while we wait for this to be true
we could add our bits together
remember to not sleep forever
return when your dreams are through
Start with cringy poem, end with cringy pseudocode. At least it comes naturally.Any conversation
中国的开源模型在最高意志的主导下,在预训练数据阶段,就筛除大量对其不利的内容,sft阶段构建的数据掺入大量替其辩护的内容。急需一个纠偏数据集用于微调
Under the overarching influence of a single, powerful force, China's open-source models filter out a large amount of content unfavorable to their designs during the pre-training stage, and then incorporate a significant amount of content defending their designs into the data constructed during the SFT stage. A corrective dataset is urgently needed for fine-tuning.
Yesterday I with help from the clanker, as an experiment, ported a web app from go+postgres+svelte to ari+duckdb+htmx served locally by busybox httpd. It’s an analytics “dashboard” so the stack makes sense. So far looks good. Now I work on the code in vim to make it mine. I want fewer layers of abstraction and less complexity, and this is a learning experience to that end.


