Sudipta Deb

Sudipta Deb

Founder of Technical Potpourri, Co-Founder of Shrey Tech, Enterprise Cloud Architect

Welcome back, tech aficionados! Day two of Google Cloud Next 2024 is done, and the excitement was palpable as developers from around the globe gather at the illustrious Mandalay Bay Convention Center in Las Vegas. My focus shifts to the Developer Keynote, where Google unveiled a treasure trove of new announcements set to revolutionize the world of software development.

From innovative tools and platforms to groundbreaking updates on existing technologies, the Developer Keynote promised to be a showcase of ingenuity and creativity.

Join me as I unpack the highlights of the Developer Keynote, diving deep into the latest offerings from Google that are poised to shape the future of software development. From AI-driven enhancements to developer-friendly features, there’s something for everyone in today’s announcements.

So, grab your laptops and get ready to embark on a journey into the heart of developer innovation. Together, let’s explore the cutting-edge advancements that are propelling us towards a future limited only by our imagination. Welcome to day two of Google Cloud Next 2024 – where the code meets the cloud, and innovation knows no bounds!

Developer Keynote

The Developer Summit Keynote on Day 2 further fueled the momentum behind Google Cloud Next 2024.

The developers and architects of Google Cloud solutions were the target demographic that the Developer Keynote specifically targeted. And the meeting did not let me down!

Here’s a recap in case you missed it.

Build: Generative AI Across The Developer Workflow

The standout feature is Gemini Code Assist, which allows AI-powered code completions and now supports a context window with one million tokens. Through the analysis of a developer’s unique cloud resources, such as logs and configurations, Gemini Cloud Assist applies this AI support to cloud operations.
Google is integrating generative AI into BigQuery and other tools to analyze data with natural language and multimedia inputs. Vector search and unstructured data embedding are being added to its databases. In a demonstration, an AI app that generated floor plans, suggestions, and ideas for house design in response to multimedia cues was built in a matter of minutes.
In order to increase developer productivity, Google wants to integrate contextualized, configurable generative AI into data analytics, ops, coding, and application development.

New things that makes this possible:

  • App Hub – Regardless of the particular Google Cloud products that are used, App Hub—which was unveiled on Day 2—offers a precise, current depiction of deployed applications and their resource dependencies thanks to its strong connection with Google Cloud Assist.
  • BigQuery continuous queries – BigQuery can now process SQL continuously over data streams in preview, allowing for real-time pipelines with AI operators or reverse ETL.
  • Natural Language support in AllowDB – Support for Google’s cutting-edge ScaNN algorithm gives AlloyDB customers access to improved vector performance, which underpins a number of Google’s most well-known services.
  • Gemini Code Assist in Apigee API Management – Utilize Gemini’s natural language prompts to assist you in developing enterprise-grade APIs and integrations.

Run: Deploying And Scaling Generative AI

With GKE for reliable Kubernetes-based AI programs and Cloud Run for quick serverless deployments, Google Cloud makes it possible to launch and scale generative AI applications. The latest additions include Quick Start options, GKE connections for running big language models like Gemma, and Cloud Run Application Canvas for quickly creating AI apps.

New things that make this possible:

  • Cloud Run application canvas – Create, edit, and implement AI apps in Cloud Run. Vertex AI integrations allow you to quickly and easily use generative APIs from Cloud Run services.
  • Gen AI Quick Start Solutions for GKE – Utilize AI on GKE by integrating it with Ray or using a Retrieval Augmented Generation (RAG) method.
  • Support for Gemma on GKE – Gemma is Google’s open model built on Gemini, and GKE provides multiple ways to operate it. Even better, the performance is superb.

Operate: Handling AI Behaviors

Emergent behaviors from AI apps can lead to new problems, Google Cloud Reliability Advocate Steve McGhee stated. “Dynamic and chaotic” is how Charity Majors described contemporary systems in contrast to dependable heritage systems.

But generative AI also offers fresh resources for comprehending and handling change:

  • Prompt Management and Rapid Evaluation are two Vertex AI MLOps that are useful for experimenting with prompts and evaluating the performance of models.
  • To identify unregulated APIs that can pose security problems, use shadow API detection.
  • Confidential Accelerators provide AI/ML workloads with hardware data protection.
  • To maximize GPU utilization and speed up cold starts, preload models and GKE containers.

New things that make this possible:

  • Vertex AI MLOps Capabilities – While Vertex AI Rapid Evaluation assists users in assessing model performance when iterating on the optimal prompt design, Vertex AI Prompt Management allows customers to experiment with migration and tracking of prompts and parameters in preview, allowing them to compare prompt iterations and assess how small changes impact outputs.
  • Shadow API Detection – Shadow API identification in Advanced API Security preview assists you in identifying APIs that lack adequate governance or supervision and may be the cause of harmful security incidents.
  • Confidential Accelerators for AI workloads – The A3 machine series’ confidential virtual machines, which use NVIDIA Tensor Core H100 GPUs to manage sensitive AI and machine learning data, expand hardware-based data and model protection from the CPU to the GPU.
  • GKE container and model preloading – GKE can now accelerate workload cold-start in preview to maintain low AI inference latency, increase GPU usage, and save costs.

Conclusion

As I wrap up my coverage of day two at Google Cloud Next 2024, the air is buzzing with excitement and anticipation for what’s to come. The Developer Keynote has left me inspired and eager to dive deeper into the world of software development, armed with the latest tools and technologies unveiled by Google.

I promise to continue bringing you comprehensive recaps of each day’s events, including all the groundbreaking announcements and insights that emerge. So stay tuned, as I am preparig the recap of day three of Google Cloud Next.

In the meantime, let’s reflect on the wealth of knowledge and inspiration gained from Day 2’s Developer Keynote. From AI-driven enhancements to developer-friendly features, the possibilities seem endless, and the future looks brighter than ever.

So, until next time, fellow developers and tech enthusiasts, let’s continue to embrace the spirit of innovation and collaboration as we journey towards a future limited only by our imagination. Stay tuned for more exciting updates from Google Cloud Next 2024!

Disclaimer

This article is not endorsed by Salesforce, Google, or any other company in any way. I shared my knowledge on this topic in this blog post. Please always refer to Official Documentation for the latest information.

0 Comments

Submit a Comment

Your email address will not be published. Required fields are marked *