Not just writing, but reading and comprehension. LLMs are a life-changing technology for a very large proportion of the population. Tens of millions of Americans are functionally illiterate, but are forced to go through life in a world that at best wasn't designed for them and at worst is actively malicious towards them.
A lot of people will look at a credit agreement they're about to sign and have no way of knowing whether it's a fair deal or naked exploitation that will financially ruin them. Thanks to LLMs, they can now point their phone camera at it and ask "is this a fair contract, or am I being ripped off". The LLM will give you an answer in speech and happily go back-and-forth with questions and clarifications for as long as you need. Whether the LLM is 90% accurate or 99% accurate at that task isn't truly critical, because it's a vast improvement over the status quo. Modern life is full of tasks that are utterly trivial for people in the top 10%, but a fraught ordeal for people lower down in the distribution.
To paraphrase a very old adage: God created man and Sam Altman made them equal.
We should be addressing _why_ so much of our population is functionally illiterate and _why_ people can't read financial agreements that can harm them. This should definitely not be left to an extremely-profit-motivated entity to equalize.
Instead, more states are eroding public education in favor of school choice and voucher programs, which is scary in a world wherein many parents would have no issues pulling their kids out of school or sending them to religious schools that consider the Bible as the authoritative text for everything, and for-profit schools would have no problem sacrificing curriculum, learning development and staff the minute it complicates revenue.
Technology should not be fixing such a societal or legislative issue. In the example you're providing, why should the user trust the LLM, but not the credit agreement? Why shouldn't the LLM point the user to a different credit agreement, just as exploitative? The company operating the LLM may have such an incentive, and it could be very lucrative.
Illiteracy in the case you are proposing can only increase, for the benefit of some.
(The argument that people can use open source models can't possibly be applied, considering you're speaking about, as you called them, the other 90%, or even the functionally illiterate.)
The last paragraph is concerning to say the least.
Comments
Not just writing, but reading and comprehension. LLMs are a life-changing technology for a very large proportion of the population. Tens of millions of Americans are functionally illiterate, but are forced to go through life in a world that at best wasn't designed for them and at worst is actively malicious towards them.
A lot of people will look at a credit agreement they're about to sign and have no way of knowing whether it's a fair deal or naked exploitation that will financially ruin them. Thanks to LLMs, they can now point their phone camera at it and ask "is this a fair contract, or am I being ripped off". The LLM will give you an answer in speech and happily go back-and-forth with questions and clarifications for as long as you need. Whether the LLM is 90% accurate or 99% accurate at that task isn't truly critical, because it's a vast improvement over the status quo. Modern life is full of tasks that are utterly trivial for people in the top 10%, but a fraught ordeal for people lower down in the distribution.
To paraphrase a very old adage: God created man and Sam Altman made them equal.
Sorry, but this is a bad take.
We should be addressing _why_ so much of our population is functionally illiterate and _why_ people can't read financial agreements that can harm them. This should definitely not be left to an extremely-profit-motivated entity to equalize.
Instead, more states are eroding public education in favor of school choice and voucher programs, which is scary in a world wherein many parents would have no issues pulling their kids out of school or sending them to religious schools that consider the Bible as the authoritative text for everything, and for-profit schools would have no problem sacrificing curriculum, learning development and staff the minute it complicates revenue.
I wholeheartedly disagree.
Technology should not be fixing such a societal or legislative issue. In the example you're providing, why should the user trust the LLM, but not the credit agreement? Why shouldn't the LLM point the user to a different credit agreement, just as exploitative? The company operating the LLM may have such an incentive, and it could be very lucrative.
Illiteracy in the case you are proposing can only increase, for the benefit of some.
(The argument that people can use open source models can't possibly be applied, considering you're speaking about, as you called them, the other 90%, or even the functionally illiterate.)
The last paragraph is concerning to say the least.