Bard becomes Google Gemini Pro 1.0. Drink the Cool AId

Last Modified: Feb 14th, 2024 - Category: Artificial Intelligence, News
Cover image for the article Google Bard becomes Google Gemini PRO 1.0

So the battle for AI is getting tougher, and Google can’t stay out of it. After a somewhat ignored launch of Google Bard, which met with little interest, Google renamed the Bard project to Google Gemini PRO 1.0.

Will this renaming turn Google’s AI service into a competitor? And can the new PRO live up to the expectations placed in it? That’s what we try to find out in this article.

From Bard to Google Gemini PRO

Google launched Bard a year ago as a heavyweight that would dominate the AI chat business. And with good reason. Let’s be clear: Google was one of the first (if not the first) company to use AI on a large scale. But let’s not kid ourselves: Whatever you see (or perceive) about OpenAI… Google is much bigger than that. And much more skilled.

And believe me, that was (and still is) a big problem for OpenAI, despite its huge success, especially with ChatGPT. If you don’t believe me, check it out for yourself:

Image showing you can't ask Chat GPT about Google Bard
Surprise: you can’t ask Chat GPT about Google Bard

I tried all kinds of prompts related to Google Bard and Chat GPT rejected them all. Whether it’s writing articles, translating, or editing grammar, Chat GPT throws a Moderation Error. Perhaps a momentary hiccup, but at the time of writing this article, this is what I get.

So how much of this rebranding is about dropping the poetry and becoming a boring (but effective) PRO? What does the Gemini PRO 1.0 rebrand mean for AI Chat users? Is Google finally unleashing its quantum supercomputer? Many questions that we’ll answer below.

Google AI Chat: a story of secrecy and failure.

As we have already mentioned, Google was already working with AI before ChatGPT. The experiments with AI started around 2008, although there are versions that say they actually started in 2004 or 2005. If you compare that to the founding of OpenAI in December 2015, you can see the difference between the two projects.

To give a little history, let’s start with the first KNOWN (I mean public) AI project from Google. This project, called Google Brain, was open to the public but was mainly used by academic users as it required a certain level of expertise. Nevertheless, it was the driving force behind many of Google’s product improvements, including Google Translate, TensorFlow, Image Recognition and medical applications. Not too shabby, huh?

Google goes Quantum

in 2006, Google scientist Hartmut Neven began investigating how quantum computing could help the company accelerate machine learning. In 2013, Google acquired its first quantum supercomputer from NASA. This was a major milestone in Google’s AI projects, as computing power rose to unprecedented levels.

While it’s possible that some of that supercomputing power was used for publicly available AI, it’s doubtful. As far as I know, Google has never reported on the development of AI based on quantum supercomputers, which is odd.

DeepMind Acquisition

in 2014, Google acquired the British startup DeepMind, which specialized in AI research and development. DeepMind then became Google DeepMind, which became the vehicle for everything AI-related, especially the products aimed at the general public. More or less what OpenAI was to do a few years later.

If Brain was the great-grandfather of Bard, Google DeepMind is its grandfather. And it’s alive and kicking. In fact, both Bard and Google Gemini Pro 1.0 are developed by Google Deepmind.

Image featuring some of the upcoming projects by DeepMind
Some of the upcoming projects by DeepMind (watch out for Phenaki!)

On a related side note, our CEO Fabio Devin created the UX framework known as Quantum UX by the end of 2014. This framework extensively uses the emerging AI models, though at that time, it was limited to simple models fueled by Google Analytics and supplemented with actual user research.

AI Modelling on Steroids: Introducing BERT

No, we’re not talking about the Sesame Street character. BERT is an acronym that stands for the language model Bidirectional Encoder Representations from Transformers (or BERT for shorts).

And what is BERT, you may ask? Well, even if it’s not very well known to the public, it’s the game changer between the old “dumb” AI and the new realistic models. This is especially important when developing AI chats. I’ll explain it with an example:

Imagine teaching a child a new language. Before BERT, you used to show them individual words without explaining how they belong together. BERT has changed that by showing whole sentences and stories, promoting an understanding of the meaning and relationships between words.

Here you can find out how BERT has influenced AI development:

Significant advancement in Language Comprehension: BERT was one of the first models to understand the meaning of words in the context of a sentence, much like humans. This was a significant advance in Natural Language Processing (NLP), the technology that enables machines to understand human language.

Opened Up New Possibilities: BERT’s improved understanding of language led to the development of various advanced AI applications. Chatbots became more conversational, language translations more accurate and even writing aids more intelligent.

Paving the way for Future Models: BERT may not have been the only groundbreaking model, but it highlighted the effectiveness of this new approach to language modeling. Many models that are popular today were inspired by and developed from the concepts of BERT.

Thus, BERT was a key innovation in AI and linguistic processing, enabling machines to understand and process language in novel and improved ways.

Google Bard: Shakespeare Would Have Hated It

Image of Shakespeare as a robot created with Google Gemini
Shakespeare as a robot created with Google Gemini’s AI chat

BERT enabled new ways of improving language generation that made many other tools like Google Bard possible in the first place.

Google Bard was a little late to market and was seen as a late response to the dominance of Chat GPT and OpenAI in an area that had always been Google’s own territory. And while it was pretty disappointing, it was the awakening of the giant. Yes, Bard had issues and wasn’t a real competitor to ChatGPT, at least not in version 4.0 (the paid one). One can argue that it was very similar to the free version 3.0 of ChatGPT and one should compare the free version to the free version rather than the free version to the paid version. But still, they were too late and it wasn’t enough.

In a very interesting article from Forbes, the sentiment about Bard was pretty much summarized (I edited it to include only the key points since it’s a large article):

‘AI First’ To Last: How Google Fell Behind In The AI Boom

Key Points:

  • Google declared itself an “AI-first” company in 2016, but rivals like OpenAI are now outpacing them.
  • Google’s AI ethics controversies and internal issues may have slowed down their progress.
  • Microsoft invested in OpenAI, giving them access to valuable data and tools.
  • Google released their own chatbot Bard and plans to integrate AI into their search engine.
  • The company faces the “Innovator’s Dilemma” of protecting existing products while innovating.
  • Whether Google can catch up is uncertain, but they still have potential due to their resources and history.

Additional Notes:

  • The article highlights Google’s early contributions to AI technology, like the “Attention Is All You Need” paper.
  • Some of the authors of that paper left Google to build competing companies.
  • Google’s cautious approach to releasing AI products due to potential risks is criticized.
  • The future of search engines and AI development is uncertain, with players like Google, Microsoft, and OpenAI all vying for dominance.

Welcome to the Gemini PRO era

With the recent launch of Gemini Advanced, Google unveiled its most sophisticated AI model to date, Ultra 1.0. This new offering, which was favored over the competition in blind evaluations, promises a leap in chatbot interaction, especially for complex tasks such as programming and creative activities. Gemini Advanced is able to better understand the context of previous prompts, enabling more in-depth conversations and personalized assistance, such as tailored tutoring and advanced programming support.

The launch of Gemini Advanced is a strategic move for Google to incorporate AI into the everyday use of technology and promises future enhancements in terms of multimodal interactions and advanced programming capabilities. Currently, Gemini Advanced is available in English in 150 countries, but there are plans to expand language options. Gemini Advanced is part of the Google One AI Premium plan, which costs $19.99 per month after a two-month trial and includes 2TB of storage and integration with Google Workspace.

Google has focused on the ethical use of AI with the launch of Gemini Advanced and has undertaken extensive safety testing and model improvements based on external feedback. The Gemini app is designed to provide users with a versatile AI assistant that can perform tasks ranging from image creation to the creation of complex text messages.

For Android users, the Gemini app is integrated with the Google Assistant and enhances the user experience with AI-powered assistance directly onthe screen. iOS users can look forward to similar features in the Google app. As the rollout begins in the US and expands globally, Google is asking for user feedback to further improve Gemini Advanced offerings.

What are the differences between Google Bard and Gemini PRO 1.0?

Having said that, it’s now time to highlight the differences between Google Bard and Google Gemini PRO 1.0.

One of these differences, which is probably of interest to most users, is that Gemini will be paid. You can still try it for free, but after 60 days you’ll have to pay and buy the Google Gemini Advanced package, which basically includes the features of Google One plus Gemini. If Gemini is as good as ChatGPT, that’s a good deal if you ask me.

As for the “improvement” of the user interface, it is really moot from a UX point of view. It’s no better (and no worse) than any other AI interface. I could list 4 or 5 improvements they could make without thinking too much, so I guess they either have technical limitations or were just lazy.

Here you have a simple and clear insight into the differences between Bard and Gemini PRO. One thing to notice: unlike ChatGPT, which offers a free version (known as 3.0) and a premium version (known as 4.0), it looks like Bard will no longer exist and Google will only offer Gemini. Hence, no free AI from Google!

FeatureBardGemini PRO 1.0
CapabilitiesBasic conversation, simple tasksComplex tasks, coding, long-form content creation, nuanced reasoning
AvailabilityFree, English onlyPaid subscription, Early Access, more than 40 languages.
FocusText-based conversationsWide range of tasks, including creative text formats, translation, content creation, information retrieval
Data accessSmaller datasetLarger dataset
User interfaceSimpleMore options and customizations

Conclusion: is Gemini Advanced a step forward?

That is a good question. As always, we’ll test it extensively and see what happens. For now, the announced changes are a huge improvement over Bard, so I really hope it works as expected.

One thing I’m thinking about: unlike ChatGPT, which has limited data, Google has all the data of the web, so technically they should be able to provide better and faster results than ChatGPT. On the other hand, it’s possible that the huge amount of data will become a bottleneck if they don’t use the quantum supercomputer to process the results. So we’ll have to wait and see. But I’ve really high hopes for Gemini Advanced!

We can improve your business!

Let us help you with the best solutions for your business.

It only takes one step, you're one click away from getting guaranteed results!

I want to improve my business NOW!