Random thought: What if all those #GhostJobs posted by recruiters are actually ways for them to gather a whole bunch of PDFs for #tokenization to train #LLMs?
@KathyReid Additional random thoughts: Wonder how many call centers are now feeding their years of recorded phone conversations to #LLMs for "training purposes"
@catsalad strong agree, and also feeding the audio, which can now be automatically transcribed, into ASR models, except that the ASR models are stratified by customer segment - so for example the ASR can now detect if the person calling is "rich white woman" or "low income older man".