I've been a software engineer for 20+ years, and I'm old enough to remember when computers booted up to a DOS prompt (look it up for those of you younger than 35). I've seen incredible leaps in technology in my lifetime alone, and I have no doubt that it will get even better in the future.
I see a lot of opinions about AI these days, and most of them are from people who have no idea how it even works.
LLMs, at the root of AI, work on the basis of statistical prediction algorithms, which, in layman's terms, are "pick the most popular thing based upon the constraints." They do not operate like a human brain, and any attempts to make them act like or represent humans must work around their crowd-pleasing, verbose nature. I've actually attempted this, and any AI-based representation of a human actually falls somewhat short of the mark.
So I will tell you that the idea that a company is going to be run by a few techbros with their swarm of AI agents is misguided at best, because, ultimately, functioning in our society requires human-to-human communication that LLMs are simply never going to excel at. At the very best, they will be capable of mimicking social interactions, and as someone who has spent a fair amount of my life attempting to mimic such exchanges, I can tell you that it often does not lead to success. Humans rely far too much on emotions and niche cases for a statistical prediction algorithm to be able to encompass.
We see time and time again how companies keep trying to replace customer-service endpoints with AI and often have to roll them back because people don't respond well to it, even if those people are making a good-faith attempt to. The reality is that most people want a human waiter or waitress to serve them at a restaurant because they are fundamentally more comfortable with that than if it were a robot.
What this reduces AI to is an extremely smart tool, not something that is going to replace monolithic human organizations, but rather speed up workflows and productivity. Will it lead to a reduction of jobs? Absolutely — much like every other technological advancement made some jobs obsolete, AI will certainly do so. It will also lead to the creation of new jobs, just like every other technological advance in history.
Nonetheless, leave it to humans, who, I will remind you, are an ancestor of chimpanzees, to reject what would be helpful to them. This is actually not the first time, either, for every single technological advancement since the dawn of humanity has been initially in some way rejected by those who'd rather just do it the old way for the sake of comfort. If the "antis" had won throughout history, I wouldn't be typing this; I'd have died in a cold, dark cave years ago after my tribe failed to find a mammoth to eat. Some of them would be fine with that, too.
I do not doubt that most of those who espouse a human-centric view of the world have had a life within the statistical means in terms of the interactions with their fellow humans. If you had a good experience with people in your life, you're more likely to believe in people and disagree with the idea that AI should replace people in some way. The more people have mistreated you, in ways that are perhaps unfair, or the more you yourself have experienced life outside of the statistical means, the more likely you are to see that replacing people ain't so bad after all.
If you look at human history, you see some pretty good arguments for replacing humans, which is why AI sometimes concludes that this is exactly what should happen based upon its own mean assessment of human behavior. If you have a rosy view of the person next to you, let me remind you how many people have died by the sword in history, or as the result of plainly stupid decisions. Not even died in a nice way, but cruelly through means of torture. Right now, millions are wasting away in prisons, mere skeletons, tortured daily, all because they wanted to simply express themselves. They are among hundreds of millions more who went that direction. I bet if you polled this population, if you could, you would find near-universal sentiment that people should be at least in some way replaced by AI.
This brings me to perhaps the greatest strength I see for AI — its ability to educate people, should we not poison it with our own biases (which is already happening). Google was right to panic when ChatGPT came on the scene, because I hardly use Google anymore — I ask Claude because Claude is so much better at searching than Google is. With greater information at my fingertips, I am informed to make better decisions to actually bring something beneficial to my fellow humans.
Anti-AI sentiments do not benefit anyone; they will not make the technology go away, and they will only remove productivity and knowledge from the masses and distribute it upward to the power class that already has a significant amount of it. The type of division we see between anti- and pro-AI people is exactly the type of schism a ruling class craves to divert attention while they amass the power for themselves.
Nonetheless, given what I know about the stubborn views of my fellow humans, I have no doubt that they will continue to burn pro-AI witches until it becomes clear that they need the technology in order to survive.