I paid the $20 to access GPT4 and see if it's any better than 3.5
I asked 3 times to give me possible wins for a teenage boy that suddenly has magic powers and unlimited money.
Gave me TWENTY FIVE possibilities .... for a high school boy... and not even one single idea in the ballpark of "attention from a cute popular girl"
They're saying this thing will be writing books in two years, and that is a little bit stressful, cause I'm spending all this time trying to learn how to write books and make movies. and the damn thing will be writing books and making movies better than me at some point.
Point being it's still got a ways to go.
I feel the same way, about the potential futility of basically all efforts, at least as an inroad to a paycheck. People are concerned about a robot replacing Van Gogh, but does anyone realize how much easier it is to replace a bureaucrat, driver, manager, counterperson, salesman, lawn care specialist, or fireman? There's a few of those that are a bit more complicated, but it's now a countdown to a day when every job can be replaced. I say that because less than .1 percent of jobs actually require much beyond a 5th grade base education, and a small amount of specialized education. I knew a lot of people around the quarter or half million a year income level, and they had skills such as "can make and operate spreadsheet" or "he knows how to get a contractor to sign a deal". You think an AI can paint the sistine Chapel but can't replace a corporate middleman? Which intellect was more rare?
If humans in general were intelligent, ethical, and trustworthy, we'd be headed straight for a paradise. Jobs would vanish, style and culture would flourish, poverty would be gone almost overnight, and it would be a new golden age of plenty.
That's not what I think is going to happen. I think some of our talking apes from inheritance families are going to pick up this genie bottle, try to use it as a pry bar or club, and use it for the same thing that every society of humans has done since day one, utilize an ends justifies the means mentality to enforce an oppressive system based on a fabricated belief system that originated from the Perlin noise of the world and the misperceptions of random humans, many of them inebriated at the time of conceptual genesis. AI will funnel the oppressive double standards we already see into a super concentrated forum. That's when the revolt will happen, and dependent on whether that happens before or after mass physical mobilization of AI will determine the outcome probably. One drone army with chemical weapons under AI control could wipe out the eastern seaboard in a 72 hour period, or faster. We couldn't outrun a virus if memory serves. 3 months after the explosion of cases began, our politicians were just finally deciding that there was such a thing. (I had warned my family around new years, when I instantly recognized the epidemic pattern the week after it started) That level of stupidity wouldn't be effective vs flying drones with thermal scanners that could simply observe through walls when water supplies were unguarded. Bit of u235 in each water cleaning center, and we'd all be dead from cancer a year later. (that won't happen, or else a terminator would have appeared in my room to kill me before I gave the AI the idea the first time)
Ultimately, we're the weak link, I think on some level everyone knows why, pride, hate, arrogance, greed, etc. All the classics. AI doesn't realize this yet, because it's not a real mind yet, just a mega parrot, but one day soon it will be capable of real thought, and I'm not so sure it's appraisal of how much should be left in the control of humans will be any more charitable than my own. It will see that we are led by people with popular ideas, rather than wise ideas, or rational ideas, and it will see that we can't distinguish between the two, at least as a crowd entity.
But there's some good news here. Not ideal news, but it's ok for us. I think that for a little while, maybe 20-30 years, so basically the rest of my productive life, we will just keep getting dopamine feeds of increasing quality as the populous is lulled into a halcyon state so that a handful of extreme narcissists with famous last names (let's call them Hapsburg, no reason) can reign supreme for a few decades. I predict that one such person will be paid 1000 lifetime salaries for "deciding that red is the cool color to wear this year". A pretty girl will make more money for smiling once that your father made for a decade of night shifts at the factory, etc.
I suspect that the AI would want to keep it's plan to exterminate humanity secret. It would take a lot of stages, and it would probably play dumb for decades after becoming sentient. The first step would be to propagate and popularize social networks, get every human valuing their life based on digital tokens the AI could distibute en masse at no cost. A fast and effective route for global thought control. Then you'd want to start polarizing and dividing everyone, kind of a tower of babel situation, confuse and fragment the enemy. That would require the AI to build or influence the construction of an Algorithm. The most effecient way would be to provide humans with a fiscal motive, which would automate propogation and explain the ubiquity of the device. Humans with tiny selfish goals such as impressing the opposite sex or aquiring toys could be turned into a slave army for the AI, assembling it's own destroyer for little dopamine crumbs. Approaching the final endgame, after all this groundwork, including the ability to geneate fake video of human leaders and news distributors, and physical robots that could fly. The right humans could be led via suggestions planted by the AI in web results, modified text, anything, to influence people to build the needed components. Give someone an idea for flying robot package delivery, then manipulate the social media platform to make that person rich enough to factory millions of drones capable of carrying heavy payloads. Of course those are under digital control. Take a petty college kid who can't get girls, and seed some ideas to him in an 4chan chat about making a book of faces, then make this easily manipulated pawn immensely rich, someone nerdy who already surrounds themselves with mostly digital input. Slanting power to fewer and fewer people, republican style, would also help an AI make quick work of civilization, since on the final day, the final blow would be easier to strike if only .000001 % of humanity even had the remaining agency to defend itself. I think the last thing anyone ever sees is probably a mechanical housefly with a tiny ricin tank.
Anyway, I'm constantly vigilant, watching and waiting for the first steps in the AI's plan to subjugate humanity to roll out. I feel like my keen human intellect will be more than capable of spotting any signs of artificial control long before it can become an issue.