You: “Alexa, what’s the weather like today?”
Alexa: “The weather today is 25 Celcius; the weather is cloudy and will remain so till 4 PM.”
Alexa is one of the most efficient AI assistants. Mainly, in my opinion, because it at least understands Indian accents better than Siri. Siri in my dad’s iPhone takes 5-8 times to understand my Indian English compared to Alexa, which can understand my Indian English better.
Regarding competition, Google doesn’t wait too long to push itself into the race. In the same fashion, Google was surely not letting OpenAI’s ChatGPT take all the attention. So, on 7th February 2023, Google launched Bard. Bard is Google’s experimental, conversational AI chat service.Sundar Pichai, CEO of Google, announced the project in a blog post. According to Sundar Pichai, the public will be able to use it more in the coming weeks. Bard employs a scaled-down version of LAMDA, a powerful AI model first proposed by Google in May 2021 and developed on the same architecture as ChatGPT. He described Bard as an “experimental conversational AI service” that will converse with users and reply to their questions. A group of “trusted testers” is now testing the program.
Bard uses Google’s Language Model for Dialogue Applications to get responses from the internet (LaMDA). Through this linguistic approach, Bard will provide more details to questions than a standard Google search. LaMDA’s lighter and second version utilizes fewer computational resources, allowing it to scale for more people to use and provide feedback.
How Does Bard Work?
The Bard can assist you in expressing your creativity and also arouse your interest. Bard “uses information from the Internet to give new, good responses,” according to Pichai. It indicates that it can answer inquiries about recent events, which ChatGPT finds difficult.
While LaMDA is based on Transformer, Google’s neural network architecture that it designed and publicly released in 2017. According to Google, GPT-3, the language model on which ChatGPT operates, was also created on Transformer.
According to the announcement, the initial version of Bard will employ a lightweight model version of LaMDA because it takes less computer power and can be scaled to many people. In addition to LaMDA, Bard will use all available web information to deliver responses. Pichai said pulling from the web would provide “fresh, high-quality responses.”
The Difference Between Bard and ChatGPT
Well, the primary distinction between Bard and ChatGPT is the data source. Bard will constantly pull information from the internet, ensuring that everything is up to date. ChatGPT’s data sources finished in 2021. Therefore it is confined to more recent research and information. By obtaining the most recent study, Bard will have more data to gather insights in real-time.
For dialogue applications, Bard uses LaMDA, whereas ChatGPT employs GPT 3.5. LaMDA was created to interpret natural language using an open-source network. It has been trained to look for patterns in phrases and between words to generate dialogue rather than individual words.
Bard even produces more information in smaller bits, while ChatGPT creates content in response to a single text input. Even though the underlying technology of ChatGPT could be more innovative, OpenAI’s decision to make the system publicly available online allowed millions of people to learn about this novel way for computers to produce text which is quite impressive.
Google’s Bard is still in its early stage of development. The competition from other AI Chatbots has even sent Google into a highly competitive mode. If you look into history, this competitive spirit Google has cultured and might have produced excellent results like Google Maps, Gmail, Google Drive, Youtube, etc., a lot of Google’s own employees admit that Bard needs to be revised to the usual standard of Google. But it’s a fact that everything takes its own pace
Though, I’m hopeful for the future. Because Google has always primarily proven itself with its excellent interface and easy usage.