WoodChuck.ai转发了
“AI” is pattern recognition and repetition. The leep from “AI” to “AGI” (Artificial General Intelligence) is massive and may never happen. Synthesis and reason are the realm of the human brain.
Marketing & communications strategist with a global portfolio of work producing billions in revenue. List curator of over 1,000 investors worth $3 Trillion. Building a CBT-based productivity app for neurodivergents.
You want the truth? There are some of you that can't handle the truth. Because the truth is this: AI, or more appropriately, "AI," does not exist. Apple knew this, and it’s a pretty good indication of how misguided people are to dunk on a company worth trillions that got there through making thoughtful, logical decisions. Recently, Ignacio de Gregorio did a great takedown of why AI, its companies, and its proponents are lying to you. I’ll post a link to his article with the research in the comments for those with Premium Medium memberships. While tech companies want you to believe the future is here, sorry, there are no flying cars or jet packs and no actual AI. Hucksters and showmen, companies and tech bros like Sam Altman want you to believe AI is AGI, Artificial General Intelligence. The AI you're being sold is just machine learning on predictive language models, trained from mostly stolen and often flawed sources, programmed to make predictive outputs. AI company’s dirty secret they’d rather you not know about is all they need to solve this, turning AI into AGI, is to cram in more data to overcome the current massive limitations. What limitations? Current AI relies on memorization, not understanding. On top of that, it's programmed on flawed data, makes inferences based on pre-fed biases, is easily fooled by irrelevant information, and can't actively memorize anything, all hallmarks of actual intelligence. But feeding it more data, will not make it suddenly AGI, and in this way, their backs are painted against the wall, and, to some extent, people are catching on. These flaws come out en mass when giving tasks for intricate planning and reasoning, especially unfamiliar tasks that stray too much from training data, demonstrating a lack of reasons, i.e., any sense of intelligence. Remember several years ago Tesla created a demo of one of its cars driving across a city, making all turns and stops perfectly without a person inside, wowing people? Then the truth came out; the route, every turn, every stop, every detail was preprogrammed. It was even filmed several times because it kept making mistakes. It was merely following a routine fed to it. Smoke and mirrors are a trait of Tesla's CEO and his promises. I bet Elon is kicking himself he didn't slap "AI" on his flawed "Autopilot," system he’s claimed would drive from coast to coast without human intervention in 2014. Here we are, 2024, and it’s still not possible. It was hype. That’s pretty much the same promise of Sam Altman and his sycophants: the AI future is here, or will be, very soon, we promise, any day now. There is a better chance of a Tesla going coast to coast without human intervention (don’t forget charging) before AGI gets here, and both are a long way off—unless you believe tech grifters. Apple will keep deepening its LLMs and machine learning, used for years. If they slap AI on it to sell it, they’ll still know the truth. AI does not exist.