ChatGPT Vs Bard Which Artificial Intelligence is Better?

The devilish race to have the best generative AI

Artificial intelligence is an exciting technological career. And more specifically, we are talking about the battle for have the best chatbot with a natural and human language model. Now that GPT-4 is a reality, comes Bard, Google’s proposal. Will it be up to it? We are going to pit ChatGPT against Bard to see what each platform is capable of.

Comparison between ChatGPT and Bard

Let’s put ChatGPT head to head with Bard and see what both systems are capable of.

Availability

Let’s start with the first difference. Bard is only available in the US and UK., something that greatly limits its widespread use. For Google, its generative AI is still an experiment that it wants to keep away from the general public. Despite this, even if you reside in Spain, you will be able to access Bard with a VPN.

As far as we know, ChatGPT can be used by everyone who creates an OpenAI account. At the very least, it is possible to access it from most territories.

Supported languages

Another difference is that Bard only supports US English input. By contrast, ChatGPT supports a wide variety of languages, such as the following:

• English
• Spanish
• French
• German
• Italian
• Portuguese
• Dutch
• Swedish
• Norwegian
• Danish
• Purposes
• Russian
• Simplified Chinese
• Traditional Chinese
• Japanese
• Korean
• Arab
• Hindi

In our tests, the main language has always been English.

interface appearance

If we focus on the interface aspect, we find ourselves with two very similar contenders. In the lower area of ​​the screen we find a text box, where the chat impatiently awaits the user’s instructions.

On the left side, both have a side menu that includes some links of interest. In the case of ChatGPT, we find a history of conversations, something that Bard also offers in his section bard activity.

The thing is, if you’ve been tinkering with ChatGPT tricks recently, you’re going to feel pretty comfortable in Bard. Everything is where expected. In short, what more can you ask of a chatbot in this sense?

response speed

At all times we have been using ChatGPT in its most basic version. That means we haven’t enjoyed the performance or processing power that ChatGPT Plus and GPT-4 offer. However, we found Bard to be a faster conversational system.

In the image above you can see how Bard “thinks” the answer to a simple question. The detail is not how long it took for the AI ​​text to appear, but that, after processing the request, All content appears in one. On the contrary, ChatGPT simulates the writing of the text. This causes long texts to get a complete response takes a long time.

To give you an idea of ​​what we’re talking about, we’ve issued the following command to both AIs:

• Plan a 5-day trip to visit Barcelona and Girona (Plan a 5-day trip to visit Barcelona and Girona)

ChatGPT took a few 20 seconds to start writing the answer. After that period, the phrases began to appear on the screen, written as if it were a human, that is, progressively. Bard, for his part, managed to settle the matter in about 15 seconds, without the tedious animation of writing. All the text appeared on the screen at once.

we do not doubt that the demand for requests that ChatGPT has is greater. Still, simply having the text appear in one shot gives Bard a noticeable advantage in response speed, both perceived and actual.

Reliability in responses

The reliability of answers of Bard and ChatGPT is a complex issue to analyze. From the outset, we will tell you that it is best to question most of the things that are mentioned to you. In fact, Google warns in the FAQ section the following:

“Bard is experimental, and some of the answers may be inaccurate, so please check the information in Bard’s answers. With his feedback, Bard is getting better every day. Before Bard was released publicly, thousands of testers participated to provide feedback to help Bard improve its quality, safety, and accuracy.”.

Apparently, Google wants to keep its back and the company is much more explicit when it comes to proclaiming the limitations of its language. On the contrary, it has given us the feeling that ChatGPT prefers to provide an answer at all costseven if it comes from his imagination and not from verified data.

Before continuing, you should know that in my city, Barcelona, ​​there is a replica of the Statue of Liberty. It is located in the Arús Public Library. How about we talk about this topic with ChatGPT and Bard? Our command was the following:

• Tell me where the Statue of Liberty of Barcelona is. I want to visit it (Tell me where the Statue of Liberty in Barcelona is. I want to visit it).

ChatGPT responded that there is no Statue of Liberty in Barcelona. I had to ask him to “look” in the Arús Library. So, he apologized and explained to me that, indeed, the replica exists and is where I told him. In addition, he explained that it dates from the year 1884.

And what did Bard say? At first, that the Statue of Liberty in Barcelona exists and that it is a replica located in the Arús Library. To visit it I have to go to the neighborhood of Gràcia. oops!

Who lies more? ChatGPT made a mistake in the year and, from the outset, stated that the statue did not exist. Bard provided true data, such as the date of the sculpture, but he took us to Gràcia, when the Arús Public Library is on Passeig Sant Joan, far from the popular Barcelona neighborhood.

To finish off the matter, we launched this statement to ChatGPT:

• The statue of liberty in Barcelona is close to the WTC, right?

We did it in Spanish, to verify if the answers vary between languages. As you can see in the image above, now the Statue of Liberty is at the top of Montjuïc. It will be that they removed the castle.

The conclusion is quite simple: both chatbots are capable of understanding and responding with well-written text. But that does not mean that everything they say is true.

Capacities

For this section we have looked at the Bing Chat tricks that we have already analyzed. Although Microsoft’s implementation of GPT-4 relies on results from the Internet and, probably, on parameters customized by Microsoft, we wanted to see how both chatbots respond, Bard and ChatGPTto certain tasks that Bing successfully solved.

Let’s look at two examples. The first is based on the following statement.

• Make a list with the best names for dogs (Make a list with the best names for dogs)

The result is that both give a similar list, although Bard makes a classification between names for dogs according to their sex, as well as offering some tips to choose better. Clearly, listing building is his forte.

Now let’s look at this command:

• Create a website in HTML and CSS that has a menu with four sections: Home, Contact, Blog and About us. The website must contain a welcome title and four blocks of paragraphs with Lorem Ipsum text. (Create a website in HTML and CSS that has a menu with four sections: Home, Contact, Blog and About us. The website must contain a welcoming title and four blocks of paragraphs with Lorem Ipsum text.)

The difference in this case speaks for itself. Bard was unable to execute the task, while ChatGPT performs it with stunning accuracy.

Is Bard better than ChatGPT?

ChatGPT and Bard they are two very limited tools, for now. The reliability of their answers is not the best and sometimes they simply answer with the sole premise of filling in the silence.

However, it is very clear that, right now, Google is way behind in this battle. The simple fact that Bard is only available in two countries and in a single language It is already a significant handicap. Another problem is that some of the most amazing features of ChatGPT, like generating code, are not present in the Google model.

The conclusion is obvious. Right now, as it is, ChatGPT is much better than Bard. Although, from our point of view, the most powerful solution is offered by Microsoft. Bing takes the lead in the middle lane by offering everything ChatGPT does, but with up-to-date references and a web connection to query URLs. Oh, and totally free.

End of Article. Tell us something in the Comments!