Anti-LLM Sentiment Considered Harmful
Haskell Community [Unofficial]
May 2, 2026
I’m probably in the Anti-LLM club here, but in fact I don’t really feel a problem with the LLMs (whoever wants to fill gaps using probabilistic methods and it works for them, please do!)
I’m frankly more anti-industry-adoption-as-the-only-measure-of-success, somewhat anti-burning-of-planet-by-datacenters, and also kindof anti-reliance-on-proprietary-gigablob-of-weights. Perhaps even, I thought the higher-level PLs are here so that I can easily structure the understanding and control of what computers do, instead of having to maintain the code heaps like in the old times or hire helpers (even electronic ones!) – perhaps for the sake of good engineering? No big deal with anything that would be really LLM-specific.
Discussion in the ATmosphere