Everyone’s heard of ChatGPT, but that’s certainly not the only option when it comes to free, publicly-available, generative large language models. If you work in tech, we think you should heavily consider using Google Bard, especially now that they’ve updated the platform to be able to generate code. But, that’s not the only thing it can do—the Google Bard AI directly integrates with many other Google services, allowing you to directly export code to Colab, export a summary to Docs, or even send a request to gmail.
How to Get Access to Google Bard
First things first! Before you can use Google Bard, you need to meet just three basic requirements:
Have a personal Google account
Be 18 years or older
Use Chrome, Safari, Firefox, Opera, or Edgium
If you meet all three requirements, then just navigate to https:/bard.google.com and click “Sign in” in the top right corner. Earlier this year, there used to be a waitlist in order to use Bard, but we haven’t encountered any waitlist in recent tries. With that said, if you do end up on the waitlist, don’t worry; in our experience you’ll be granted access in the next 2-3 days.
If you don’t meet all the requirements, there’s still a chance you can use Google Bard. Let’s go over solutions in order:
If you don’t already have a personal Google account (where have you been?!), you can make one right now by clicking this link.
If you’re not 18 years or older, then unfortunately you’re out of luck. You can’t use Google Bard with a Google account managed by Family Link or with a Google Workspace for Education account designated as under the age of 18. This is a strict requirement, and they don’t make any exceptions.
If you don’t normally use a Chromium-powered browser (or Safari), Google will deny you access to Bard. Like Microsoft with their AI-powered Bing, these companies want you to use their own software to gain access to “experimental” features. Fortunately, browsers are super easy to download and, when in doubt, download Firefox.
What is Google Bard?
Now that you know how to access Bard, we’re sure the next big question on your mind is: “What does Bard stand for, Google?” Well, we’re here to answer all questions, no matter how big or small. Google doesn’t usually publicize why it names products the why it does, and this time is no different--but we’ll take a crack at it anyways.
A bard is an old English word for a poet or storyteller---someone who creates art with language. Since Google Bard is a generative large language model, the name “Bard” is a play on the fact that this is software that creates responses using natural language.
Google Bard vs ChatGPT
There’s two big players in the generative language model space: ChatGPT by OpenAI and Bard by Google. But hold up, we know what you’re thinking—what about AI-powered Microsoft Bing? Isn’t that another chatbot that’s super similar to Google Bard? Yes, you’re right, they are similar, but, AI-powered Microsoft Bing actually uses ChatGPT 4.0.
Google Bard and ChatGPT have a lot in common: they’re both generative large language models and can be used to do similar things from simple chatbot-like tasks to more complex tasks like generating code. Here’s how the two products match up head-to-head:
Was created by Google
Uses the Language Model for Dialogue Applications (LaMBDA)
Sources dataset from Common Crawl, articles, books, Wikipedia, and real-time searches
Was created bv OpenAI
Uses the Generative Pre-Trained Transformer (GPT) Model
Sources dataset from Common Crawl, articles, books, Wikipedia, and real-time searches (only available on GPT-4, which is paid).
GPT-3.5 is free, but users must pay a subscription to use GPT-4
While Bard was released after ChatGPT, Google has a lot of experience with natural language processing from work associated with improving Google Search and Google Assistant. In fact, Google Research invented and open-sourced the neural network architecture that’s used by many language models, including GPT 3.5 and GPT 4.
How to Use Google Bard AI
Now that we know all about what Google Bard is and where we can access it, we should probably tackle the question, “Well, how do I use Google Bard?!” You’re in luck! There’s no right or wrong way to use products like Bard. It’s a language model that’s built to have human-like conversations; if you can give it a task in words, it can probably do it (or at least help you accomplish it yourself). But, we know that’s really broad—it’s probably easier to understand if we split the things that Bard can do into three categories: explanatory, conversational, and creative. We’ll give a brief explanation of each category as well as an example question or task along with a snippet of Bard’s answer.
Explanatory: Bard can explain all kinds of things from giving the definition of a word to answering challenging and open-ended questions about life or the universe.
Q: “Who painted the Mona Lisa?”
A: “The Mona Lisa was painted by Leonardo da Vinci, an Italian polymath of the High Renaissance who was active as a painter, draftsman, engineer, scientist, theorist, sculptor and architect.”
Conversational: Bard can engage in a conversation with you on practically any subject. It’ll remember your responses and be able to impress you with how it can handle complex topics thoughtfully.
Q: “You are a web developer who was teleported to the 18th century. What's the first thing you would do?”
A: “[The] first thing I would do is try to find a way to communicate with people from my own time…”
Creative: Bard can create poems, narratives, musical pieces, emails, and even code.
Q: “Write a haiku about desert sunsets.”
A: “Sunset blazes red, Ashes of day fade to night, Desert sky aglow.”
As you can see, Bard can do basically anything you ask of it—but it does have several limitations. Let’s discuss:
For starters, Bard does not fact-check itself when giving answers. We cannot overstate this enough: generative AI should never be used as a source of truth. For example, in the “creative” section above, we showed Bard’s answer to the prompt, “Write a haiku about desert sunsets.” In the answer (and feel free to check for yourself), Bard gives us a poem with three lines of 5, then 7, then 5 syllables. This is great! It’s indeed a haiku; however, that was not Bard’s initial answer. This is what we originally received:
Sunlight fades to red,
The desert sky ablaze with color,
A perfect end to day.
If you count the syllables in this poem, we have 5, then 9, then 6 syllables in the lines, which is not what we asked for! Ouch, a Google Bard error. Only after we informed Bard that the second and third lines had too many syllables did it come back with a poem in the correct format.
As we’ve already mentioned, Bard can do basically anything you ask. Obviously, it’ll do some things better than others—but it’ll always try its best. And, Google is continuously updating Bard so that it’s able to do more things. Here are a few notable changes:
Bard can now generate code in languages such as C++, Go, Java, and Python
When generating Python code, you’re able to export the code directly to Google Colab
Bard can export content directly to Google Docs and Gmail
Bard can cite sources directly in the response
Bard can display images
Bard is Google’s most recent product release and we think it’s also one of the coolest project’s in the company’s history. Google has been working on generative large language models for decades now, and it’s great to finally see and interact with a product that’s the culmination of so much time, effort, and care.
If you’re interested in working in the tech industry and creating awesome products that help make the world a better place, then we think you would be a perfect fit for one of Ironhack’s bootcamps. We offer bootcamps for just about every role in tech, and there’s never been a better time to join the industry. Want to learn how to make the most of ChatGPT, Google Bard, or other generative language models? We can’t wait to see you in class!