When ChatGPT was released late last year, it quickly gained a lot of attention because it made its AI engine freely available to the public. Suddenly, anyone could type in a question, and ChatGPT responded with fresh, human-like answers in seconds. From writing essays about the First Crusade to short poems about Al's Gore love for Toyota Prii (plural for Prius), ChatGPT spewed out answers that Google and Bing never could.
While traditional search engines add links to websites that best match a person's search query, ChatGPT searches large datasets and uses large-scale language models (LLMs). provide answers to people by creating sentences that mimic human answers. It is described as a powerful autocorrect. This has thrown both Microsoft and Google into turmoil, given that ChatGPT had an estimated 100 million active users as of January, making it the fastest growing web platform of all time. Microsoft's Bing, which previously had less than 3% of its search market share, quickly adopted ChatGPT and integrated AI into search. Microsoft actually licenses GPT technology from OpenAI to Bing. Since then, traffic has increased by about 16%.
Other products also integrate various forms of generative AI, such as the "Copilot" tool in Microsoft Word, Excel, and PowerPoint, and AI capabilities in Google Workspace tools such as Gmail and Docs. Snapchat, writing assistant Grammarly, and WhatsApp are also using AI.
However, not all AI chatbots are created equal. The test below compares answers from the paid version of ChatGPT (compared to the free version, his 3.5), which uses GPT-4, with answers from both the Bing search engine integrated version of ChatGPT and Google's own Bard AI. did. We compared the systems. (GPT, by the way, stands for Generative Pretrained Transformer.) Bard is currently in invite-only beta, and while Bing is free, you must use Microsoft's Edge web browser.
Poem
The fun part of using an AI chatbot is giving it silly prompts and seeing what it spits out. Watching a chatbot create rhymes and beats in real time is a really interesting exercise.
Among Bing, Bard, and ChatGPT, OpenAI's service is the best poet. ChatGPT not only has richer prose, but also more creative rhymes and diction. Bing and Bard's poems seemed lazy, but ChatGPT has created something that feels like they took their time and thought into each poem.
The invitation to write a poem about an online influencer who gradually realizes that she's not that important is meant to be both funny and self-help. Only ChatGPT was able to get to the heart of the existential crisis faced by this fictional influencer and end it on a positive note that still felt real.
Interestingly, Bing allows people to increase their level of creativity. Bing's balanced execution of this verse felt trite and trite. In "Creative Mode," Bing opted for more colorful language and less pressure. It came close to ChatGPT, but still wasn't quite there. The bard's poetry felt lazy in comparison. Many words were repeated and little attention was paid to rhyme or meter.
ChatGPT dominated this exercise.
Main difference
Bard, Bing, and ChatGPT all aim to provide human-like answers to questions, but each performs differently. Bing uses the same GPT-4 technology as ChatGPT, but can go beyond text and generate images. Bard uses his own Google model called LaMDA, which often produces answers with less text. (Google CEO Sundar Pichai said Bard will move to PaLM, a more advanced dataset, in the near future.) All of these bots can occasionally make factual errors, but three bots The least reliable of these was Bard.
Although ChatGPT and Bing both use the same technology, entering the same query in both will not yield the same results. This is part of the nature of generative AI. Unlike traditional search, which aims to highlight the most relevant links, AI chatbots create text from scratch by creating new answers from records. For example, if you ask the chatbot to write a poem about Pikachu's love of ketchup twice in a row, he will get a different answer each time. Another reason why posting the same question on ChatGPT and Bing can give different results is that Bing adds its own layer on top of his GPT-4. “We have developed a unique way to harness the full power of OpenAI models,” said a Microsoft spokesperson. “We call this collection of skills and techniques the Prometheus model.”
The Prometheus model combines Bing's search index with his GPT-4, allowing it to provide up-to-date information, unlike the ChatGPT dataset, which only contains information up to 2021. Bing also allows you to develop a range of conversational styles that are balanced, creative, and accurate. A Microsoft representative could not comment on the quality of ChatGPT compared to Bing, but said the engine benefits from all the improvements OpenAI is making to GPT-4. The representative also said Bing leverages Microsoft's Azure AI supercomputing technology to integrate search, chat, and the Edge browser. Google and OpenAI did not respond to requests for comment.
Recipe: Chai Tres Leche
Tres Leches Cake with Chai combines parts of South Asian and Latin American cuisine into a moist, spice-filled cake. Rather than ask an AI chatbot to bake a simple chocolate cake (there are tons of recipes online), we thought a more specific method might be more difficult.
Break down complex topics
It’s another thing for an AI chatbot to provide information on complex topics. What's even better is that you can extract this information for different audiences. For this test, we asked Bing, Bard, and ChatGPT to explain quantum physics to fourth graders.
Bird created most of the text for this query, but the language was more complex and probably not fully understandable to his fourth graders. Bird fell into the same trap, using difficult words such as "subatomic" and "proportional" that might be too difficult for elementary school students. None of the chatbots performed particularly well in this test, but ChatGPT provided the most obvious answer.
this is just the beginning
Currently, ChatGPT (paid version) is the best chatbot currently available. There are more human and detailed answers than Bing's and especially Bard's answers. However, these products are constantly evolving. As Google, Microsoft, and OpenAI feed AI more data and continue to optimize it, we should see improvements. Google would benefit most from switching from LaMDA to PaLM. The current version of Bard is simply not good enough. If new developments occur, we will update this guide accordingly.
Until then, please use ChatGPT.