Raw LLM Responses

It’s a bummer that there *are* some actual explanations of AI in here, and discussion of the industry, but it’s so muddled with just, well, nonsense that this video disappoints. The response to “what about people who think it’s autocomplete?” of “well I think it’s smarter” pretty much says it all. This is just a giant opinion piece. Which I guess is fine but doesn’t belong in being presented as anything but. Philosophy at best. There’s nothing backing this up. Just breezing over “it’s been trained to get thumbs up from humans” is another disservice because that ALL IT IS. There’s *plenty* of hard data there on models doing that, and the problems those answer cause. Just because we have more computing power than we’ve ever had to steal the internet as a source and give answers based on that huge data set, doesn’t mean it understands ANYTHING. That’s just pure science fiction. It’s fine to speculate on that science fiction future, and how it might be heading there, but let’s separate speculation from things we have evidence of. And the evidence isn’t there. And it muddles the *actual* issues we should discuss to start attributing powers that it doesn’t have to it. The same human brain that looked up in the stars and linked random stars into constellations, are taking guessing software and attributing intent and direction to a thing that’s just trying to optimize to the often black box parameters given to it (which change every iteration of the “model). And no one in the current generation of AI “thought leaders” is going to get us to superintelligence; they are almost all hucksters peddling snake oil of a product in search of a need. If it actually was on the road to “superintelligence” they wouldn’t need to sell it as “cut all your staff” or “Oh, yeah we’ll allow porn AI now.” If there are two branches “crap for today” and “getting to superintelligence”: Show your work. Is *anyone* seeing that? And if “intelligence” really comes down to, “being able to regurgitate the internet back to people in a way that tricks them into believing it is thinking” then we need a new definition.