Raw LLM Responses

These types of publications make me laugh when you consider how far we actually are from any resemblance to functional AI. For example look at military AI. You program a drone to eliminate hostiles. Ok, so who are the hostiles. Well let's RFID tag our troops so it doesn't target them. Now let's feed it 10,000 pictures of what a human can look like from various aerial angles and program it to constantly match its camera feed to this database. Now define it only one specific geographic area (aka the battlefield). Program any match of greater than 90% resemblance to human to be a target and engage. This is what AI is so far. This AI is not capable of any independent thought. It can't make a moral decision, it can't act on its own authority and it can't function outside of its own script. Even random bugs would only result in the program maybe targeting all humans. But its not capable of thinking 'hmm, maybe if I develop a factory and produce 10 more of my own drones, I can act from multiple angles.' True AI needs to capable to learning information that it is not aware of in the current state and then input that in consideration for future decisions. A bulk of AI in modern functions basically works on "let's feed it a ton a raw data and then let it draw a correlation from that to apply to a new source". This is essentially just a script drawing probable "successful" pathways (no better then just randomly picking) and masking it as "intelligence". A human is far better at this job and and the only real competition AI provides is on just how much more information it absorb. The day real AI is created, we wouldn't have a say in who gets the final decision as it would be capable on independent thought, giving no consideration to what rules humans impose on it.