In a significant development this May, Google unveiled Google Bard, an AI-powered tool designed to augment writing, research, and coding activities. Google Bard, powered by Google's LaMDA technology and trained on an extensive dataset of text and code, is set to make a mark in AI-assisted productivity tools.
Google Bard's capabilities span generating a wide variety of creative content, including poems, scripts, and code; translating languages; and providing insightful answers to various questions.
Sounds very similar to ChatGPT? Perhaps because the pioneering work on generative AI models was developed at Google and was open-sourced. So why was Google slow off the mark with Bard? Reportedly Google’s Bard team has been working on a war footing since ChatGPT 3.5 was released last year. This was indeed an existential threat to their search business. Sundar Pichai also announced how AI is bein integrated across the board inside several offerings, including their Office suite (sheets/ docs et al.). It remains unclear what impact Bard will have on their search revenues, but now that Bard, the genie, is out of the bottle, it can't be put back!
How to use it?
Although still under development, the tool has demonstrated considerable potential.
Google Bard is accessible through the Google Cloud Platform. Users must create an account and enable the Google Bard API to benefit from its functionalities. It can also be accessed directly via https://bard.google.com/.
What is Bard?
Bard is a large language model (LLM) chatbot developed by Google AI. It was introduced in 2022 and is still under development. LLMs are trained on a massive dataset of text and code. This allows Bard to communicate and generate human-like text in response to a wide range of prompts and questions, like writing different kinds of creative content, translating languages, and answering your questions in an informative way.
Here are some of the things Bard can do:
- Follow instructions and complete requests thoughtfully.
- Answer your questions comprehensively and informally, even if they are open-ended, challenging, or strange.
- Generate different creative text formats of text content, like poems, code, scripts, musical pieces, emails, letters, and more.
Here are some more details about Bard:
- Bard is built on top of LaMDA, Google's language model for dialogue applications. LaMDA is a generative model that can create realistic and engaging conversations.
- Bard uses a transformer architecture, a neural network well-suited for natural language processing tasks. Transformers can learn long-range dependencies between words, which allows them to generate more accurate and coherent text.
- Bard is trained on a massive dataset of text and code (1.56 trillion words, which is about 10 times larger than the dataset used to train GPT-3). This dataset includes books, articles, code, and other text forms. The dataset is constantly updated, allowing Bard to learn new things and improve its performance.
I tested Google Bard with coding while creating an App Script app. Although it performed decently, it made more errors compared to ChatGPT. This could be attributed to the complexity of the specific example I used. ChatGPT could generate the code I wanted pretty quickly, although it also needed iterations.
Interestingly, because of access to real-time information, Google Bard has a slight edge over ChatGPT in its default state, without plugins. For ChatGPT to access real-time information, I need to use a paid version and enable the Bing plug-in.
This means Bard has more access to recent events and facts. Google continues to be a formidable competitor in the AI space.
Plus, it is free to access real-time information (I said that already ). On the Internet, “free” is a powerful business model.