Meet Bard, Google’s Comeback To ChatGPT | by Yasmeen Naseer | Feb, 2023
Since it was quietly launched last November, ChatGPT has been the talk of the town – and for good reason. It is able to answer complex questions with coherence and clarity and has many dreaming of a revolution in education, work, business and everyday life. People called it the Google killer and Google itself considered it important enough to ask the company to code red and bring in Google’s founders, Larry Page and Sergey Brin, to help deal with it.
Image source: Unsplash.
This is not the first time I have covered the war between Google and ChatGPT and in my earlier post on the subject I was quite clear that despite the fact that ChatGPT has risen in popularity and amassed millions of users in record time, it is not going to end of Google and I also explained why.
The architecture for GPT-3 was introduced in a 2017 paper written by Google researchers. The paper, called “Attention is All You Need,” was presented at the Neural Information Processing Systems (NIPS) conference and has since been cited more than 62,000 times. It is considered one of the most influential papers in AI and has been used to build modern language models, including GPT-3. However, Google’s development of its AI did not stop in 2017, and since then it has also developed the techniques to improve the performance of transformer-based models, including using massive training data sets, fine-tuning the model on different dialogue-related tasks , and using a special pre-training method called “conditional masked language modeling”.
It’s been less than 3 weeks since my article went live and Google has already shared some really exciting news:
“Two years ago, we unveiled the next generation of language and conversational capabilities powered by our Language Model for Dialogue Applications (or LaMDA for short).
We’ve been working on an experimental conversational AI service, powered by LaMDA, that we call Bard. And today we’re taking another step forward by opening it up to trusted testers before making it more widely available to the public in the coming weeks.” Sundar Pichai on the Google Blog
Unlike ChatGPT, which is only able to provide answers based on information until 2021, Bard uses information from the web to provide up-to-date, high-quality answers to user queries. According to Pichai, it is initially being rolled out with Google’s lightweight model version of LaMDA. This means that Bard will require significantly less computing power than LaMDA, allowing Google to scale it up to more users. As OpenAI did with ChatGPT, Google also aims to leverage external feedback along with internal testing to ensure Bard’s responses are high-quality, safe and accurate.
Bard represents the next logical step in Google’s mission to “organize the world’s information and make it universally accessible and usable.” That’s because, in Pichai’s own words, “AI can deepen our understanding of information and turn it more efficiently into useful knowledge — making it easier for people to get to the bottom of what they’re looking for and get things done.” Pichai goes on to explain this in more depth in the following words:
“When people think of Google, they often think of turning to us for quick factual answers, like ‘how many keys does a piano have?’ But increasingly, people are turning to Google for deeper insights and understanding – like, “is is the piano or guitar easier to learn, and how much practice does each need?” Learning about a topic like this can take a lot of effort to figure out what you really need to know, and people often want a diverse range of opinions or examining perspectives.
AI can be useful in these moments and synthesize insights for questions where there is no one right answer.” — Sundar Pichai on the Google Blog
Google, which initially did not focus on commercializing its natural language processing technology, was propelled into action by OpenAI’s launch of ChatGPT and Microsoft’s USD 10 billion investment in OpenAI along with Microsoft’s plans to integrate ChatGPT into its full range of products to integrate, including its number 2 search engine: Bing. Google, aware of its reputation, initially proceeded extremely cautiously with LaMDA, by its own admission, given the inherent susceptibility of NLP models trained on text scraped from the web to exhibit bias and hate speech. repeat. Some Google AI researchers feared AI and called it sentient, others left the company frustrated with its reluctance to commercialize the technology.
However, things have changed significantly since then and while Google hasn’t given an exact date on when Bard will roll out, it has said that it plans to roll out AI-powered features in Search soon. These features will take complex information and distill it into an easy-to-understand format so users can quickly understand the big picture of whatever it is they’re researching or trying to learn.
My childhood was spent with bulky computers that relied on slow dial-up internet; computers have raced forward ever since at an unfathomable pace that shows no signs of slowing down. It is indeed an exciting time to be alive and I can’t wait to see how these new competitive developments will shape the landscape of work, business and education as we know it.