The great big Ai LLM thread. Github code, blogs & opinions, walkthru's, trainer's & more

community off-topic
,
81

Something to increasingly watch out for.

82
83

Different tools, different prompting techniques. Why? Different human languages, some LLM’s will work better if the prompt is written in the native language by a native language speaker.

English is different to languages that are encountered as you move east towards Russia, those languages are more instructional. Chinese is surprisingly closer to English (subject–verb–object) than Russian, so that favours LLM’s like Qwen because its Chinese built & found on alibaba.com. Chinese could be considered one of the first emoji languages, until the west caught on because of early mobile phone sms charging plans.

So choosing an LLM model is more than just choosing the one with the greatest number of parameters.

84

I knew about the XZ backdoor with this video. Worth checking it out. Good for non technical people

BTW, the social engineering involved in this hack is obscure and fascinating at the same time.

85

Yeah I saw or read about this a while ago. I liked the Veritasium video on the rogue satellite knocking out GPS. Surprisingly basic science.

Is AI ruining our skills? Early results are in — and they’re not good

Reliance on artificial-intelligence tools degrades the abilities of physicians and software engineers, studies show.

86

I wonder how quickly Ai can alter Favicon’s to make them go rogue.

https://www.reuters.com/technology/norway-imposes-near-ban-ai-elementary-school-2026-06-19/

87

Pretty decent video, but dang that music was over the top. :slight_smile:

I did appreciate the info though.