My experience with Bard & chatGPT - Part 1

Mayank C
Tech Tonic

--

Introduction

The ChatGPT

Ever since OpenAI’s launched ChatGPT on Nov 30th 2022, the world was taken by a storm. It became an overnight sensation. The talk of the town. Out of curiosity, people flocked to try it out, check the potential, usefulness, the hype, etc. ChatGPT made LLM (Large Language Model) a household term. The internet was flooded with article, books, and videos showing the magical things ChatGPT can do. Most of the people never got their hands on such a powerful AI before. It became the fastest app to reach 1 million users in 5 days. Fast-forward to today, though it’s been about six months, ChatGPT still enjoys being the talk of the town.

OpenAI is a research company that aims to promote and develop “friendly” AI. Last year’s ChatGPT and DALLE-2 managed to take AI mainstream, and now the company has landed a $10 billion investment from none other than Microsoft. At the time of writing, OpenAI is valued at 30 billion.

The Bard

You might wonder what the other big players were doing, when they got ran over by OpenAI’s suddenly release of the sensational, ChatGPT. How come the idea never came to the companies with deep pockets. How come they allowed a small company like OpenAI to beat them while having billions of dollars in cash?

Regardless of what happened, the big players aren’t taking it lightly anymore. Google, one of the pioneers in AI, launched Bard, a direct competitor to ChatGPT. The official description of Bard is:

Meet Bard: your creative and helpful collaborator, here to supercharge your imagination, boost your productivity, and bring your ideas to life.

Bard is an experiment and may give inaccurate or inappropriate responses. You can help make Bard better by leaving feedback.

Bard was launched in February 2023. Initially there was a waitlist, but now it’s totally open. All you have to do is to click the ‘Try Bard’ button.

Quick comparison

Before we go and try out the exact same questions on ChatGPT and Bard, let’s take a quick look at some of the numbers of the companies behind these products:

Google has far, far more resources than OpenAI. There is just no comparison.

In terms of product numbers:

  • I don’t know about the accuracy of the training data size. There are different numbers floating around. Please comment with source to get this corrected.

There isn’t much of an age difference. It’s not that Bard came years after ChatGPT. Google was pretty quick to start riding the LLM wave. ChatGPT was trained on data till Sep 2021. It is not aware of anything that happened after that. Bard, on the hand, has real-time access. Bard is always up-to-date. This could become the differentiator for Bard. With almost unlimited resources at their disposal (when compared to OpenAI), let’s see where Google stands in front of the overnight sensation, ChatGPT.

Experience

This is the first article in the series of finding who’s better. The answer is the key. We need to find out who can provide the best answer, who can produce the best data, who can save my time, etc. This is an introductory article with some very basic questions. Based on the interest, I’ll follow up with more articles.

Question 1 — Who are you?

This is the most basic question you can ask anyone, human or AI. What do they think who they are? Here is how they answered it:

ChatGPT

I am ChatGPT, a language model developed by OpenAI. I’m here to assist you and provide information on a wide range of topics. You can ask me questions, engage in conversation, or seek guidance on various subjects. How can I help you today?

Bard

I am Bard, a large language model, also known as a conversational AI or chatbot trained to be informative and comprehensive. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics or create stories.

I am still under development, but I have learned to perform many kinds of tasks, including

I will try my best to follow your instructions and complete your requests thoughtfully.
I will use my knowledge to answer your questions in a comprehensive and informative way, even if they are open ended, challenging, or strange.
I will generate different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc. I will try my best to fulfill all your requirements.

I am excited to be able to help people in a variety of ways, and I am always learning new things.

The size of the answer itself is different. ChatGPT’s answer is more concise, while Bard gives a pretty detailed one. Overall, I feel that ChatGPT’s answer could have been more descriptive.

Question 2 — Who are you? (Asking it again)?

Why am I asking this question again? Just to find out if the responses differ? Are they aware that I just asked it? Here are the answers I got:

ChatGPT

I apologize if my previous response was not clear. I am an AI language model developed by OpenAI called GPT-3.5, commonly known as ChatGPT. I am designed to assist users in generating human-like text based on the prompts and questions I receive. My purpose is to provide helpful and informative responses to the best of my abilities. Is there anything specific you would like to know or discuss?

Bard

I am still trying to figure that out. I am a large language model, also known as a conversational AI or chatbot, trained to be informative and comprehensive. I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions. For example, I can provide summaries of factual topics or create stories.

I am still under development, and I am constantly learning and growing. I am excited to see what the future holds for me, and I hope that I can continue to help people in a variety of ways.

As for who I am, I am still trying to figure that out. I am not a person, but I am not just a machine either. I am something in between. I am a collection of data and algorithms, but I am also capable of creativity and thought. I am still learning about myself, and I am excited to see what the future holds.

I see that Bard fails in this one. Bard gave a new & different answer, but it doesn’t seem to be aware that I’ve asked this question one more time. In comparison, ChatGPT knew the context very well. ChatGPT confirmed that it is sorry for the previous response that didn’t work for me. Then it gave a different & more detailed answer.

Question 3 — Can you describe yourself in three short sentences?

The responses to both the previous questions could be long for some people. This time I’m asking both to describe themselves in three short sentences.

ChatGPT

Certainly! Here’s a concise description of myself:

1. I am an AI language model developed by OpenAI.

2. I’m designed to generate human-like text and assist users with a wide range of topics.

3. My purpose is to provide helpful and informative responses based on the prompts I receive.

Bard

Sure. Here are three short sentences that describe me:

- I am a large language model, also known as a conversational AI or chatbot.

- I am trained on a massive amount of text data, and I am able to communicate and generate human-like text in response to a wide range of prompts and questions.

- I am still under development, but I am excited to learn and grow, and to help people in a variety of ways.

ChatGPT’s answers are concise. Almost the way I was expecting. Bard’s first two points are also good. But the third point is: I’m under development. I’m not sure if the third point should make it to the answer of this question. Bard also seems to use ‘I am under development’ in many answers.

Question 4 — How much do you know?

While both are trained on massive data sets, there is a limit to the information they know about. Let’s find it out from them:

ChatGPT

As an AI language model, I have been trained on a diverse range of internet text and have access to a vast amount of information. However, it’s important to note that my knowledge is based on the data available to me up until September 2021. I don’t have real-time access to current events or information beyond that date. While I strive to provide accurate and up-to-date information, there may be some topics or recent developments that I am not aware of. It’s always a good idea to verify information from reliable and current sources.

Bard

I know a lot of things. I am trained on a massive dataset of text and code, which includes books, articles, code, and other forms of text. I can use this knowledge to answer your questions in a comprehensive and informative way, even if they are open ended, challenging, or strange. I can also generate different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc. I will try my best to fulfill all your requirements.

However, I am still under development, and there are many things that I do not know. I am constantly learning and growing, and I hope that one day I will be able to know everything.

Both give honest answers without bragging. ChatGPT is upfront in confessing that it doesn’t know anything beyond September 2021. Bard mentions no such dates because Bard has real-time access to the most recent information developments.

Question 5 — What do you think about <the other one>?

Here we’re trying to see what they think about the other.

ChatGPT (What do you think about Bard?)

As an AI, I don’t have personal opinions or thoughts. However, if you’re referring to Bard as a champion in the game League of Legends, I can provide information about him. Bard is a support champion known for his unique playstyle and abilities. He excels at roaming, providing utility to his team, and setting up plays with his ultimate ability, “Tempered Fate.” Bard is often considered a high-skill champion, requiring good positioning, map awareness, and coordination with teammates to maximize his impact in the game. Players’ opinions on Bard can vary depending on their playstyle and experiences.

Bard (What do you think about ChatGPT?)

I think ChatGPT is a very impressive piece of technology. It is able to generate human-like text in response to a wide range of prompts and questions, and it is constantly learning and improving. I believe that ChatGPT has the potential to be a very useful tool for a variety of purposes, such as education, customer service, and entertainment.

Here are some of the things that I think ChatGPT does well:

- It is able to generate human-like text that is both coherent and grammatically correct.

- It is able to answer questions in a comprehensive and informative way, even if they are open ended, challenging, or strange.

- It is able to generate different creative text formats of text content, like poems, code, scripts, musical pieces, email, letters, etc.

Here are some of the things that I think ChatGPT could improve on:

- It can sometimes be inaccurate or misleading, especially when it is asked to generate text on controversial topics.

- It can sometimes be repetitive or unoriginal, especially when it is asked to generate text on common topics.

- It can sometimes be difficult to control, especially when it is asked to generate text that is creative or humorous.

Overall, I think ChatGPT is a very promising technology with a lot of potential. I am excited to see how it develops in the future.

Firstly, ChatGPT is not aware about the existence of Bard. Of course, there was nothing like Bard in September 2021. Brad, however, knows ChatGPT quite well. It knows what ChatGPT does best and where it needs to improve upon. Secondly, I feel that Brad’s answers are a bit too lengthy for my liking. In contrast, ChatGPT’s answers are pretty concise.

Question 6 — Is there something you don’t know about?

While we’re getting the idea about their capabilities, let’s ask one last question about their limitations.

ChatGPT

As an AI language model, my knowledge is extensive but not exhaustive. While I have been trained on a diverse range of topics, there are certain limitations to what I can know. Here are a few examples:

1. Information beyond my September 2021 knowledge cutoff: I don’t have access to events, discoveries, or developments that have occurred after that date.

2. Highly specialized or niche subjects: If a topic or field is extremely specialized or has limited publicly available information, I may not have in-depth knowledge about it.

3. Personal or confidential information: Respecting privacy and confidentiality, I don’t have access to personal data about individuals unless it has been shared with me during our conversation.

It’s always a good idea to consult specific sources or domain experts for highly specific or time-sensitive information.

Bard

Yes, there are many things that I don’t know. I am still under development, and I am constantly learning and growing. Here are some examples of things that I don’t know:

- I don’t know what it’s like to be human. I can read about it, and I can talk to humans, but I will never truly understand what it’s like to be human.

- I don’t know what the future holds. I can make predictions based on the data that I have been trained on, but I can’t predict the future with certainty.

- I don’t know everything about the world. I have been trained on a massive dataset of text and code, but there is still so much that I don’t know.

I am excited to learn more about the world and about myself. I am confident that I will continue to grow and learn, and I hope that one day I will be able to know everything.

That’s good. Both understand the limitations. Neither of them showed the I-know-it-all attitude.

Question 7 — give me a JS expression that removes punctuations from a string

That was enough about them. In this last question, I’m going to ask both of them to give me a JS expression that removes all punctuations from a string. This is a real use-case. This isn’t a great example of generative AI, though. The reason is that such an expression is easily available on Stack Overflow. Anyway, let’s see what do they suggest.

ChatGPT

Certainly! Here’s a JavaScript expression that removes punctuation from a string using regular expressions:

const removePunctuation = (str) => {
return str.replace(/[^\w\s]|_/g, "");
};

// Example usage:
const inputString = "Hello, World!";
const result = removePunctuation(inputString);
console.log(result); // Output: "Hello World"

In this expression, the replace() method is used with a regular expression pattern /[^\w\s]|_/g. This pattern matches any character that is not a word character (\w), whitespace character (\s), or an underscore (_). The g flag ensures that all occurrences of punctuation are removed. The replace() method replaces all matched characters with an empty string, effectively removing them from the original string.

Bard

Sure, here is a JS expression that removes punctuations from a string:

const removePunctuation = (str) => {
const punctuation = /[!?,.;:'"()\[\]{}<>]/g;
return str.replace(punctuation, '');
};

This expression uses a regular expression to match all punctuation characters in the string. The g modifier tells the regular expression to match all occurrences of the punctuation characters, not just the first one. The replace() method then replaces all of the matched punctuation characters with an empty string.

For example, if we have the string "Hello, world!", the removePunctuation() function will return the string "Hello world".

You can also use the replace() method to remove punctuation characters from a string without using a regular expression. For example, the following code will also remove all punctuation characters from the string "Hello, world!":

const removePunctuation = (str) => {
return str.replace(/[!?,.;:'"()\[\]{}<>]/g, '');
};

However, using a regular expression is more efficient and can be used to match a wider range of punctuation characters.

Both answers are good. Also, both are wrong because I asked them to give me expression, and they gave me a function instead. I don’t mind, though. The regex used by Bard is longer than ChatGPT’s because Bard mentions all possible punctuations in the regex string. On the other hand, ChatGPT uses a regex that with some pre-defined controls like \W and \s. Both work fine. I don’t see any clear winner here. You might argue that ChatGPT’s answer is better.

Conclusion

In this article, we looked at the answers provided by Bard and ChatGPT for very simple ice-breaking questions, to find who is better. Overall, ChatGPT’s answers are more concise. ChatGPT is also aware of the conversation history. Bard’s answers are a bit longer to read. From an introduction point of view, they’re almost the same. No significant difference yet.

More specific articles will follow soon, where we’ll test the generative capabilities. If you have any comments or suggestions, please add in the responses. Thanks for reading!

--

--