Comment on Lies, Damned Lies, and LLMs: AI is a Con
Powderhorn@beehaw.org 3 weeks ago
Confidence mixed with a lack of domain knowledge is a tale as old as time. There’s not always a con in play – think Pizzagate – but this certainly isn’t restricted to LLMs, and given the training corpus, a lot of that shit is going to slip in.
It’s really unclear where we go from here, other than it won’t be good.
jarfil@beehaw.org 3 weeks ago
That’s why AI companies have been giving out generic chatbots for free, but charge for training domain-specific ones. People paying for using the generic ones, is just the tip of the iceberg.
The future is going to be local or on-prem LLMs, fine tuned on domain knowledge, most likely multiple ones per business/user. It is estimated that businesses are holding orders of magnitude more knowledge, than what has been available for AI training. Will also be interesting to see what kind of exfiltration becomes possible, when one of those internal LLMs gets leaked.
Powderhorn@beehaw.org 3 weeks ago
I’m sure that, as with Equifax, there will be no consequences. Shareholders didn’t rebel then; why would they in the face of a massive LLM breach?
jarfil@beehaw.org 3 weeks ago
It’s going to be funnier: imagine throwing in tons of data at an LLM, most of the data will get abstracted and grouped, most will be extractable indirectly, some will be extractable verbatim… and any piece of it might be a hallucination, no guarantees! 😅.
Courts will have a field day with that.
Powderhorn@beehaw.org 3 weeks ago
Oh, yeah. Hilarity at its finest. Just call it a glorified database and a day.