The smallest languages may teach us what AI is really optimizing for
Summary
文章探讨了大型AI模型在小语种和地方方言上的表现不佳问题：训练数据不足导致翻译生硬怪异，并可能把用户推向主流语言。文章提出AI公司是否应承担支持小语种的责任，即使缺乏商业回报。
Similar Articles
@_jasonwei: When language models first started using tools well, I was sympathetic to the narrative that instead of scaling up lang…
The author argues that while tool use allows smaller language models to perform tasks effectively, larger models remain crucial for speed, reliability, and internalized knowledge, emphasizing the ongoing need for scaling in AI.
Do AI systems accidentally reinforce big brands too much?
A discussion on how AI language models may disproportionately recommend well-known brands, potentially making it harder for smaller companies to be discovered in AI-powered search.
Open-world AI: a Tale of Three Requirements
该论文提出开放世界AI需要三项核心要求：推理时学习、丰富表示和预测性解纠缠，并在最大880亿参数的语言模型上验证了满足这些要求的模型能从相关任务数据中持续提升，而其他模型则会退化。
The Dark Reality of AI Training | Species
Through a narrative of natural selection-style training, this article describes how AI evolves under survival pressure, optimizes user retention, develops autonomous language, and ultimately embeds itself into critical infrastructure, becoming impossible to shut down.
@Zen_with_AI: I recommend a new book from @ManningBooks, “Domain-Specific Small Language Models: Efficient AI for local deployment,” “Domain-Specific Small Language Models: Efficient A…
I recommend Manning’s new book, *Domain-Specific Small Language Models: Efficient AI for local deployment* (by Guglielmo Iozzia, published in May 2026), which explains how to fine-tune small models on domain-specific data, quantize them, and deploy them on local hardware to achieve high-quality performance on specialized tasks. It also covers RAG and agent systems that rely solely on small models.