Google Bard, the AI chatbot developed by tech giant Mountain Valley, faced a disappointing reception upon its initial launch. To exacerbate matters, the chatbot even resorted to plagiarizing content, including in the Google Bard vs ChatGPT comparison. Consequently, the situation appeared bleak.
However, the tides have turned at the Google I/O 2023 developer conference, as the company showcased an extensively upgraded version of Google Bard AI, equipped with a plethora of remarkable enhancements. Curious to learn about these astounding advancements? Read on for a comprehensive overview of the new features below.
Google Bard Powered by PaLM 2
In a significant transformation, Google Bard has transitioned from being powered by a fraction of LaMDA, one of the company’s large language models (LLMs), to now relying entirely on PaLM 2, the next-generation language model developed by Google. This recent upgrade introduces substantial improvements to the reasoning capabilities and coding structure of Google Bard.
The integration of PaLM 2 unlocks a wide array of new features that were previously unattainable. Google highlights in its blog post that PaLM 2 benefits from an expansive dataset that encompasses scientific papers and mathematical expressions.
Consequently, Google Bard exhibits enhanced proficiency in complex logical reasoning and mathematical problem-solving. Moreover, PaLM 2 operates at a faster pace than its predecessors, resulting in quicker responses from Bard. The new LLM encompasses various other advancements, further augmenting its capabilities.
Google Bard Serves Multiple Drafts
While ChatGPT is known for its prompt responses, it can sometimes feel repetitive due to its limited set of answers. On the other hand, Google Bard boasts an exceptional feature called “drafts,” which sets it apart. Drafts allow users to obtain multiple responses from the AI chatbot for the same query. Each response generated by Bard is unique, offering different perspectives based on the given prompt.
By default, Google Bard generates three distinct versions of the answer to a query. This grants users the freedom to select the response that best suits their needs. You can choose a single response or mix and match elements from multiple responses to tailor the answer to your liking.
The draft feature is seamlessly integrated into Google Bard, requiring no special steps from the user. For those seeking varied interpretations from the AI, the draft feature in Google Bard proves to be a valuable asset.
Google Services Integration
In addition to its existing chatbot functionality, Google Bard has an ambitious roadmap for further enhancement by integrating various services into its platform. With the goal of streamlining workflows, Google aims to establish seamless connections between Bard and other services, including Google Sheets, Docs, and Maps. This integration will enable these services to work harmoniously, infusing Google Bard with a host of new features.
During a recent keynote, Sissie Hsiao, the General Manager for Bard, exemplified this integration by demonstrating how she leveraged the bot’s capabilities. She instructed Bard to find college programs based on her interests. Once Bard presented the results, Sissie requested it to showcase them on a map, and the AI swiftly displayed them using Google Maps.
Taking it a step further, she then asked Bard to organize the results neatly into a table with various columns for college, locations, degrees, names, and more. Bard successfully created the table and seamlessly exported it to Google Sheets, preserving the formatting.
This achievement signifies a significant advancement in Bard’s ability to facilitate cross-platform collaboration. Although not all features are currently available, Google has plans to introduce these functionalities to Bard in the near future.
Built-in Image Generator
In the pursuit of offering comprehensive services, Microsoft and other companies have achieved significant success by incorporating native features such as the Bing AI Image Creator into their portfolio. These invaluable tools, coupled with state-of-the-art AI art generators, have become indispensable. While Bard has previously lagged behind in this particular aspect, a transformation is on the horizon.
This innovative capability is made feasible through Adobe Firefly, a collection of creative generative AI models renowned for producing high-quality images in real-time. Google plans to integrate Firefly’s model directly into Bard, granting users effortless access to the service.
Consequently, users will be able to make image requests through prompts and receive image outputs generated by Adobe Firefly, thereby enhancing the creative potential of Google Bard.
Support for Extended Languages
Despite the multitude of features offered by OpenAI’s GPT-4 LLM, such as multimodal responses, it falls short when it comes to language support, covering only slightly over 26 languages. Although this is already a considerable number, it fails to encompass the vast array of languages spoken worldwide. However, Google surprised everyone with its recent announcement that Bard will soon provide support for over 40 languages.
This new addition significantly expands the capabilities of the AI bot, allowing it to engage in full-fledged conversations with users around the globe in their preferred language. These 40+ languages will include widely spoken ones, ensuring broader support for diverse communities.
Notably, Google emphasized that Bard will soon be available for interactions in Korean and Japanese as well. With extended multilingual support, Bard’s ability to communicate across various languages will undoubtedly become one of its standout features once the feature is fully implemented.
Bard AI Can Code for You
The FAQ page for Bard AI confirmed this limitation, stating that coding support had not yet been implemented. Although Bard could provide basic HTML responses, the generated code was often incorrect and unreliable. However, the recent Google I/O 2023 event has unveiled exciting news—Google Bard AI is now equipped with programming features.
It can now execute a wide range of coding tasks, including code generation, debugging, and code explanations. Demonstrating its capabilities at I/O, Bard AI effortlessly generated the “Scholar’s Mate move in Chess” using Python and even offered a comprehensive explanation for the code. To ensure accuracy, Google Bard will provide citations for the rendered code, allowing users to double-check the information.
Upon completing the code, users will have the option to export it to either Google Colab or Replit for execution and obtaining results. It will support over 20 programming languages, including Java, C, HTML/CSS, C++, Kotlin, Dart, and more.
Google Bard Can Now Read Images
Although Google Bard may not possess the multimodal capabilities found in OpenAI’s latest GPT-4 LLM, it now possesses the ability to intelligently interpret images and generate corresponding responses. This achievement is made possible through the direct integration of Google Lens into Bard. For those unfamiliar, Google Lens is a cutting-edge image recognition technology developed by the company. With this integration, Bard becomes capable of comprehending various types of images and formulating appropriate responses.
Google recently demonstrated this capability by providing Bard with a picture of two dogs and requesting it to generate a humorous caption for the image. Within a matter of seconds, Bard analyzed the image and produced several amusing captions, showcasing its creative prowess. Although other use cases were not showcased during the presentation, it is evident that Google Bard’s image recognition capabilities are among its most noteworthy features.
Bard Gets More Visual
Upon Microsoft Bing AI Chat’s public release, it introduced an array of impressive features, including the ability to respond with images. Google has now followed suit and equipped Bard with the capability to provide visual responses. In the near future, when presented with various queries, Bard will leverage Google Search intelligently and offer contextual visual information alongside text-based results.
These combined results will be conveniently displayed within the chat interface itself, ensuring easy access for users. Google asserts that these accompanying images will provide users with a more comprehensive understanding of their search outcomes, a sentiment we wholeheartedly agree with.
While the feature is not currently live, the integration of image insertion within Bard is expected to become available to users in the following weeks. Furthermore, it will also be introduced to Google Search as part of the upcoming “Generative AI” functionality.
Export Bard Chats with a Click
While Google has recently introduced an export button for coding outputs, the company is taking it a step further in terms of sharing and utilizing outputs.
From now on, you will have the ability to export your Google Bard conversations directly to services such as Google Docs and Gmail. This feature applies to all types of messages generated by Bard.
Google has introduced these one-click options with the aim of simplifying workflows and making the entire process of handling AI outputs even more efficient. By enabling direct exports to popular Google services, Google Bard aims to enhance productivity and streamline the utilization of AI-generated content.
Google Bard Tools (Plugins)
The integration of services extends beyond Google’s own products, as Bard AI aims to become even more useful through collaborations with external companies and their extensions.
If you’ve been following recent developments in AI, you may recall that OpenAI’s ChatGPT introduced plugin support, a significant advancement that enabled the bot to benefit from additional experiences. Google is not lagging behind in this regard. In addition to incorporating Adobe Firefly, Bard will also receive its own plugin experiences through partnerships with various companies.
In the near future, Google plans to collaborate with notable services such as Spotify, Kayak, OpenTable, Instacart, Wolfram, Khan Academy and ZipRecruiter to offer selective experiences within Bard.
This partnership will expand the range of Google Bard features, opening up possibilities in productivity, leisure, education, and much more. By combining these external collaborations with Google’s integration of its own services, the forthcoming tools for Bard will empower the AI bot to accomplish even more tasks and cater to a broader range of user needs.
Bard AI Gets Dark Mode
While some users may consider it a minor addition, the audience at Google I/O had a different perspective. It is now official that Google Bard has introduced a brand-new feature: dark mode. Following the trend of dark modes in various applications, Bard’s dark mode presents an aesthetically pleasing combination of dark and grey shades.
To switch to the newly introduced dark mode, users simply need to click on the convenient dark theme button located on the bottom left sidebar. With this simple action, users can experience one of the most anticipated and appreciated features of Google Bard. So, go ahead and enable this highly sought-after feature for a visually appealing chatbot experience.