ChatGPT says
Here are two explanations ChatGPT gave me when I asked it to explain “intelligence.”
This first one was to be written for a ten-year-old:
Intelligence is the ability to learn, understand, and solve problems. It's like having a really powerful brain that helps you figure out things that might be tricky or difficult.
Just like how you learn new things in school or from your parents, animals also learn from their surroundings and experiences. For example, birds know how to build their nests, and monkeys know how to use tools to get food.
People are considered to be very intelligent because we can do things like learn different languages, create new things, and solve complex problems. But there are many different types of intelligence, like being good at music, sports, or art. Everyone has their own unique kind of intelligence, and that's what makes us all special and valuable in our own way.
This second one was the same request, but meant for a college student.
Intelligence is a complex and multifaceted concept that can be difficult to define. Generally, intelligence refers to the ability to acquire and apply knowledge, think abstractly and solve problems, and adapt to new situations.
Intelligence can be measured using a variety of standardized tests, such as IQ tests, which aim to assess cognitive abilities like logical reasoning, pattern recognition, and memory. However, intelligence is not just about having a high IQ score; it can also refer to other forms of intelligence, like emotional intelligence (the ability to understand and manage one's own emotions and the emotions of others), social intelligence (the ability to navigate social situations effectively), and creative intelligence (the ability to generate novel ideas and solutions).
Some theories suggest that intelligence is largely determined by genetics, while others suggest that environmental factors, such as upbringing, education, and life experiences, play a larger role. There is also ongoing debate about the extent to which intelligence can be improved through education and training.
In summary, intelligence is a complex and multifaceted concept that encompasses a range of cognitive and non-cognitive abilities, and there is ongoing scientific research to better understand its nature and determinants.
I found the output interesting—setting aside to what extent I would agree with the content. It is odd that it uses “we” when referring to “people” in the first one, implying it is one of us, a fellow human. Getting a little too chummy, there, aren’t we?
To end my session, I asked ChatGPT if it is trained by humans. The reply:
Yes, ChatGPT is a language model that was trained by humans. Specifically, it was trained using a process called "unsupervised learning," in which the model was given a massive amount of text data and learned to generate human-like language by identifying patterns and relationships within the data. The training data was pre-processed by humans, and the model was designed and developed by a team of researchers and engineers. However, once the training process was complete, the model became an independent machine that generates responses based on its learned patterns and does not rely on human input or supervision during the conversation.
Altogether, rather fascinating. There’s no doubt that what makes the output text so compelling is that it is what I would say is linguistically competent. As a piece of software riding on electrons in computer chips, it doesn’t “know” anything except what it’s been programmed to do, which is compile a text according to a set of language rules and based on information retrieved from a very vast database. And this it does quite convincingly, making it seem real entirely because of how we as humans receive its language output. The language follows the rules of grammar and syntax, of sentence, paragraph, and essay structure. We anthropomorphize the rest, with our brains wanting to perceive it as a sentient being, as another real person.
If you have your own questions for the bot, you can go to the site and set up a free account to play around with it.
One potential everyday use I’ve already seen mentioned—I think by Tyler Cowen—was to have ChatGPT explain theories or concepts mentioned in other discussions, sort of like you might use an encyclopedia if you could just “talk” to it. It will give you the conventional wisdom of the subject matter in a legible style. It will make for a very sophisticated internet search engine.
Of course, the output can only be as good as the input. GIGO still applies: garbage in, garbage out. Too much reliance on the AI at the start may wind up reinforcing thought conformity and group think within a given discipline. On another hand, it may help us explore ideas from far afield without having first to acquire a specialist’s expertise, but in a smarter way than available now using a plain internet search engine. Or so I think.
Note on commenting: I’ve changed the comment sort order to “New first” as the default on my end. I only recently discovered this setting. You can still select a different sorting order, but it will go back to showing the new comments first when you reload the comment page.
Help, it won't let me edit my own remarks now. Hey Substack, what the--??
Oh. So this one I can edit. Just not the others. Again, Substack, what the--???
Ahhh...so this explains why a couple of comment sections I follow automatically display "New First" when I load the page, and all the rest show "Top First', my never having seen a "Chronological First". I once spent no small amount of time quite some time ago trying to discover if there was a setting I could use on my end to make my own personal, individual 'preference' the default. Nope.
Yet another example of the grip of illiberal, anti-democratic forces in which we find ourselves these days, exacerbated by tech and the digital age. Did anybody ask what the masses here might prefer? Was a free and fair election held to decide the matter? Again, nope.
The Big Shot just decides, with no real regard for the actual will of the people, and the rest of us just gotta' live with it, with no recourse to right this wrong. What about the Chronological People? They may be a small minority, but what about their right to choose?
Oh, sure, you say...you can still click what you want each time you load the page. But that's not *real* choice or *real* freedom, since it requires a hot second's worth of effort to get what one wants instead of having it automatically dished up on demand vis a vie the powers that be. Like all authoritarian regimes, it's only just enough to more or less sufficiently pacify the dissenters and keep them from pouring into the streets in rabid protest against the regime. But since politics often makes for strange bedfellows, I'd not be surprised if one day the Chronoligicalers and Top Firsters form a coalition and mount an effort to cast off this yoke of tyranny.
Meantime, I've long been a New Firster, since the Top Firsters are such a biased bunch in their desire to relegate the less smart, less witty and less articulate among us to a lower notch on the totem pole, requiring more effort from folks like me to find, read and engage with folks like me. And considering what I've occasionally found at the top of the heap in the way of big numbers next to those little hearts in a few places, I'm not so sure the fix isn't in on occasion.
So, I'm down with your policy decision there, Mr. Big Shot. I know deep down in my Sunday School heart of hearts it was arrived at unfairly, but deep down in my heart I'm really quite lazy. And it serves my interest. So, good on you!! From my perspective, it's all hail the BS!!
Speaking of which, I'm sort of tempted to ask that ChatGPT thingy to explain American politics. Only because I love the smell of blown fuses and melting wires in the morning. Smells like...victory.
https://www.youtube.com/watch?v=vRp7tYWnJJs