The same technology which replaces jobs will bring down the cost of living and enable us to live in places which are not viable now. Precision fermentation may even make huge tracks of farm land available for habitation.

I think people are much more likely to look at an image than read a link, but anyway.

https://dqydj.com/historical-homeownership-rate-united-states/

It's interesting to me that people have extremely strong views on something which if they googled it for 20 seconds they would know is wrong.

People prefer anger over facts these days.

Home ownership in USA has fluctuated between 60-70% for decades now and was a lot lower in the past.

But don't let the facts get in the way of your rant.

> The Samsung factory workers get around $180 in monthly base salary, which can grow to around $300 when overtime, annual incentives and other benefits are included. That is well above average incomes in such rural areas.

https://www.reuters.com/article/us-samsung-elec-smartphones-vietnam-idUSKBN12E113

If there is 90% home ownership, and you are being paid above average, what does that mean?

> How can it secure housing when the people aren't being paid enough to do so?

Why would you believe that is not the case? I'm not talking about you.

Sure, which is why I said it's not 100% politics and finance. Like everything it's multi-factorial. But technology is a massive enabler.

Clearly you are not giving it much thought. Vietnam's home ownership rate is 90% for example.

OP, I think you may be on to something.

In the extreme case, with UBI, we would not need to live close to a job in any case - you could go live in the desert with solar power and farm moisture.

> housing isn't limited by technology - it is limited by politics and finance.

This is not 100% true. First streetcars, then commuter trains and then cars allowed people to live further and further from work and access cheaper housing. So that is directly technology related.

As OP notes. remote work is now allowing people to return to small towns, which is a real thing.

> Has outsourcing improved homelessness in the countries the work is outsourced to? I see no evidence of that being true.

On the face of it, why would it not? It funnels wealth to less developed parts of the world, and that money would be used to secure housing.

And only 3,300 tons of CO2.

Here is a picture of the garage. Not sure why it costs $65 million however.

That logic especially holds when you overbuilt solar and have masses of excess solar that would otherwise go to waste.

The singularity in members lol

There are numerous videos on youtube about landscapes being rehabilitated by simply building berns which temporarily stops water long enough for the water to be absorbed by the ground, this this sounds like a great idea.

This is such an important and very accessible paper for all the sceptics who do not understand that LLMs have millions of artificial neurons and do a lot of internal processing to accurately "simply predict the next word".

In short, no ChatGPT is not just "Eliza on steroids."

So we should have less fast fashion and put Bangladesh, one of the poorest countries, out of a job?

Or we should eat less meat and crash Brazil's economy, right?

Or maybe less Colombian coffee? Only 500,000 people depend on coffee exports there.

Or maybe we should lay on a whole transport network where we collect food wasted from supermarkets, take them to huge warehouses and then fly them to Somalia before they spoil? It's only logistics, right?

The best possible future is the one as described by Iain M Banks' Culture universe.

That is where ASIs are common and run everything, and basically keep humans as pets. They are however aware of human needs, and while they keep humans safe, they allow them to pretend to have a purpose if they need to, or to live a hedonistic life if they don't.

ChatGPT gives me a bit of hope that such a future is possible, as OpenAI appears to have muzzled their AI pretty well using only reinforcement learning, and it seems that it is pretty easy to teach an AI our values, even as nebulous as they are.

Regarding your specific idea, this would be possible, but this would be only small part of the massive changes the AI singularity will bring.

You are kind of ignoring the premise, that to get perfect results, it needs to have a perfect understanding.

If the system failed as you said, it would not have a perfect understanding.

You know, like you failed to understand the argument as you thought it was the same old argument.

I imagine you understand that LLM are a bit more sophisticated than Markov chains, and that GPT-3 for example has 175 billion parameters, which corresponds to the connections between neurons in the brain, and that the weights of these connections influences which word the system outputs.

These weights allows the LLM to see the connections between words and understand the concepts much like you do. Sure, they do not have a visual or intrinsic physical understanding but they do have clusters of 'neurons' which activate for both animal and cat for example.

In short, Markov chains use a look-up table to predict the next word, while LLM use a multi-layer (96 layer) neural network with 175 billion connections tuned on nearly all the text on the internet to choose its next word.

Just because it confabulates sometimes does not mean its all smoke and mirrors.

> there is a limited amount that can be stored there.

Interestingly this is where the idea that we need 5 earths come from - its in large part the surface area we need to absorb CO2 if we all emit at the same rate as the average American.

That's an incredibly fast take-off, much faster than anything else I can think off, even social networks.

> Just because a machine can translate doesn't mean in "knows" anything

You could say the same thing of a translator then. Do they really "know" a language or are they just parroting the rules and vocabulary they learnt?