If AI works you'll be able to make more stuff with fewer people. While that could lead to unemployment historically it's not gone that way. You get more stuff.
Like agricultural employment has gone from ~70% to ~2% but the people who would have dug potatoes make cars, houses, aircraft and the like.
That's quite an unsubstantiated leap. The world has gone through plenty of digital transformations and the number of people in poverty has only _shrank_.
It's hard not to make that leap when so many layoffs are (according to PR releases anyway) attributed to AI adoption. Even if the reality on the ground is that many of these workforce reductions are to make the balance sheets look better (presumably as a bet on AI), it's impossible to ignore the accelerating wealth gap, especially in the context of the gutting of regulations and state actors leveraging world events on prediction markets. We will not be given a fair deal if we simply wait for our benefactors to provide one.
Yes, but during those transformations, the CEOs of the companies selling the products involved weren't actively and aggressively marketing them as being able to replace all the humans they employ.
You can't have it both ways: either LLMs are an amazing, revolutionary technology that can replace many human jobs in unprecedented ways, or it's going to be a mild transition that really only helps people.
> the CEOs of the companies selling the products involved weren't actively and aggressively marketing them as being able to replace all the humans they employ
The assembly line was explicitly about replacing skilled with relatively unskilled labor.
That was exactly what a great many things were marketed as, such as the jacquard loom and dynamite.
What actually happened in each case was that employment went up for a good long while, as the efficiency boost to the sectors touched made investment far more viable. Eventually successive rounds of automation did reduce employment in each of weaving and mining, but it wasn’t an overnight catastrophe as initially advertised or feared.
It isn't the first time a new technology has been pitched to replace many worker's jobs, both successful and unsuccessful versions of the promise have come to pass several times.
I think what they are saying is "that something can replace a job does not inherently imply the next step is poverty". From that perspective, you can absolutely have it both (and many other combinations of) ways.
The number of people in absolute poverty has shrunk, but the proportion of national income held by the wealthy has increased, so economic mobility is declining. There are many reasons for this, but typically deployment of technology is a capital expense and employers aim to realize all the gains from their investment, notwithstanding the upskilling and/or deskilling effect it has on workers, who are treated as fungible economic units rather than people. Nobody likes this except capitalists.
Unemployment rampant. All production remains in the hands of a few. All power (tokens) remains in the hands of a few. Goods are cheaper but no one can buy them. Path to the upper class now guarded closely by tokens, potential avenues for entrepreneurs diminish rapidly. Own an AI or compute, get someone to give you tokens, or live in poverty.
Distribution of abundance in current time is close to evil, America reducing entitlements and support (not expanding). Rampant waste. No reason to think any of this will change.
I'm confused. It seems the parent comment is saying AI proliferation could make cost of goods drop orders of magnitude and you say it's detached from reality because people don't want goods, they want housing food and gas?
> Cost of goods and services drops by orders of magnitude at every point in the supply chain.
That sounds great, but how are LLMs supposed to achieve this? You can't just say "AI will make a utopia". You have to present a vision for how it will get us there.
I'm tired of hearing about how AI will solve all the worlds problems. I want to see actual progress towards achieving these goals. And for the most part that hasn't manifested. Most people would consider AI to have had a net negative impact on their lives.
How? The biggest cost of most products comes down to energy cost and the profit margins of each proccess and middleman. Actual labor costs are already a pretty small portion of most products and even if you mine and smelt twice as much material per worker with AI somehow, that is at best a few percentage off the final price. And adding in AI processing isn't going to reduce energy costs or increases wages.
> would 100% expect a commencement speaker to be hyping me up
That’s what this speaker was trying to do. The problem is it was stupid and dishonest. It could have been done properly. But none of that will rise to the level of a roadmap. If you’re looking for a roadmap at commencement, you were failed at multiple steps before.
The funny thing is that it's not even true. People invested in AI just glee at the thought of common men in abject poverty, so this is the marketing that stuck.
Shows you don't need to have red skin and horns to delight in the suffering of starving people.
If AI works you'll be able to make more stuff with fewer people. While that could lead to unemployment historically it's not gone that way. You get more stuff.
Like agricultural employment has gone from ~70% to ~2% but the people who would have dug potatoes make cars, houses, aircraft and the like.
The problem isn't AI, the problem is a generation addicted to social media that sells their attention to the highest bidder.
How expensive do you think it would be to convince 30 million people of something that wasn't true?
Who might benefit from a generation of Americans being pessimistic about their future?
More specifically, who might benefit from a generation of Americans being anti-AI?
How would the cost compare to the potential benefits?
That's quite an unsubstantiated leap. The world has gone through plenty of digital transformations and the number of people in poverty has only _shrank_.
It's hard not to make that leap when so many layoffs are (according to PR releases anyway) attributed to AI adoption. Even if the reality on the ground is that many of these workforce reductions are to make the balance sheets look better (presumably as a bet on AI), it's impossible to ignore the accelerating wealth gap, especially in the context of the gutting of regulations and state actors leveraging world events on prediction markets. We will not be given a fair deal if we simply wait for our benefactors to provide one.
Doom is popular right now. People want to feel terrible and pessimistic and their media diets only reinforce this.
To be fair, the labour market for a recent humanities graduate is quite crap right now, too.
3 replies →
To be fair the last 60 years have seen wealth distribution consolidate at the top of this country at rates never seen in human history.
Yes, but during those transformations, the CEOs of the companies selling the products involved weren't actively and aggressively marketing them as being able to replace all the humans they employ.
You can't have it both ways: either LLMs are an amazing, revolutionary technology that can replace many human jobs in unprecedented ways, or it's going to be a mild transition that really only helps people.
> the CEOs of the companies selling the products involved weren't actively and aggressively marketing them as being able to replace all the humans they employ
The assembly line was explicitly about replacing skilled with relatively unskilled labor.
That was exactly what a great many things were marketed as, such as the jacquard loom and dynamite.
What actually happened in each case was that employment went up for a good long while, as the efficiency boost to the sectors touched made investment far more viable. Eventually successive rounds of automation did reduce employment in each of weaving and mining, but it wasn’t an overnight catastrophe as initially advertised or feared.
It isn't the first time a new technology has been pitched to replace many worker's jobs, both successful and unsuccessful versions of the promise have come to pass several times.
I think what they are saying is "that something can replace a job does not inherently imply the next step is poverty". From that perspective, you can absolutely have it both (and many other combinations of) ways.
[flagged]
16 replies →
The number of people in absolute poverty has shrunk, but the proportion of national income held by the wealthy has increased, so economic mobility is declining. There are many reasons for this, but typically deployment of technology is a capital expense and employers aim to realize all the gains from their investment, notwithstanding the upskilling and/or deskilling effect it has on workers, who are treated as fungible economic units rather than people. Nobody likes this except capitalists.
Then it should easy to show a world where we are all not in abject poverty. We’re waiting.
Cost of goods and services drops by orders of magnitude at every point in the supply chain.
That being said we already have relative superabundance and we're more miserable than ever, so it's not clear that more of it will cheer us up.
Unemployment rampant. All production remains in the hands of a few. All power (tokens) remains in the hands of a few. Goods are cheaper but no one can buy them. Path to the upper class now guarded closely by tokens, potential avenues for entrepreneurs diminish rapidly. Own an AI or compute, get someone to give you tokens, or live in poverty.
Distribution of abundance in current time is close to evil, America reducing entitlements and support (not expanding). Rampant waste. No reason to think any of this will change.
I'm voting for your future.
This is the kind of commentary that is completely detached from reality: people want housing, people want food, people want gas.
It's not great that we can buy iphones (and AI is going to cause all electronics to be scarce, so much for abundance there)
I'm confused. It seems the parent comment is saying AI proliferation could make cost of goods drop orders of magnitude and you say it's detached from reality because people don't want goods, they want housing food and gas?
Housing food and gas are goods...
Or did you mean something completely different?
5 replies →
> Cost of goods and services drops by orders of magnitude at every point in the supply chain.
That sounds great, but how are LLMs supposed to achieve this? You can't just say "AI will make a utopia". You have to present a vision for how it will get us there.
I'm tired of hearing about how AI will solve all the worlds problems. I want to see actual progress towards achieving these goals. And for the most part that hasn't manifested. Most people would consider AI to have had a net negative impact on their lives.
How? The biggest cost of most products comes down to energy cost and the profit margins of each proccess and middleman. Actual labor costs are already a pretty small portion of most products and even if you mine and smelt twice as much material per worker with AI somehow, that is at best a few percentage off the final price. And adding in AI processing isn't going to reduce energy costs or increases wages.
> show them a future that doesn't leave them in abject poverty
To be fair, this isn’t the commencement speaker’s job.
If this is sarcasm, it's amazing. If not, depressing.
I would 100% expect a commencement speaker to be hyping me up for what comes next.
> would 100% expect a commencement speaker to be hyping me up
That’s what this speaker was trying to do. The problem is it was stupid and dishonest. It could have been done properly. But none of that will rise to the level of a roadmap. If you’re looking for a roadmap at commencement, you were failed at multiple steps before.
[flagged]
For someone taking about basic logic, you're making quite the leap in logic by assuming they used LLMs to do every single bit of schoolwork.
[dead]
> You can't have it both ways.
Yes you can. They use AI and also despise it because it will turn the world into one big caste system. Ones with access to compute, and ones without.
The funny thing is that it's not even true. People invested in AI just glee at the thought of common men in abject poverty, so this is the marketing that stuck.
Shows you don't need to have red skin and horns to delight in the suffering of starving people.