Raw LLM Responses

Software cannot be intelligent because machines cannot be suspicious. Computers are perfectly naive. This is why they will never recognize sarcasm or humor. All it takes to change the mind of a computer is to locate a sufficiently critical bit in its decision space and change it from 0 to 1 or vice-versa. Basically, you just simply tell it "no, that's wrong" and it says "Ok" and starts believing the opposite without question. There is a very good reason the access to some software is protected from people using schemes typically referred to as "trust verification". It is because software is by nature perfectly trusting of everything it is capable of decoding. It recognizes both lies and truth as truth. For critical systems it is imperative that only trustworthy individuals have access to tell it things that it will reliably accept as fact without questioning. AI is nothing more than a software development technique which leverages the computer to help program itself in order to alleviate some of the load from designers & programmers, to catch mistakes humans are more prone to make, and to discover relationships and dependencies humans overlook when designing. Additionally, it is unlike traditional software development which is 100% predetermined during design. Using AI some of that pre-planning normally carried out prior to starting development is deferred and handled by the software during its implementation. All of which produces a more robust and flexible result than would otherwise be achieved using traditional software development techniques which rely on only humans to predetermine, program & code the apps. It is cheaper & easier to maintain, upgrade, and repurpose, plus it takes less time to make it. The trade-off is that AI is indeterminate whereas traditional is 100% determinate. Whenever AI has insufficient information to come to a conclusive decision it simply makes a wild guess. Traditionally developed software will not do this, instead simply spitting out an error message and refusing to continue. If there is no human around to inform AI of the correctness of its guesses, it will forever remain confused and error prone. In fact, with AI all of its decision making is done using wild guess. Educated guess to be sure, but random. It is extraordinarily rare for AI to decide anything with 100% confidence. It chooses the most correct action based upon a computed probability of it being correct. AI is only ever probably correct, never (rarely) definitely correct. AI never invents things for itself to do. Like all software it takes requests and produces answers based on them. Given no inputs all AI becomes idle and inert just like every other app. Incapable of thinking up something to do unless asked by a person. Not artificial, not intelligence, just another app. It is computer assisted software design & maintenance. No AI ever decided to be a chat bot, they are told (designed & programmed) to operate as chat bots. No AI chatbot will ever independently decide to stop chatting and do something else instead unless told to do so by some intelligent being with enough expertise in something else to teach (design & program) the AI how to do it correctly.