Yet Another Update in the World of Technology…Google Bard is Now Gemini…
Google Bard is now known as Gemini and is the latest AI tool revolutionizing the workday for literally everyone. While you may be puzzled by the connection between AI, a poet and a zodiac sign, fear not, for the Meta Mind Shift Mastermind is here to guide you through the technological landscape, but more on that later…
Rapid Evolution: From Bard to Gemini to Gemini 1.5
Google has been (to the angst of all of its sphere) unconscionably slow to roll out its emerging tech over the last couple of years, sitting on the sidelines to organizations like Meta, OpenAI and StabilityAI, but lately its begun making up for lack luster performance amongst emerging tech. In fact this year, specifically over the last week, we’ve experienced rapid evolution of its AI capabilities, reflected in a series of updates and rebrandings. First a little history, then the details.
Google first revealed its AI equivalent of ChatGPT, naming it “Bard” in 2022. Bard launched to humdrum reviews in 2022 as Google’s alternative to ChatGPT, enabling natural conversations across over 40 languages. Update #1, Google’s conversational AI is now, after months of improvements, Bard has not only been rebranded as Gemini but also came forward with more advanced capabilities.
The Gemini update also introduced Gemini Advanced, powered by the larger Ultra 1.0 model. Gemini Advanced can handle more complex tasks like coding, reasoning, and creative collaboration. At the same time, Gemini jumped into an app and became widely available through new Android and iOS apps.
But wait theres more! Evolution continues as just a week later, the company unveiled Gemini 1.5 – representing a “dramatic improvement” in AI performance leaving tongues of techies wagging faster than a doggie on a hot day.
Gemini 1.5 Pro introduces a more efficient “Mixture of Experts” (MoE) model architecture, allowing for faster training and deployment. More notably, it supports an experimental 1 million token context window – a 10X increase over the original Gemini. This expanded context enables Gemini 1.5 to process vast amounts of text, audio, video and code in a single prompt.
To understand the revolutionary impact here:
- Original ChatGPT: 2048 tokens (Nov ’22)
- ChatGPT Plus: 4000 tokens (Feb ’23)
- GPT-4 Turbo: 128,000 tokens (Dec ’23)
- Gemini 1.5: Up to 1 million tokens (Feb ’24)
The token count demonstrates how rapidly context sizes are expanding. In just the span of a year, the “state-of-the-art” has gone from 4000 tokens to 1 million tokens – a 250X increase, yes that’s right two-hundred-and-fifty times improvement in processing capability.
Based on Googles release notes, Gemini 1.5 Pro performs on par with the larger Gemini 1.0 Ultra model, despite being more efficient. It also shows impressive skills in analyzing long texts, reasoning across modalities, and problem-solving complex code.
According to Google they are committed to responsible AI development, and say Gemini 1.5 underwent extensive ethics and safety testing before release. (how extensive or who was involved in assuring that extent are the areas of concern) For now, access to Gemini 1.5 is limited to select developers and enterprise customers. But these rapid advancements promise more helpful AI applications for everyone on the horizon.
The evolution from Bard to Gemini to Gemini 1.5 demonstrates Google’s relentless pace of progress in conversational AI…even though we’ve been woefully disappointed with most of their timing, this one seems to be driving the point home that they are still in the game. Each iteration brings significantly more advanced capabilities. If Google is determined to lead the AI race (they’ve got a lot of catching up to do in terms of cleaning the perception that they are behind) we can expect even more groundbreaking updates as they continue to innovate.
NOW THINK ABOUT YOUR BUSINESS
You may be avoiding AI but are your competitors? What would intuitive access to 57 fields of knowledge do for your business? Here you are, manually slogging through data, meetings and emails, while your rivals capitalize on AI’s potential. Did you see Gemini’s Massive Multitask Language Capabilities? Think about realistic use cases in your business where Gemini can handled massive inputs across text, audio, data files, etc. in different languages and formats – and produce actionable outputs for your teams. Perhaps its a key to unlocking new markets across cultures, borders, and countries.
Massive Multitask Language Understanding (MMLU) refers to the ability of an AI model to handle multiple language-related tasks simultaneously on a large scale. This means that the model can perform various language understanding tasks, such as translation, summarization, question-answering, and sentiment analysis, all at once. Essentially, MMLU enables the AI model to comprehend and process language in a wide range of contexts and scenarios, making it highly versatile and capable of tackling diverse challenges in natural language processing.
ChatGPT, Stability AI and now Gemini are already here, they are not going back in the box but they are guaranteed to be changing Google’s core products and the world around you. Digital transformations are already underway at nearly every major company, and small businesses (like yours) are also considering this revolution. Is your workforce ready to professionally embrace AI, or do they only understand the power of Instagram filters and TikTok’s AI?
With each groundbreaking update, organizations including Google are pushing the boundaries of what is possible. Gemini puts advanced capabilities in reach for businesses seeking a competitive edge. The choice is yours – will you lead your company into the future, or allow your rivals to capitalize on AI’s potential? The time for digital transformation is neigh!
Note: Gemini 1.5 Pro comes with a standard 128,000 token context window. But starting today (February 15, 2024), a limited group of developers and enterprise customers can try it with a context window of up to 1 million tokens via AI Studio and Vertex AI in private preview.
