After the success of ChatGPT, Google has been trying to catch up to Microsoft with their AI chatbot, Bard. Initially, it had a rocky launch due to a factual error in their advertisement of Bard AI. This caught people’s eyes and caused the company to have a downfall in the stock market. Then came GPT-4, a language model even more powerful than GPT-3.5 which raised the hype and popularity of Microsoft and OpenAI.
Google has been letting people try their Bard AI by joining a wishlist although it was restricted to some specific regions. On May 10 during the Google I/O 2023 event, the company finally made the experimental version of Bard open for all. Now, anyone can try out the chatbot.
With improved math, logic and reasoning skills, Bard can now help generate, explain and debug code in 20+ programming languages — and coding has quickly become one of the most popular things people are doing with Bard. #GoogleIO pic.twitter.com/Vcx5X6y0w7
— Google (@Google) May 10, 2023
Bard is open to all
Bard was initially developed on LaMDA Language Model but it was later updated to the Pathway Language Model. It was trained on 540 billion parameters and claims that it can keep track of 10 sentences to maintain a consistency of dialogue. For each prompt, it generates 3 drafts of responses.
Bard is still an experimental version, meaning there is still a lot of room to improve but it is still pretty good, at least in my experience. There are light and dark theme options, which can be changed by going to the side panel. The “Updates” section in the side panel lists the changelog.
Bard will soon become more visual in a couple of helpful ways. #GoogleIO
1️⃣ Bard will incorporate relevant images in its responses. ↓ pic.twitter.com/My6VIBARUg
— Google (@Google) May 10, 2023
Google announced that they are introducing multi-modality in Bard soon. It will be able to take images as input similar to GPT-4. While it is vulnerable to fake and misinformation found on the web, it is also a better alternative than ChatGPT which doesn’t have accurate information after 2021. The AI chatbot will be able to create images with the help of Adobe Firefly as well.
Despite being in the experimental stage, Bard shows potential. Google has always used its AI in other services, like Photos, Maps, etc. Now that they are using AI as a standalone service, like Microsoft, it is time to see how the future of AI will unfold.
YouTube: Bard | Google I/O 2023
Photo credit: The feature image used is owned by Google and has been provided for press usage.
Source: Matthew Sparks (NewScientist)