, and references a Pew poll where Americans find AI helpful but bad for society. What bothers him: LLMs speak in an uncanny "LLM-voice," confidently bullshit, and offer only fake remorse when corrected. He also distrusts the Silicon Valley "brogrammer" culture that shapes them. His life hack is avoiding people he doesn't like or trust, and LLMs act exactly like those people. The post names no specific models, parameters, or technical approaches — it's a personal stance piece.

Why it matters: Martin Fowler's byline carries weight, and the 'I don't like LLMs' stance is a minority voice in the current hype cycle — worth surfacing. The piece has concrete arguments (LLM tone, confident bullshitting), not just raw emotion. Deduction: it's ultimately a personal essay wit...