"Add economic value" is the crux of the problem, don't you see?
If it increases the GDP by 5%, but the result is a massive redistribution of wealth to the infrastructure holders at the expense of millions of people losing their livelyhood is that worth it?
You don't have to answer, most of HN would say yes because we (including me) are part of the class of people that will also benefit as society's demand for technologists continues strong.
And there is also the argument that all of this has happened before with every technical revolution - from the cotton gin to the car to Excel.
So this is where the question of LLM's true capabilities becomes essential. Because if everyone approached them cautiously and adopted use cases gradually and increased individual people's productivity, before realizing that in time whole jobs would cease to be essential - no problem.
But instead - and we've already seen this happen with ML and customer service - whole industries are eager to absolutely decimate their fixed human costs - and all their flaky human problems - to replace them with LLMs, and let the chips fall where they may.
THIS is why so many people are pushing back to challenge the true utility of LLMs and not over-inflate their capabilities. Because we can see the inevitable wave that will make society and technology worse for consumers, and worse for most of the employees in the short term. It will be better for the "Economy" (as measured solely by GDP), and will make some new millionaires and double up some billionaires.
But my (and others') claim is that it will not create a better world for humanity. And it may decimate artistic creativity for a generation before we recover.
Parents already see it as terrifying to figure out how to teach children to handle all the HUMAN-generated misinformation over the last decade. It's only going to get worse when it becomes machine-generated.
> If it increases the GDP by 5%, but the result is a massive redistribution of wealth to the infrastructure holders at the expense of millions of people losing their livelyhood is that worth it?
Yes. It would be totally worth it. I am fine if my job gets 100% replaced by an LLM.
I do not know if your opinions are a good representation of what the critics have been saying but if it is a good sample then I would say such critics have very poor grasp of basic economics or how the world works.
If the infrastructure holders get very rich very fast it is not a bad thing, just invest in their stocks and you can get very rich too. It is that simple.
But what you need to ask is what exactly people do when they get very rich ? They just invest it into other ventures that can make them very rich. LLMs were result of few people having lot of money to spend. GPT4 is supposedly cost $100M to train. It is $100M capital that created it. Imagine what $100B can create if people have that much money.
If AI destroys entire sectors of jobs that would be a noble goal. It means human beings wont have to waste their time on problems that are worthless like how we look at folks who moved from toilet to toilet collecting human waste to dispose it off.
I would also like to point out that his change is "inevitable". If US tries to regulate it, the AI companies will move to China, Russia, North Korea or Saudi. Imagine a nation like North Korea with no real human capital suddenly producing AI generated content, code and software that Americans end up buying using VPN.
> If the infrastructure holders get very rich very fast it is not a bad thing, just invest in their stocks and you can get very rich too. It is that simple.
I don't think you can get to criticize anyone for having a poor grasp of basic economics.
Otherwise it just sounds like you're saying that there shouldn't be a need for anyone to anyone to be poor - they simply need to invest in the stock market.
> Imagine what $100B can create if people have that much money.
I don't need to imagine. I see what Elon Musk did with $44B, which is now worth ~$15B.
> If AI destroys entire sectors of jobs that would be a noble goal. It means human beings wont have to waste their time on problems that are worthless like how we look at folks who moved from toilet to toilet collecting human waste to dispose it off.
I broadly agree with that, except you skipped my entire point where I'm clarifying that I'm not concerned about the menial jobs, but rather the ones that are about to (already started actually) be replaced by INFERIOR versions of product quality. Modern AI/LLM/Generative AI tools are not replacing toilet collectors. They're replacing journalists, writers, artists, and musicians.
We're not replacing burdensome labour work, we're replacing the most uniquely human tasks of creativity.
If it increases the GDP by 5%, but the result is a massive redistribution of wealth to the infrastructure holders at the expense of millions of people losing their livelyhood is that worth it?
You don't have to answer, most of HN would say yes because we (including me) are part of the class of people that will also benefit as society's demand for technologists continues strong.
And there is also the argument that all of this has happened before with every technical revolution - from the cotton gin to the car to Excel.
So this is where the question of LLM's true capabilities becomes essential. Because if everyone approached them cautiously and adopted use cases gradually and increased individual people's productivity, before realizing that in time whole jobs would cease to be essential - no problem.
But instead - and we've already seen this happen with ML and customer service - whole industries are eager to absolutely decimate their fixed human costs - and all their flaky human problems - to replace them with LLMs, and let the chips fall where they may.
THIS is why so many people are pushing back to challenge the true utility of LLMs and not over-inflate their capabilities. Because we can see the inevitable wave that will make society and technology worse for consumers, and worse for most of the employees in the short term. It will be better for the "Economy" (as measured solely by GDP), and will make some new millionaires and double up some billionaires.
But my (and others') claim is that it will not create a better world for humanity. And it may decimate artistic creativity for a generation before we recover.
Parents already see it as terrifying to figure out how to teach children to handle all the HUMAN-generated misinformation over the last decade. It's only going to get worse when it becomes machine-generated.