New Scientific Discovery - ChatGPT Is In Fact “Bullshit”!

[AgentPete]

I like this paper in “Ethics & Information Technology” a lot:

ChatGPT is bullshit - Ethics and Information Technology

It makes abundantly clear what many of us have known for a long time – that all the talk, hot air and excited hand-waving about “intelligence” is just grift-speak. Your computer / iPhone / whatever is no more “intelligent” with added-ChatGPT than your toaster is without it.

Actually, I’d argue that Douglas Adams’ Hitchhiker's Guide To The Galaxy is far more “intelligent” than anything that derivative ChatGPT can ever produce.

Depends on how you define “intelligence”, of course. Pretty broadly, in ChatGPT’s case.

The paper cited above makes it painfully clear that this is not real intelligence in any meaningful sense, and we are anthropomorphically wrong to use words such as “hallucinating” when the damn program is simply bullshitting.

Good for the scientists involved, I say.

 

[TimRees]

I recently watched an interview with David Bowie where he was told he was more intelligent than someone else (I can't remember who now), and his response was (not verbatim): It depends how you measure intelligence. As a creator, I just do... It comes from the heart and I don't think about it too much until i have to package it...

I found that interesting. Through my life I have learned that I process information differently to others around me and I think we are all unique in our looks and our thoughts, thus expressions... Will "individual" AI units ever be as unique as each human person is an individual?

This is a relevant interview segment from Bowie:

 

[Peyton Stafford]

Depends on how you define intelligence. An ability to understand and adapt to changing circumstances, rather than repeating strings of words that once worked. ChatGPT is Captain Queeg in the 21st century. I should add that AI is the cryptocurrency of the 2020s. And we all know about Sam Bankman-Fried.

 

[LJ Beck]

Depends on how you define intelligence.

Exactly.

There are different levels of AI. And some say "AI" has hit a wall in their level of intelligence.

I love the description in the article of these LLM programs as "Bullshit Machines" and the debate is if they're "soft bullshit machines" or "hard bullshit machines." hahaha!! That is super awesome.

@TimRees Will "individual" AI units ever be as unique as each human person is an individual? - Not for a very long while. Well, hopefully never. Other discoveries need to be made first, hardware like quantum computing, and a better understanding of what we're trying to get them to mimic... us. Our brains. We just don't know enough yet.

I thought this article was good. It goes through the levels of AI.
4 Main Types of Artificial Intelligence: Explained

One of the things it says is:
Understanding, as it's generally defined, is one of AI's huge barriers. [We have] The type of AI that can generate a masterpiece portrait still has no clue what it has painted. It can generate long essays without understanding a word of what it has said. An AI that has reached the theory of mind state would have overcome this limitation. This type of AI has yet to be developed.

 

[Pamela Jo]

Another facet of the argument.

​

·


There's an interesting meme going around where the general concept is that AI should be the thing doing the laundry and dishes, while the humans get to do the art and recreation, but instead it's the other way around. What I love about this meme is that I'm not sure people realize just how old of a problem this is for humans.
In Bertrand Russell's short essay, "In Praise of Idleness" (1935) he quickly maps out how the overall strategy of these industries on the whole was to tout an idea invented under Feudalism, which is the nobility of labor as it was espoused by people who had never experienced labor in their entire lives. They then make the argument that the new technology will make our lives easier and give us more time for other things. But then when the technology gives us more time for other things, that time gets filled with more work in a celebration of high productivity.
This all gets carried down to the birth of AI, and while it's being claimed that it will help us in every aspect of our lives, one can't deny the obvious of what it's already doing, which is taking some of the jobs most meaningful to humans, the arts, while making more time for us...to do more work. It's part of the cycle of technology where the people get shafted on their end, from the long standing justification of technological development.
Russell argues that we now (over 90 years ago) have a right to be idle. Humanity has built upon itself a lineage of technological advances that have given us great power, and yet instead of understanding what it takes to truly have meaningful lives, the demand for more work and more productivity has increased with the development of new tech.
So I think people have every right to be outraged by AI, it is the intelligent accumulation of the great long con invented by medieval royalty, that without hard work you would have no purpose, and so to be a true and good citizen you must dedicate your life to serving at the whims of the overlords, and you should be thankful because it's never been so great.

 

[Pamela Jo]

Exactly.

There are different levels of AI. And some say "AI" has hit a wall in their level of intelligence.

Having encountered "copilot" several times when searching for info it has hit me in the face like a brick pie-why this race to control what info we receive?

What I want is different references -what copilot wants too do is answer my q for me. And how many people will just accept that wo thinking? And then you have them convinced they have "Done their research."

This after decades of becoming more and more conscience that information I know to be true because I found it back in the days of going to the library to do research - is not available online. For example that the sewing machine was actually invented by a woman who solved the problem of how to hang the needle, but the patent was bought and marketed by Singer.

We are entering a new Dark Ages, as defined by the dying of the light of libraries and truth seeking from humans across the world. We will now be informed what the truth is as defined by oligarchs who control AI.

 

[AgentPete]

Brilliant stuff, PJ. I need to read Russell.

The development cost of commercial AI is proving so stratospherically high that it’s obvious where their target markets will be focused – on the most expensive and most easily replaceable human jobs. EG the average salary for a surgeon in the US is $438,111 and there are 55k+ of them in the US, so expect to be operated on by an AI system soon… cost saving $24 billion p.a. And so on, throughout the jobs market, focusing on either replacing very expensive (and skilled) people, or large numbers of grunt workers. Actual benefit to the consumer – zero. But they will sell it to us very convincingly.

I so agree re “idleness” although the very word carries with it a sense of disapproval from the work-ethic crowd. Could we not call ourselves flâneurs, instead?

 

[Sedayne]

I recently watched an interview with David Bowie where he was told he was more intelligent than someone else (I can't remember who now), and his response was (not verbatim): It depends how you measure intelligence. As a creator, I just do... It comes from the heart and I don't think about it too much until i have to package it...

I found that interesting. Through my life I have learned that I process information differently to others around me and I think we are all unique in our looks and our thoughts, thus expressions... Will "individual" AI units ever be as unique as each human person is an individual?

This is a relevant interview segment from Bowie:

I've often thought that everything went wrong after Bowie left the planet.