Post

Four Pillars Gaming published hundred of AI-generated profiles over the past four months and when people asked to be taken down, they doubled down, defended their process, and refused. The ramifications of hallucinatory LLM "histories" threaten to erode truth and destroy the historical record.

rascal.newsAI-generated profiles of TTRPG designers ignite fury across social media4 Pillars Gaming’s new website adds AI-generated slop to an already ephemeral online historical record.

301 reposts684 likes

I can't imagine anyone that's complained about that site wanting to stay on that site at this point, least of all if it involves paying them 10% the cost of writing.

Why would anyone actually believe that editing is easier than writing when the writer is prone to error and hallucination? That is not a time-saver, but it is a common justification for using Ai to substitute real work.

It's not even "would that be better than AI", getting rid of the AI isn't even up for debate, it's just taken as a non-negotiable obvious prerequisite which is very funny even ignoring everything else wrong here.

An excellent recap, analysis and investigation in the 4 Pillars debacle. Nicely done Lin! Reading your piece I found myself wondering if there's some mental illness at play with Whitman. This mixture of intense dedication to his project – such as it is – despite intense pushback is so bizarre!

I had concerns when talking to one of the Universities linguistic staff some years ago, "If the meaning of words change so quickly that you cannot reference a common frame, would communication break down?" AI is doing this with all online knowledge.

Everything that gets online becomes grist for LLM's to throw up as "Information" Law student runs hypothetical court cases through ShatGPT and they are now part of the data Shat uses when idiots look for cases. The world burning to create GIGO.