The current AI boom feels different from previous advancements in technology and makes me very skeptical of the future. Like way back then we used to have icemen, ppl who would deliver large blocks of ice daily. Then we created refrigerators that made icemen obsolete, but we created even more jobs for the economy like refrigerator factory workers and appliance repair technicians.
AI has been taking away jobs via automation so you would think that means we need more people in tech to push AI forward, but it's the opposite. AI is even replacing the tech workers, especially the junior ones and now the new grad from Harvard can't even find a tech job easily. Idk what's gonna happen now since so many companies no longer want jrs since they feel like AI coding assistants like copilot basically replace them. Experienced engineers are doing better, but technology moves so fast and you need to double down on being an expert in AI to even start competing for jobs now. Not to mention the rise in vibe-coding via tools like replit that allows non-tech people to build apps via text prompts only (although the code is crap and requires rework to be production ready). I'm working an AI job now, but my passion is gone. Only people who live and breath tech can survive in this field for long.
We didn't have just icemen. In the colonial age, we had people like the Boston tycoon who filled large ships with ice and lugged it to all the European colonies in tropical Asia, South America and the islands. But, that is how technology works. It's purpose is to fulfill needs. Once some needs are fulfilled, we just find new needs. Eventually, humans will be like the elves of LOTR - infinitely long-lived and dying only if violently killed. Which will force space exploration and colonization. This is all well known and inevitable.
So, I am not troubled by "AI taking away jobs" as the media puts it. It has to happen. And this has happened over and over in human civilization. As humans have always adapted, we must again. Like there were silent movie stars; then came the talkies. Most adapted. Only some faded away. Humans will just find new industries where humans are a must.
I don't think that is the problem with AI. The problem with AI is that what is now sold as AI is not really AI. I come from a background in cognitive science among other things. What they are doing now, is not the purpose for which the AI field was even created. They haven't solved the hard problem of consciousness. They haven't figured out human thinking to make machines actually think. Theory of AI, where the neuro, cognitive, philosophy and logic and the computing types all work together, has a long way to go. All they are doing is machine learning, or applied statistics. Finding repeated patterns and hoping to extrapolate a conclusion. The other thing that's happened is, they have gotten better at processing natural language. This is all pretty much faking intelligence and not simulating it. Whenever an app or a bot does something cute, it's like the Seinfeld episode where Elaine admitted to faking it and went "Fake!" "Fake!" "Fake!" Fake!" They just can't sell a lot of crap labeled "Statistics Inside", so they sex it up and say, "AI Inside". We can see the limitation like this: Let's say you are Beethoven. You have just completed the Fifth and are sitting down to compose the Sixth. You would try to go in much a novel way as possible, based on your composing brief. But, if there were an to mimic Beethoven. It's totally and precisely modeled the Fifth. The app then tries to produce the Sixth. It's necessarily going to derive it from the Fifth, or it would be a variation. It is not going to magically figure how Beethoven might approach a Sixth from a very different creative direction. Or, it can't model Beethoven but only specifically model Beethoven's process for the Fifth alone. But, the task of AI is to do the latter.
I don't know... I have been around for donkey's years and seen the field of AI start from virtuous roots in Philosophy and then take a couple of bad turns and turn into this data-interpretation-extrapolation mechanism.
I actually got recruited into the Ivy League quite young, without ever applying. But, I was around when they started changing what they wanted to see in AI....
It started with some MIT folks (new retired) try some redefinitions and claim "victory", then their counterparts in Stanford who were actually getting somewhere in doing it right also turned out... two blinks and two decades later... here we are.