Raw LLM Responses

Right now, we're using AI as a means to convert information, largely to allow humans to communicate more effectively. The path looks like this: \`Human has idea -> asks AI to develop it -> asks AI to make it presentable to another human -> presents to other human -> other human makes decision\` Soon, the person receiving the information will usually use AI to interpret it. \`... -> presents to other human -> other human uses AI to change it and support decision -> human makes decision\` Then, the other human will delegate the decision. And already, the original human used AI to get the idea. So, we can easily end up with \`AI recommends idea -> human askes AI to develop and make presentable to human -> human presents to other human -> other human uses AI to recommend decision -> other human uses AI's choice\` Why bother with the humans at all. It's just slowing things down and costing money? \`AI recommends idea\` -> \`AI communicates idea in it's own code\` -> \`other AI makes decision\`