Raw LLM Responses

I'm a huge Hinton fan (his academic work), but his recent stage does not seem very helpful. Yes, AI poses dangers, but what are we gonna do about it? Sit back, learn plumbing, call your local representative once day? I mean that's certainly not nothing, but why not offer an approach that leverages knowledge and research, instead of more or less idling around, waiting for doom to knock on your door? Look at his student, Ilya, who quit OpenAI and moved on to start a company all about safe AI. If you're worried about your career, AI may not not take programming jobs for 20 more years. I do both AI and software engineering, but anyone that knows a little about code knows that current AI is very far from taking any jobs. And agentic frameworks look promising, but to me it looks like this in need of mostly human intervention, so progress there may be in the order of decades. I can only recommend to keep studying computer science, math, and understand the "threat" that is coming. To me, this is a no-brainer. Plenty people worried about not having jobs, and plenty people worried about AI as a threat. Software engineering may no longer be a high paying job in the future, but it needs more people. And it especially needs people who *know* what they're doing, rather than going through their entire education by having AI do their chores, and them not learning anything beyond prompt engineering. The world will be chock full of such people, who got their degrees without effort, but also don't have skills beyond prompting. And if you can't afford a degree, then just teach yourself. It's never been easier than now, where AI can teach it all to you. What matters is what you ask the right questions and pay attention, and never blindly trust anything the AI says. If you're not willing to touch math with a ten foot pole, then you could still go into red-teaming AI. Learn how to hack it, make it your career to exposes its weaknesses. There will be no shortage of that in the future. You don't even need to be employed, there are bounty programs for ethical hacking. The worst thing you can do is let yourself be depressed by the potential of a bad future that may not even happen. Remember that Hinton's position is not that AI is likely to be a detriment to humanity, but that this is a possibility that shouldn't be ignored. It may still turn out largely positive. How this goes is influenced by CEOs and politicians, sure, but certainly also in part by AI engineers and security experts of the future.