I recently attended the Google Cloud Day ‘24 in Lisbon and got to spend a day exploring Google’s plans for the future. The event featured several keynotes, each focusing on different topics, but all sharing a common theme: Generative Artificial Intelligence (GenAI). The highlight of the event was Google’s project Gemini, their cutting-edge GenAI model.

Gemini is now seamlessly integrated across Google’s suite of services, from Google Docs to their security products. At first, I though it was just another tool for us mere mortals who haven’t mastered the intricacies of technology. However, as I delved deeper into the keynotes, I discovered that there was much more to Gemini than meets the eye.

Most companies building large language models offer a prompt - a place where you can ask questions and receive answers. Recently, these models have also evolved to accept pictures as input, providing answers with a surprising level of confidence (notice I did not say correctness, but models are improving at lightning speed in this regard). This capability alone is a significant leap compared to what we had in the past. Moreover, the speed with which these models respond to our questions is, quite simply, impressive.

Despite these impressive advancements, most projects still require you to go out of your “confort zone” to get your answers. For developers like myself, this means leaving the familiarity of our IDEs to find the information we need. While someone could certainly develop a plugin for VSCode or IntelliJ to streamline this process, making it seamless and integrated into our workflow, there is still room for improvement. But what about everything else? Imagine the benefits if these models were integrated across the entire suite of software you use daily. Picture having this advanced AI functionality embedded in everything from your email and calendar to your document editor and photo manager. Would it not be awesome?

And those are precisely the cards Google is playing. Instead of developing another AI model and providing a prompt for questions, they have done what they do best and have integrated it accross their whole suite of apps, making AI a central part of any modern app.

This sounds good in practice, but is even better when you see it work in practice. The example that made this clear for me was of someone trying to create a new market pitch. At first, they demonstrated how Gemini could generate a polished document with just a few prompts. In less than a minute, a new document had emerged from nothing but a few simple prompts. But what happened next was what I found most impressive. After writing the document, there was also the need to create a presentation for the market pitch. Thanks to the integration with Gemini, creating the presentation was as simple as pointing to the google document that had just been generated and wait for the slides to be generated. In addition, Gemini was also able to generate related images that could be used for marketing purposes.

This demo truly showcased the real power of AI for me: integration across products is everything. It made me realize that the goal is no longer to create the AI model, but to embed these advanced features within the apps we use every day. Achieving this seamless integration is something only a few companies can do, given the extensive ecosystem required. Google’s tight integration of Gemini across its numerous services almost guarantees its success, as it has the potential to become the easiest LLM to use. While ChatGPT currently holds the upper hand in terms of speed and accuracy - at least from my brief experience with Gemini - Gemini’s deep integration into Google’s ecosystem might soon outpace it.

Naturally, this integration also brings up important topics such as security and privacy, which are major concerns as LLMs become more popular. Google has emphasized strong security measures, likely aiming to differentiate itself from competitors like Microsoft, whose recent leaks have raised serious concerns about data protection.

Google appears to be prioritizing security from the start, even going as far as acquiring Mandiant, a leading security company. This acquisition was highlighted at the Google Cloud Day ‘24, showcasing impressive promises for a robust security product. Mandiant is now an integral part of what Google calls Google Threat Intelligence, offering a comprehensive range of security features and benefits to companies. With this extensive offering, Google is trying to position itself as a reliable choice for businesses looking to leverage advanced AI while maintaining stringent security protocols. With such strong integration of AI across its services and a steadfast commitment to security, Google is aiming to lead the way in transforming how we interact with technology in our daily lives, and it certainly seems to be working.

Comments