This AI cloud: How Google Gemini will help everyone build things faster, cheaper, better (2024)

This AI cloud: How Google Gemini will help everyone build things faster, cheaper, better (1)

When it comes to generative AI, Google has become a major player with its Gemini offerings. Users experience Gemini with every search, it's available as part of Gmail, there are coding and development tools that benefit from Gemini, and more.

I had the chance to talk with Richard Seroter, chief evangelist for Google Cloud, to learn more about where Google sees AI taking us, both at the developer and the consumer level.

Let's get started.

Also:Sorry bean counters: AI should bolster creatives, not replace them

ZDNET: Can you share your journey to becoming the chief evangelist for Google Cloud?

Richard Seroter: It's a long and dramatic story. Actually, it's not. I joined Google Cloud four years ago as our first external "outbound product manager" to engage with customers and internal teams on our app development and modernization products.

I was asked to lead Developer Relations about two years ago, and last year I also got the pleasure of adding our Cloud Documentation team to our group.

Now, I have the privilege of leading a talented group of engineers, tech writers, and product managers who help people find, use, and enjoy Google Cloud.

ZDNET: Can you share some examples of how AI-powered coding assistants have improved productivity for developers?

Seroter: The goal is helping teams ship faster, more efficiently, and with higher quality. AI-powered coding assistants can do this by reducing context switching -- staying in the IDE [the development environment] for more activities thanks to AI chat and inline code generation -- and creating code by expressing intent, not needing to recall every aspect of code syntax.

Also: What is Gemini? Everything you should know about Google's new AI model

A good AI assistant also speeds up onboarding of new skills, helps you find areas of the codebase to improve, and gets rid of repetitive tasks.

For example, using AI assistants to quickly generate database connection code or data objects is a huge timesaver! While developers aren't spending all day coding -- far from it! -- there's a legitimate productivity benefit for junior and senior developers.

We've seen some really great use cases of this with customers including Turing and Commerzbank who both presented at our recent Google Cloud Next conference.

ZDNET: What measures are in place to ensure the reliability and accuracy of AI-generated code? As I showed in my coding test articles, Google Gemini failed several coding tests.

Seroter: While products like Gemini Code Assist are backed by rigorously trained models and a series of filters to check results before returning to the user, the answers aren't always perfect.

Also: I caused Google's Gemini 1.5 Pro to fail with my first prompt

This is one of the reasons we are building (and pre-announced at Next) our Code Customization feature that lets you ground the answers in your private codebase. We will continue improving and refining our underlying models, while taking customer feedback on areas to improve.

ZDNET: How does Google Cloud's partnership with Stack Overflow and other platforms enhance its AI tools? In particular, what are you doing to ensure that the vast amount of incomplete or incorrect information in Stack Overflow is kept out of the knowledge base?

Seroter: The data from all our partners (like Stack Overflow, Snyk, and others) offers additional knowledge that enables us to meet developers where they are and to provide more comprehensive answers to our users' questions.

Also: Google Cloud adds Stack Overflow's knowledge base to Gemini AI

As part of our overall data processing strategy, we filter and evaluate all data contributions, including data from our partners. We use a combination of techniques and tools to ground our responses regardless if a third-party data source is used or not, and we continuously validate and monitor the response quality via automated and manual batteries of tests.

ZDNET: How does Google ensure the security and privacy of customer code when using Gemini Code Assist?

Seroter: Google does not train our model based on prompts entered into Gemini Code Assist. We publish documentation on how we encrypt prompts in transit, and our overall privacy commitment.

Also: Google unveils Gemini Code Assist and I'm cautiously optimistic it will help programmers

We also cite sources wherever possible, provide indemnification, and offer secure access for your perimeter network using VPC Service Controls.

ZDNET: How does Google Cloud address potential biases in AI models used in development tools?

Seroter: We work hard to stick to Google's AI principles and have vigorous protections in place applied during training and response filtering. We also offer multiple feedback mechanisms (in the IDE, elsewhere) for users to flag anything deemed offensive or inaccurate.

ZDNET: Can you discuss the impact of AI on the future of software engineering and development practices?

Seroter: We expect it to positively impact virtually every role in software development and delivery. Teams will use AI-infused systems to analyze data to craft requirements, create prototypes, set up dev environments, write and update code, generate test plans, review code, deploy applications, provision and optimize infrastructure, troubleshoot issues, and secure their systems.

Also: From AI trainers to ethicists: AI may obsolete some jobs but generate new ones

You'll see faster models providing contextual answers -- factoring in local knowledge repos and code bases -- to the full app delivery lifecycle. Our practices will have to keep up as we think of creation and curation activities, need to serve up AI-friendly platforms to build teams, and even how we test non-deterministic systems.

Research shows that developers are looking for AI to help make engineering more efficient; it's not about fundamentally changing workflows. At least not yet, but stay tuned.

ZDNET: How does Google Cloud's AI technology help in managing and reducing technical debt in software projects?

Seroter: Technical debt comes from many directions, and sometimes is "good" debt accrued by teams getting value to market. But AI-assisted tooling and AI overall can help teams apply best practices earlier, and fix existing debt faster.

A well-trained AI assistant like Gemini Code Assist can generate and validate code as the developer goes along, ensuring a limited number of compromises and future debt. For existing codebases, our huge context window in Gemini 1.5 lets teams explore full codebases in the hunt for issues to resolve.

ZDNET: What future advancements do you foresee in AI-assisted development tools on Google Cloud?

Seroter: We're excited to bring to market the items we announced in preview at our recent Google Cloud Next conference. The full codebase awareness that comes from offering Gemini 1.5 as a base model means teams can perform complex modernizations or code exploration at unheard-of speed.

Also: Meet Gemini 1.5, Google's newest AI model with major upgrades from its predecessor

And the ability to customize responses based on code in GitHub, GitLab, or Bitbucket means that teams can get even more trust and context from their AI-generated results. Look for more grounding options that help developers get timely and relevant assistance.

We also see that AI-assistance doesn't just live in the IDE, or benefits developers. The overall Gemini in Google Cloud investment brings AI assistance to BigQuery users trying to generate complex queries, Cloud SQL users explaining a massive store procedure, security pros analyzing threats, developers creating low-code integrations or APIs, and so much more.

An AI-assisted cloud helps people "build" all sorts of things faster, cheaper, and with higher quality.

ZDNET: Can you discuss the role of AI and machine learning in enhancing cloud services?

Seroter: While I get excited about building software, most software spends its days being operated. We announced Cloud Assist as part of Gemini in Google Cloud, and this upcoming service will transform how teams manage their cloud services.

Also: Meet Google Threat Intelligence, Google Cloud's security solution with Gemini Pro

From personalized optimization suggestions to helping people pinpoint problems and get a system back online faster, tools like this could fundamentally change how you operate (cloud) services.

At the same time, Gemini in Google Cloud is designed to make the overall cloud easier to use. Getting AI-generated threat summaries in Security Command Center is powerful. Seeing AI-assisted log summaries in Cloud Logging makes the product easier to use.

Offering an always-present AI chat window in our Cloud Console means you don't have to context switch to ask a product question or get clarity on a CLI command. These sorts of experiences provide help where you're at, which stands to make a big impact on day-to-day cloud use.

ZDNET: What are some innovative use cases of Google Cloud that have impressed you recently?

Seroter: I'm inspired by the broad research-based work of Google, and also how customers apply technology to solve their specific problems. Recent Google work with AlphaFold 3 predicts the structure and interactions of the molecules of life. That's remarkable. We're mapping new aspects of the human brain. Awe-inspiring stuff.

Companies like Chugai Pharmaceutical take advantage of some of our research by deploying cloud systems to accelerate drug discovery. I'm impressed by teams from Aviator looking at helping developers become more productive across the entire dev lifecycle.

And companies like Goldman Sachs are democratizing data access by open-sourcing their data platform. Google Cloud customers solve big problems and constantly make outsized impacts in their industries.

Also: Why AI solutions have just three months to prove themselves

What do you think?

Are you using Gemini now, either with search or email, or as part of your coding process? Did Richard's answers help you understand more about Google's perspective on AI? Let us know in the comments below.

You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at, on Instagram at, and on YouTube at

Artificial Intelligence

  • Transparency is sorely lacking amid growing AI interest
  • What is a Chief AI Officer, and how do you become one?
  • How Adobe manages AI ethics concerns while fostering creativity
  • 6 ways OpenAI just supercharged ChatGPT for free users
This AI cloud: How Google Gemini will help everyone build things faster, cheaper, better (2024)
Top Articles
Latest Posts
Article information

Author: Pres. Lawanda Wiegand

Last Updated:

Views: 5887

Rating: 4 / 5 (51 voted)

Reviews: 82% of readers found this page helpful

Author information

Name: Pres. Lawanda Wiegand

Birthday: 1993-01-10

Address: Suite 391 6963 Ullrich Shore, Bellefort, WI 01350-7893

Phone: +6806610432415

Job: Dynamic Manufacturing Assistant

Hobby: amateur radio, Taekwondo, Wood carving, Parkour, Skateboarding, Running, Rafting

Introduction: My name is Pres. Lawanda Wiegand, I am a inquisitive, helpful, glamorous, cheerful, open, clever, innocent person who loves writing and wants to share my knowledge and understanding with you.