ChatGPT's Rational Journey: Exploring AI's Evolution in Reasoning
06:49 Share

We all know that ChatGPT is capable of coming up with some pretty complex pieces of content, right? IT can write about virtually anything you ask IT for, whether it's a hundred active or not.

But the real question that a lot of people are asking is this ChatGPT have a reasoning capability? Is this thing actually able to think? Can IT sit there in compute what it's actually saying? Or is IT just kind of thrown out words on the podcast? They really dive into that, really dive into how ChatGPT works. I'm and answer this important question.

So first off, it's important to know that ChatGPT the GPT stands for general pretrail transformer and then breaking that down further will tell about the transformer I transformer is something that was invented by um a group of researchers at google back in twenty seventeen is kind of funny that this technology was invented by google um and the transformer is essentially a model that is able to accurately put out really um powerful pieces of content that are is able to write natural language and full sentences. So first, I think if we want to understand how this works, we need to kind of define what we mean by reasoning. So in general, reason, reasoning refers to the ability to use logic and critical thinking.

I'm a more analyzing information and then from that to be objected drug conclusions or to, you know, like solve problems. So IT involves making connections between different pieces of information. So I essentially you identifying patterns and relationships and then you're using these insights you get to make um your decisions that should be informed, right? So reasoning is pretty critical um it's a pretty critical obviously aspect of human intelligence and it's one of the main features that I would say distinguish humans from animals first, for example. So when IT comes to ChatGPT, I feel like this question is pretty important um because the question of reasoning is is you know its capability is a pretty complex topic um so on the one hand, the model has obviously demonstrated a very impressive ability to generate text.

You know we all see that every day when we put questions in and get outputs um and so that they appear to be reasoned in logical right like I mean i've asked that they write me an essay or an article and IT IT would look like IT came out with a logically written article going from point a to point b and kind of coming up with arguments Y X Y N Z is possible. So for example, when asked a question, ChatGPT um can often provide a well reasoned answer that takes into account relevant information and provides a clear explanation of its thinking. right. I could also say so that's the one side of IT. Um and then on the other side, you know there's a lot of researchers that have argued that chat G P apparent reasoning ability is largely a result of its statistical processing power rather than to reasoning.

So this is kind of getting down to the question and it's kind of interesting um because you know a while ago in the news there is a google researcher tested out google chat by and he was say, you know it's sentient because um IT was you know saying all sorts of things to him and you know using the something like chat, we can kind of start to understand because back the time the technology wasn't actually released to the public, we can started understand why he might say IT sent and right like you can ask IT to say anything or act anyway and I could do IT but is IT actually sent in? Is that actually reasoning? Those are bigger questions.

So to answer the monthly tragedy, t is essentially trained on a really vast amount of text data um and IT uses you know responses that are statistically likely to be accurate. So in other words, it's able to generate text that appears to be reasoned in logical because it's been trained on a large amount of text that has already been you like someone wrote this the text logically and maybe using reason when they did IT. So does not necessarily mean it's doing the thinking um IT just means that it's able to IT was trained on you know data where someone had to do thinking.

And now when IT replicates that content or that data, IT looks like IT has been doing the thinking. So I guess to Better understand issue, I would say um we could look at couple examples of of chat g pities reasoning capabilities. So one area where chat chip tee has been particularly, I would say impressive is in answering in trivia questions, right? The thing was trained on the whole internet so shouldn't know answers to strive questions.

So if you ask IT though, what's the capital of france is going to tell you, paris, along with info about the city in history that can stuff, you know. And also, if you ask that, you know, like who wented the telephone chat, pet can provide you with a correct dancer about that. So in these examples, IT might seem like charge pet is demonstrated like reasoning ability, but all it's really doing is taking info about a topic, processing that information in using IT to generate a well recent response, pretty much pulling IT out of its database of, you know what IT already knows.

The answer is it's getting the information am and is able to produce that. So I would say you know for looking at some areas where ChatGPT falls short. Um you know let's say we asked ChatGPT if a train leaves the station a ten A M and it's traveling in sixty miles an hour in the a train leave station b at eleven P M traveling seven miles right like like my invention is like, what time do the two trains? Meat is like a classical math problem that just essentially requires some logical thinking to solve.

ChatGPT is likely to struggle pretty hard with this question and a lot of people you shared screen shots about like why is ChatGPT so bad at math, even some basic stuff um in its because IT has not specifically been trained on math solving problems and so well might be able to generate a response that contains some relevant information. Um you might talk about the distance with the speed of the two trains, it's pretty unlikely to be able to actually reason through the problem in a way that a human would do so. Um I would say all in all, you know ChatGPT looks pretty impressive.

IT does not have reasoning capabilities IT not actually thinking about you know how to solve different problems or things you're saying it's essentially just kind of giving you back information that are been trained on. And that's not to say that in the future people aren't going to you know integrate complex math solving problem models into this or other models into this to help do that. I think when we were looking at being GPT, you know your bings chatbot that's coming out soon um mixing ChatGPT.

And I think we're going to see maybe a little bit more of a hybrid between, I wouldn't call a reasoning, but it's able to uh do A A couple of more complex things, is able to search internet live and do some things like that. I think we'll start moving in the direction um of these tools actually doing computational like reasoning, thinking about things by at the moment he does not do IT. And anyone that says that you know ChatGPT is I would say it's probably misguided at this point.