Google launched Bard in February, and compared to its competitors (Bing and ChatGPT), it somehow didn’t differ much, except for a demonstration with false answers, which led to a sharp drop in the company’s stock. Today, at the Google I/O presentation, they presented an updated chatbot that has been running for some time on PaLM 2, the next generation of the company’s own large language model.
Google is canceling the waiting list for Bard* and making it available to users in 180 countries – for now in English and with support for Japanese and Korean. In the near future, the chatbot will support about 40 languages, including Ukrainian.
For users from Ukraine, the chatbot site gives a message saying “Bard is not yet supported in your country” – although the attempt to access via VPN was successful.
The chatbot becomes multi-modal, meaning it can work with images in requests and provide them in responses (say, if you are looking for information about a tourist place, Bard will show you popular attractions). It can also now display results using Google Maps and structure the data in a table.
Subsequently, extensions will be launched for the chatbot – including from Google’s own programs and services, such as Maps, Sheets, Mail and Docs (soon also make it easier to export Bard responses to Mail and Docs). Third-party partners include Adobe’s Firefly image generator, with Wolfram Alpha and more coming soon.
Google is also expanding Bard’s coding capabilities, based in part on Google Lens. As the company notes, this includes not only the ability to generate code, but also explaining snippets of code and helping users customize it.
In previous iterations, Bard was already able to export Python code to Google Colab. The company is now expanding this feature to allow Python code to be exported to Replit, a popular browser-based IDE.
Oh well, Bard is also going to the “dark side” and getting a dark mode.