I don't think AGI is right around the corner

Apparently 54% of American adults read at or below a sixth-grade level nationwide. I’d say AGI is kinda here already.

https://en.wikipedia.org/wiki/Literacy_in_the_United_States

The ability to read is all it takes to have AGI?

Does a country failed education system has anything to do with AGI?

very cool. now let's see the LLM do the laundry and wash my dishes

yes you're free to give it a physical body in the form of a robot. i don't think that will help.

What percentage of those people could never read above a certain grade level? Could 100% of humans eventually, with infinite resources and time, all be geniuses? Could they read and comprehend all the works produced by mankind?

I'm curious.

Yet, those illiterate people can still solve enormous amounts of challenges that LLMs cannot.

No but they could probably read better. Just look at the best education systems in the world and propagate that. Generally, all countries should be able to replicate that.

The stat is skewed wildly by immigration. The literacy level of native born Americans is higher. The population of foreign born adults is nearly 20% of the total adult population, and as you can imagine many are actively learning English.

It’s not skewed much by immigration. This is because the native-born population is much larger.

See: https://www.migrationpolicy.org/sites/default/files/publicat...

51% of native-born adults scored at Level 3 or higher. This is considered the benchmark for being able to manage complex tasks and fully participate in a knowledge-based society. Only 28% of immigrant adults achieved this level. So yes immigrants are in trouble, but it’s still a huge problem with 49% native-born below Level 3.

The immigration is actually working to boost literacy levels. Americans have been falling off for a long time.

From an economics perspective, a more relevant comparison would be to the workers that a business would normally hire to do a particular job.

For example, for a copy-editing job, they probably wouldn't hire people who can't read all that well, and never mind what the national average is. Other jobs require different skills.

Life is a lot bigger than just economics.

See here for example: https://data.worldhappiness.report/chart

The US economy has never been richer, but overall happiness just keeps dropping. So they vote for populists. Do you think more AI will help?

I think it’s wiser to support improving education.

I don’t know whether it will or not. I seems like people are worrying about the economic impact of AI though?

Yes if you measure AGI against median human.

In my mind “literate” is not “hand complex tasks”, it’s a basic ability to read and write.

Seems like the standards have changed over time?

Considering the amount of digitalization in society, more government regulations, etc. I think basic literacy alone does not guarantee you can participate in society effectively.

Why do you think a US citizen represents the median human?

Let me guess, you are from USA.

I'm not.

Then why should we use Americans as the average human?

The US citizen is probably well above the median human. There are many countries where the education system is abysmal, many people are stunted because of inadequate nutrition, etc.

If it is well above it also not a good average human.

What is your point?

That in raw IQ terms LLMs already beat the median human, American or otherwise, and have thus achieved some definitions of "AGI".

That doesn't mean they're capable of completing many human tasks, much less improving themselves, which is generally considered the bar for "real" AGI/super intelligence.

But in the past, literacy was typically defined as "bare minimum".

It's fine if we want to change it to "sufficiently master language to do a white collar job", but if the standard changes we shouldn't be surprised fewer people meet it.