Key Infrastructure Takeaways from Google Cloud Next 2024
There were a mere 218 announcements of new products or services at last week’s Google Cloud Next 2024 conference in Las Vegas, a good portion of those designed for consumer use.
For example, Gemini for Google Cloud now provides users with multimodal tools for integrating text docs, video, images and audio into presentations, enabling people to be speedier about being productive and creative. This high-end bot also serves as a writing and coding coach, an assistant creative designer, an expert adviser, or even a data analyst.
Thus, Google Cloud Next 2024 brought forth some realistic day-to-day benefits for consumers. Nonetheless, this conference was more about enterprise infrastructure than anything else.
AI Hypercomputer and BigQuery
Two of those key infrastructure announcements were around AI Hypercomputer (a workload-optimized infrastructure differentiated from its peers) and Google BigQuery Data Canvas.
“AI Hypercomputer has the broadest and deepest model catalog (enabling access to first-party, third-party and enterprise-ready open source models) all built and optimized,” Gartner VP and Chief Analyst Chirag Dekate told The New Stack. “Gemini for Google Cloud and Gemini for Workspace (that run on AI Hypercomputer) are differentiated and aggregate the infrastructure and Gemini model innovations; they also activate new ecosystem flywheels and an agents’ ecosystem that brings all of the GenAI innovations to life in enterprise context, thanks to the broad partner ecosystem integration.”
How important an advancement is BigQuery Data Canvas (now integrated with Gemini AI)?
“BigQuery Data Canvas is a core-differentiated capability that enables enterprises to create AI-ready data strategy involving discovery, transformation, query and visualization of data using natural language,” Dekate said. “Here is the key impact: Using BigQuery Data Canvas, enterprises can accelerate value creation from their data ecosystem without need for highly specialized skillsets. Gemini-integrated BigQuery enables enterprises to fully utilize both structured and structured data which when coupled with the broader ecosystem creates a disruptive advantage.”
Beyond being a “must have” for cloud developers, Dekate said, “it showcases how Google can leverage its AI DNA to create differentiated enterprise experiences and deliver tomorrow’s ‘must have’ experiences, today.”
AI Coding Assistants Taking off Big Time
By 2028, 75% of enterprise software engineers will use AI code assistants, way up from less than 10% in early 2023, according to Gartner, Inc. Sixty-three percent of organizations are currently piloting, deploying or have already deployed AI code assistants, according to a Gartner survey of 598 global respondents in the third quarter of 2023.
So there you have it: a legitimate software app trend. Google’s response to this is Gemini Code Assist, which has evolved from Duet AI for Developers. Code Assist’s conversational chatbot not only finds the code you’ve been looking for but also offers thoughtful suggestions on alternatives. It supports more than 20 programming languages, and more are being added over time.
“Gemini Code Assist is one of the top coding assistants we’ve tried,” Kai Du, Director of Engineering at Turing, reported on the Google site. “Our early experience with it has been very promising with productivity gains of around 33%. We’re trying out newer features right now like indexing and debugging, which we expect to push productivity even higher.”
Designed to Save Bottom-Line Dollars
CFOs will want to check in on potential time and labor savings as a result of deploying AI coding assistants.
“Calculating time savings on code generation is a good place to begin building a more robust value story,” Philip Walsh, Senior Principal Analyst at Gartner, said in a media advisory. “To convey the full enterprise value story for AI code assistants, software engineering leaders should connect value enablers to impacts, and then analyze the overall return to the organization.”
New features currently in private preview in Gemini Code Assist include:
- Full codebase awareness: This allows developers to perform large-scale changes across an entire codebase, including adding new features, updating cross-file dependencies, helping with version upgrades, and comprehensive code reviews. This capability is powered by Google’s Gemini 1.5 Pro model, which offers a 1 million-token context window, the largest in the industry, Google said.
- Code customization: Enterprises can connect their private codebases to tailor Gemini Code Assist to assist developers in providing context-aware code generation. In addition, Google will provide connections for Gemini Code Assist to reach multiple source-code repositories, including GitLab, GitHub and Bitbucket.
Other Noteworthy Infrastructure News
Cloud Next news also included:
- Google Kubernetes Engine (GKE) with Cloud TPU v5p: GKE now supports Cloud TPU v5p and TPU multihost serving, allowing for more efficient deployment/management of AI workloads in containerized environments.
- Cloud TPU v5p General Availability: Google’s Cloud TPU v5p machine learning accelerators became generally available, offering improved performance for AI workloads.
- N4 Machine Types: These new Compute Engine machine types are built on 5th-Gen Intel Xeon processors and offer a balance of performance and cost for general-purpose workloads.
- A3 Mega Compute Instances: These new instances powered by Nvidia H100 GPUs offer double the GPU-to-GPU networking bandwidth of previous A3 instances, ideal for high-performance computing tasks.
- Confidential Computing for A3 VMs (Preview): Coming later this year, Confidential Computing will be available for the A3 VM family, enhancing security for sensitive workloads on GPUs.
