Raw LLM Responses

> I imagine alot of the AI stuff would be closed systems which would make it pretty hard to hack. Iran thought the control software for the aluminum tubes they were using to refine uranium hexaflouride were pretty hard to hack. Hezbollah thought the supply line for their pagers to avoid Israeli surveillance of cell phone signals were hackproof, too. The more power we give to autonomous systems, and the more complex those systems are, the more exploitable points of failure there will be in those systems. Do you think there aren't Chinese or Russian mathematicians, cryptanalysts, and engineers poring over downed airframes, captured signals intelligence, and other materiel in their client state's war with the US? We were certainly doing technical exploitation on captured Russian war systems in Ukraine. And I assure you, all *those* parties have much, much tighter security than the tech companies, because counterintelligence is not a loss leader and does not increase the market share you're going to threaten politicians and woo VC money with.