Google has been hailed as one of the major leaders in artificial intelligence over the past decade, but it has always been reluctant to bring AI-based products and features to the public. So the tech community has been waiting a while to see how they interact with the bot (ChatGPT), which has caused a stir since its inception and is currently used by Microsoft in most of its products.
Google's response came on February 6; When it first announced its (cold) bot Bard, it wasn't open to the public at the time as its testing was limited to a small group of trusted testers, and then Google opened the door to having one on its March waiting list of people who wanted to use the Bard bot to join.
While the news everyone was waiting for arrived at its annual developer conference last weekend (Google I/O 2023), the focus was on generative AI - when Google announced that its chatbot Bard is open to the public without having to wait for 180 hours. Be impressed with the queue at two states.
The peculiarity of the Google Bard robot is that its current work is based on the new language model (PaLM 2) proposed by Google during the conference, which allows it to support multiple languages, which greatly improves its ability to understand, generate and recognize. Accurately translate text and now support code generation, debugging, and image understanding.
It should be noted that Bard bot has significant differences from other generative AI-based chatbots. Google Bard bot can access the internet and provide answers about current events while ChatGPT cannot access the internet as it is only available in the 2021 Information Training.
Therefore, the Google Bard bot competes more directly with Microsoft Bing Chat. It is based on the modern GPT-4 language model, and can access the Internet and provide real-time information.
Here are the main new features of Google Bard bot:
1- Multi-language support:
The Google Bard bot is based on a new language model (PaLM 2) that has been trained extensively on multilingual text, allowing the bot to support multiple languages. Google announced that it now supports three languages: English, Korean, and Japanese, and the list of supported languages will be expanded to include 40 other languages, including Arabic, in the coming period.
2- It improves logical thinking and math skills:
When training the model (PaLM 2), Google focused on improving its logical, deductive, and mathematical skills as it was trained on a large amount of mathematical and scientific texts, scientific articles, and websites with mathematical expressions.
Therefore, a cool robot has no difficulty in dealing with math problems and can solve advanced math problems and equations with great accuracy.
3- The ability to process images:
According to Google, the Bard bot will soon support image processing in prompts and responses, since the bot will display images and text in its responses and users can also insert images into prompts to ask or write to them. Where the robot uses the services (Google Lens) to understand the content of the images.
Google provides an example. Say you want to write a funny short story about your dog, you can upload a picture of your dog and write an instant message like “Write a funny comment about these two dogs”. Using Google Lens, Bard analyzes photos, learns dog breeds, and creates creative captions in seconds.
Therefore, the ability to process images and understand their context will be one of the features that will make Google Bard a strong competitor to any other chatbot based on the GPT-4 language model.
4- Create and fix the code:
PaLM 2 provides improved support for code generation and debugging, and the model has been trained in 20 programming languages, including popular ones like JavaScript and Python, as well as lesser-known ones like Prolog, Fortran, and Verilog.
Therefore, Bard can help you with programming and software development tasks including: generating code from text, debugging code you write, and interpreting code.
You can also easily export the code written by the bot to Google Docs or the Google Colab platform to write and run the code.
In addition, Google Bard bot can help you interpret the code. This can be useful if you are new to programming or need additional help understanding the code that the bot is writing.
5- Acceptance of the voice prompt:
Google Bard bot supports voice search so you can query longer queries faster. In the upper right corner of the chat box, you will find a microphone icon that you can click on and record your conversations.
6- Generate an image from the text:
Several companies offer tools for creating images from text, including Microsoft, which introduced an image creation feature (Bing Image Creator) based on its work on an artificial intelligence (DALL-E) model that generates images from text indexes. New Bing search engine and Edge browser.
Bard bots don't have this feature, but that's about to change as Google announces that its bot will soon be integrated with Adobe's tool (Firefly) to generate images from text descriptions, making it easy for you to turn ideas into high-quality work can. Images that you can use as is or edit with photo editing tools and apps.
7- Export chat history in one click:
Below the Bard bot replies, Google has added a new share button. With it you can export responses to Google Docs, save as draft in Gmail, or export code to platform (Google Colab). Google says that it will soon be able to export code to the Replit platform as well.