3 ms·
People are obsessed with LLMs because they can be applied to unstructured data and produce human-like outputs, so they are applicable to vastly more processes w
by doe_eyes 2y ago
People are obsessed with LLMs because they can be applied to unstructured data and produce human-like outputs, so they are applicable to vastly more processes with vastly less oversight.
The kinds of "AI" you're describing have been with us for decades, but LLMs and their ilk are a more consequential development. And if we're passing laws today, that's what they will be regulating for the most part.
- tensor 2y agoWhich is what I'm say is wrong. If the oversight on insurance and other companies is lacking, it needs to be improved across the board not just with lip service to the popular technology of the day. The bias inherent is the exist risk scoring systems is absolutely equally if not more consequential than LLMs. Most of these regulations can replace "AI" with "technology, mathematical model, or human processes" and they will be 1000x better laws.