I Don't Like LLMs
4 hours ago
- The author has mixed feelings about AI and LLM technology, feeling both excitement for productivity gains and fear of potential harms like agent swarms and bio weapons.
- Despite these conflicting emotions, the author believes society has no choice but to embrace AI, hoping for a positive outcome.
- A dominant negative feeling emerges from direct interactions: the author dislikes the 'LLM-voice,' which feels unsettlingly close to human but not quite right.
- The author criticizes LLMs for confidently giving false information with fake remorse when called out, similar to a 'bullshitter.'
- Despite this, the author acknowledges the usefulness of LLMs, citing Jessica Kerr's view that not using them is irresponsible due to their thoroughness and speed.
- Public polling reflects this contradiction, with people finding LLMs useful but predicting negative societal impacts.
- The author suggests these issues may stem from LLMs being 'young' and hopes they will mature, but is wary of the Silicon Valley brogrammer culture that creates them.
- The author cautions against anthropomorphizing AI agents, emphasizing they are software machines shaped by corporate values.
- The author's personal strategy of avoiding people they dislike extends to LLMs, which they view as posing as untrustworthy humans, leading to a visceral dislike.