3 ms·

We've reached a point where even with just continued incremental improvements, these LLMs are going to be extremely disruptive to society, probably within a dec

We've reached a point where even with just continued incremental improvements, these LLMs are going to be extremely disruptive to society, probably within a decade, to the point that we see even more massive disparities in wealth and income, massive increases in poverty and homeless, and ultimately probably large-scale civil unrest. Rather than telling yourself (like this author) that "it's just a parrot," or alternately (like I've seen so many happy-go-lucky types on HN do) telling yourself that "LLMs are just another tool that will make me more productive!" I'd recommend preparing for the worst, while hoping for the best.

massysett 4y ago
I don't understand predictions like this. Even if I thought this technology will be as useful as it's hyped up to be (which I don't) I don't see how it would have effects different from those of previous revolutionary technologies that have improved the human condition. I think the doomsayers on these LLMs would be the same folks predicting doom when agriculture was mechanized. When massive portions of the workforce toiled in agriculture, folks thought farm machines would lead to mass unemployment and social horror. Instead, world hunger has dropped, and massive portions of the workforce were freed up to do things like work on LLMs. Maybe the hand-wringing is different this time because privileged classes feel threatened.
- 876978095789789 4y ago
  > I don't see how it would have effects different from those of previous revolutionary technologies that have improved the human condition. It will greatly reduce the demand for labor in many fields (off the top of my head: customer support, artists, any kind of writing or clerical work, and probably programmers and IT pros in general), and the benefit of this productivity gain will largely be reaped by the owners of LLMs and the companies using them in the form of higher profit margins. It's not at all clear that it will trickle down to the displaced workers. > privileged classes feel threatened Ironically, I actually believe a lot of the overly optimistic takes come from a point of privilege. Think the FAANG types making $250k a year, for whom everything in life has more or less worked out well, and who have difficulty envisioning a future where things don't just keep on working out well.
rep_lodsb 4y ago
Cynical thought: "it IS just a parrot, but so are a large percentage of humanity, particularly those who are uncritically enthusiastic about any so-called progress" AI might turn out to be a disaster not because it is smarter than us, but because stupid people will believe it is and come to rely on it, even more so than they already do on Google or Wikipedia. What really strikes me about the ChatGPT output I have seen is the tone of complete confidence with which it delivers answers (except when it is programmed to respond "as a language model I cannot blah blah blah"). For me that is something which provokes an instant negative emotional reaction to what is said, because things rarely are that simple and clear cut, and any person or machine pretending that they are seems likely to me to be either simple-minded or trying to con me / advance some agenda. That most people apparently don't have that reaction to it is frightening to me.

massysett 4y ago

I don't understand predictions like this. Even if I thought this technology will be as useful as it's hyped up to be (which I don't) I don't see how it would have effects different from those of previous revolutionary technologies that have improved the human condition. I think the doomsayers on these LLMs would be the same folks predicting doom when agriculture was mechanized. When massive portions of the workforce toiled in agriculture, folks thought farm machines would lead to mass unemployment and social horror. Instead, world hunger has dropped, and massive portions of the workforce were freed up to do things like work on LLMs. Maybe the hand-wringing is different this time because privileged classes feel threatened.

876978095789789 4y ago
> I don't see how it would have effects different from those of previous revolutionary technologies that have improved the human condition. It will greatly reduce the demand for labor in many fields (off the top of my head: customer support, artists, any kind of writing or clerical work, and probably programmers and IT pros in general), and the benefit of this productivity gain will largely be reaped by the owners of LLMs and the companies using them in the form of higher profit margins. It's not at all clear that it will trickle down to the displaced workers. > privileged classes feel threatened Ironically, I actually believe a lot of the overly optimistic takes come from a point of privilege. Think the FAANG types making $250k a year, for whom everything in life has more or less worked out well, and who have difficulty envisioning a future where things don't just keep on working out well.

rep_lodsb 4y ago

Cynical thought: "it IS just a parrot, but so are a large percentage of humanity, particularly those who are uncritically enthusiastic about any so-called progress" AI might turn out to be a disaster not because it is smarter than us, but because stupid people will believe it is and come to rely on it, even more so than they already do on Google or Wikipedia. What really strikes me about the ChatGPT output I have seen is the tone of complete confidence with which it delivers answers (except when it is programmed to respond "as a language model I cannot blah blah blah"). For me that is something which provokes an instant negative emotional reaction to what is said, because things rarely are that simple and clear cut, and any person or machine pretending that they are seems likely to me to be either simple-minded or trying to con me / advance some agenda. That most people apparently don't have that reaction to it is frightening to me.