Commercial Pilots

Sure, but it only has to be good enough to make those n journeys *for *some period of time in the unspecified future*. It doesn't matter that they can't right now, just like it didn't matter when self-driving cars were failing -- nobody cared because nobody was using them. Say it takes the next 15 years to perfect: the 5 years after that can act as the aforementioned proving ground for the concept (and 15 years is an eternity in AI). By that point, some people will be willing to take the risk, and I can easily see that snowballing. (The latter is contingent on a matter of pontification in a top-level comment I made recently [sort by latest]: there still being a travel industry or need for one in 20 years, about which I'm sceptical. My model of the world predicts a vast uptick in transport accidents due to human error because of suboptimal hiring practices, which, if it continues to come to fruition, will likely make people surprisingly willing to experiment with autonomous travel, IMO.) The last sentence is right inside my wheelhouse. Humans don't really "develop" AI in the way you conceive of it. They create a very crude model of a simulacrum of a human brain ("neural network"), give it some material, and let it do its thing. These days, models often don't even have to be overseen in a substantive sense by humans ("unsupervised learning"). I mean, they *are* sanity-checked, obviously, and many continue to undergo a bootstrapping process of supervised classification, but the developer of a large AI model understands maybe 5% (to be generous) of what their model is actually doing and how it works. The rest is the part that makes AI what it is: the ability to function in divergent ways that the programmers didn't anticipate (including truly polymorphic, non-deterministic code -- I argue ND but that's a matter of terminology and epistemology, and intense debate). A good AI model will, as in so many areas, "learn" how to deal with uncertainties ad hoc with a reaction speed and complexity of visuospatial reasoning outside the limits of the human brain. It's this ability to consistently outperform the human brain (going back 30 years in chess, more recently in Go etc.) that makes AI both so valuable and so dangerous. Governments and militaries aren't pouring all their money into AI just so that their drones can't navigate to their targets with remarkable consistency and precision. I'm evaluating and advising on many different professions here based on my own experiences with AI, computer vision, and other specialised areas of expertise. I'm also a fox-style thinker (vs hedgehog; a useful and proven cognitive model for prediction), so I know a lot about a lot of different things. All of that leaves me free to be brutally honest about everyone's fields without bias. Most people here are coming at votes and comments from the perspective of *their own* current or desired field without much understanding of the technology (mine is short: AI is going to destroy CSci and adjacent fields). I want to be clear that I despise AI and see it as one of the most plausible apocalypse scenarios we have. I gave up CSci because I fundamentally disagreed with what we were doing (now I'm a neuroscientist). I believed that all technology was inherently neutral until used; now I see that the elite class can't be trusted to use neutral but high-risk technologies ethically. All I'm trying to do is bring a realistically cynical perspective to the site so that people actively resist further AI encroachment (I don't even use ChatGPT). And hopefully help young people find employment in the most stable areas. Sadly, at this point, I'm extremely bearish on the idea that there are going to be enough jobs for more than a small subset of the population. (General advice to kids reading this: do NOT go to the most expensive universities to run up enormous bills studying something that can plausibly be automated or you'll just debt trap yourself. In so far as university is still a good idea at all at a time when the trades seem more viable than ever, only go if you absolutely need a degree for your career, e.g. law, and go for "good enough" in a cost-benefit calculation. Bear in mind that there are diminishing returns at the top end. That Harvard degree will put you in hoc for half a mil out the gate but it isn't going to increase your earnings by enough to make it pay over a 75th percentile institution when amortised over your career -- which could be cut short at any time.) I see the future being some rather dystopian AI technocracy with UBI and all kinds of strangeness, including a society consisting of two strata (excluding the perpetual elite): the upper (ironically new working) class and the UBI class (though very early signs from the UK suggest that the difference isn't going to be as large as I fear, which may be good or bad depending on where the top and bottom converge). A lot of even more qualified people than me are saying much the same things. We're going to live through the most interesting time in history -- possibly one of the hardest. My only wish is for the best of luck to all in navigating it.