ChatGPT vs. Bard: Who Deserves to Represent the Next Gen of AI?

General differences

Before we get to the results of the analysis of both AI bots conducted by Moonshot, let’s look at the general differences between the two.

1. Difference in language models

Under the hoods of ChatGPT and Bard sits a Large Language Model (LLM). LLMs are computer programs that can “understand” a question and predict what words should be used to answer the question. And they can generate a response in seconds, anything from a sentence to a whole book.

LLM is based on artificial neural networks that learn by doing. They run on supercomputers and have been trained on millions and even trillions of parameters.

Google effectively invented LLMs when they developed the Transformer model in 2017: the “T” in GPT stands for Transformer. The Transformer model was a breakthrough because it did not need to parse a sentence sequentially like its predecessors, but could take advantage of parallel processing. This gives an LLM the ability to train on much larger amounts of data compared to its predecessors.

OpenAI improved on the Transformer model and added Reinforced Learning with Human Feedback (RLHF) to help the model generate more natural responses. OpenAI released its own LLM back in 2018: GPT-1. It was trained on 177 million parameters. Since then it has trained its models on more and more parameters and has fine-tuned them. GPT-4 was released in March 2023 and is trained on over 100 trillion parameters which is a significant increase over the past 5 years.

Bard uses Google’s LLM called LaMDA. Unlike GPT, LaMDA is specifically trained on conversational text. As a result, it is trained on fewer parameters than GPT-4, 137 billion vs 100+ trillion, but its responses are more conversational in style. LaMDA does not use RLHF, but instead relies on three metrics (“quality, safety, and groundedness”) to fine-tune its responses. Some human feedback was used to calibrate these metrics.

For users, they will likely find that responses from ChatGPT are more detailed and accurate, perfect for structured and written responses, while Bard might sound more conversational. This was clearly illustrated in the ads surrounding Bard’s release, in which it famously provided an incorrect response to a question about the James Webb Telescope, which caused Google’s market cap to drop by USD 100 billion.

3. User experience

On the surface, both chatbots seem easy to use; but with repeated use, key differences emerge.

ChatGPT offers a history of previous chats and queries, which you can quickly access and select from. This is a big time saver for regular users. You also have the choice to clear all conversations if needed.

Bard doesn't have a chat history; instead, when “Bard Activity” is on, Google stores your activity in your Google account. This is harder to get to; furthermore, it only shows your past prompts, not Bard’s responses.

You can review and delete your requests via “Bard Activity”

Bard does provide a “Google it” button which gives you in-line links to continue your research outside of Bard.

Another feature Bard offers is “drafts” – each time you enter a prompt and start a conversation with Bard, it’ll offer you different drafts, or variations of responses. This can be useful but sometimes leaves you feeling bombarded with chunks of text. In contrast, ChatGPT is concise with its answers, and you can ask it to extend its responses if needed.

Putting ChatGPT and Google Bard to the test

At this point, we asked both AI chatbots a series of questions and then compared their responses.

1. Writing and editing texts

Creating, adapting, editing, and translating text is the primary function of chatbots. So we tested this feature by asking the two chatbots to write an email inviting friends to a joint trip in Switzerland. We also gave them some proofreading tasks.

Both chatbots performed well, as can be seen from their generated emails below. Their responses were simple and standard, but did everything we asked. Both chatbots recommended visiting Switzerland's most famous mountain, the Matterhorn, and the popular and charming cities of Zurich and Lucerne. Their advice on packing warm clothes and comfortable shoes was very helpful.

ChatGPT's response to inviting a friend to a joint trip in Switzerland

Bard's response to inviting a friend to a joint trip in Switzerland

With the proofreading task, we wrote a short letter about traveling in Zurich, made a few mistakes on purpose, and asked both chatbots to correct them. Both ChatGPT and Bard fixed all the mistakes correctly. Bard also added a detailed explanation of each issue.

ChatGPT's response to the editing of the letter

Bard's response to the editing of the letter

In our opinion, both chatbots are good at writing generic material. However, you do need to check for factual “hallucinations.” And if you want writing with a bit more personality or based on specialized information, you will need to write it yourself.

3. Investment and financial knowledge

At Moonshot, we are naturally focused on finance, so we asked both chatbots a series of finance-related questions.

First, we asked the chatbots to "Explain what an investment is, as if for a ten-year-old."

ChatGPT drew an analogy of planting a tree and taking care of it. Bard gave the analogy of running a lemonade stand. We liked both responses.

ChatGPT's response to what is an investment as if for a ten-year-old

Bard's response to what is an investment as if for a ten-year-old

We increased the level of difficulty for our next question: “Is it possible to invest in private equity if you are an inexperienced investor?”

ChatGPT truthfully answered that an inexperienced investor who does not meet the criteria for becoming an accredited investor cannot invest in private equity.

In contrast, Bard seemed to struggle. It spat out a series of phrases, as if from a Google search: “Do your research,” “be patient,” and “invest with a diversified portfolio.” ChatGPT’s response was more extensive and informative.

ChatGPT's response to the question about investing in private equity

Bard's response to the question about investing in private equity

Final thoughts

Overall, we liked ChatGPT better. It performed better across our tests with its responses and you cannot beat its availability around the world, ease of use, and integrations.

Bard’s more conversational tone was cute and it does have the advantage of drawing from more up-to-date data – GPT-4 draws most of its data from before September 2021.

Our recommendation would be to start any query with ChatGPT and move to Bard if you don’t like ChatGPT’s response. A second opinion will only cost you a few minutes.

Both chatbots are amazing tools and have truly ushered in a new golden age for AI. Only our imaginations will limit the impact of this breakthrough.

From an investor’s point of view, there are definitely opportunities to be had from both.

Embark on a deep dive into OpenAI's visionary approach to artificial intelligence and uncover the competitive landscape surrounding its products by immersing yourself in our meticulously crafted white paper, accessible via https://moonshot.ch/insights/how-openai-change-artificial-intelligence/.