By Manuel Allegue,  founder y CEO de Zentricx

The advances in generative AI are so rapid that the last Next was just four months ago; however, the industry is transforming non-stop and at a giant pace, a fact I could witness at this 2024 edition.

Thomas Kurian, CEO of Google Cloud, was responsible for presenting the most important developments of recent months, while Sundar Pichai, CEO of Google and Alphabet, reinforced in his keynote how in 2023, the world was just beginning to imagine how generative AI could transform businesses, but today, it is a revolution in progress. “To continue expanding its opportunities, we must see it in action with implementations like agents,” he pointed out.

Indeed, agents are what help users achieve specific goals, from online shopping to healthcare, and they can understand information in different formats: processing video, audio, and text, connecting and streamlining different inputs, learning over time, and facilitating transactions and business processes. Logically, this path toward agent integration is the driving force behind the revolution in AI-optimized infrastructure, models, and platforms, around which most of the announcements revolved.

During the sessions, it was explained how Gemini, Google’s AI, can simplify work tasks by expanding its abilities to create agents capable of coding more efficiently using natural language, managing cloud applications, gaining deeper insights from data, and identifying and resolving security threats more efficiently.

For example, within Google Workspace, Google Vids was introduced, an AI assistant for creating, writing, producing, and editing videos in real-time, with highly useful tools for automatic note-taking, translation, and message summaries.

Gemini 1.5 Pro, for its part—now available in public preview on Vertex AI, Google’s enterprise-focused AI development platform—was presented as “Google’s most powerful and capable generative AI model.” It’s a model that supports 1 million tokens, 11 hours of audio, 30k lines of code, and up to 700K words. This “text-to-image mode” further enhances with live image creation, advanced photo editing, and watermarking.

Another interesting announcement was Vertex AI Agent Builder, a new tool to help companies easily and quickly create conversational agents: “You can instruct and guide them in the same way you do with humans to improve the quality and accuracy of model responses,” Kurian commented. To achieve this, the company uses a process called “grounding,” where responses are linked to an instance considered a “reliable source.” In this case, it’s based on Google Search.

Among the standout innovations is AI Hypercomputer, a “supercomputing architecture” that employs an integrated system of performance-optimized hardware, open software, leading machine learning frameworks, and flexible consumption models. A highlight here is Google Axion, Google’s first custom CPU based on Arm, designed for data centers, which, as explained, “offers 30% better performance than other Arm-based instances from competitors like AWS and Microsoft, and up to 50% better performance and 60% better energy efficiency than comparable X86-based instances.”

Unified databases compatible with multiple clouds were also announced, featuring new vector search capabilities and support for natural language and access to Google Search results, meaning models are trained with the latest and highest quality information from Google Search.

The multimodal capabilities of Gemini, integrated with Google products, have been the common denominator in virtually all sessions, but Google also announced a series of new products and functions centered on AI for cybersecurity, such as “Threat Intelligence,” which can analyze large portions of potentially malicious code and allows users to perform natural language searches for ongoing threats or indicators of compromise.

Beyond the centrality of this event and its announcements, the reality is that generative AI is indispensable for the transformation of organizations, but at Zentricx, we know firsthand that we need to support many organizations in understanding the business impact of these solutions. At this point, it is necessary to reinforce both the theoretical part and the implementation and concrete use cases of generative AI, with the understanding that our clients expect tangible returns and concrete results from their investment in this technology.

By Manuel Allegue,  founder y CEO de Zentricx