More How I AI episodes

What launched at Google I/O 2026 (30-minute day 1 recap) thumbnail

What launched at Google I/O 2026 (30-minute day 1 recap)

Published 20 May 2026

Duration: 00:33:52

Google's recent AI advancements highlight agentic AI capabilities, the Gemini 3.5 model family (including fast multimodal Flash), Antigravity IDE 2.0 for coding, and creative tools like video-generation Omni and design apps Stitch/Pameli, alongside noted technical and usability challenges.

Episode Description

Today is day one of Google I/O 2026, and I walk through every major announcement livefrom the new Gemini 3.5 model family to Anti-Gravity 2.0, Google...

Overview

The podcast focuses on recent advancements in AI tools and features unveiled by Google, emphasizing their applications for developers, creative professionals, and consumers. Key highlights include the launch of the Gemini 3.5 model family, notably Gemini 3.5 Flash, which demonstrates superior speed and performance in multimodal tasks and agentic reasoning. This model is tailored for developer workflows, such as agentic coding frameworks and integration with IDEs like Antigravity IDE 2.0, which now includes sub-agents, Git integration, slash commands, and a CLI for terminal-based coding. Google also introduced Gemini Omni, a video-generation tool capable of creating photorealistic content from references, along with interactive editing features for refining videos through text prompts and character customization.

Additional tools like Pameli (for brand and marketing content generation) and Stitch (a browser-based design tool with real-time AI editing) were discussed, highlighting standardized design systems and streamlined workflows. However, the podcast also addresses challenges, including technical limitations in AI-generated image and video quality, incomplete features like non-functional avatar tools, and confusion arising from overlapping product names and branding. Googles efforts to integrate AI with Workspace services (e.g., calendars, docs) and no-code app development via AI Studio were noted, though early access and feature readiness remain limited. The discussion underscores the tension between ambitious feature promises and the need for refinement in usability and consistency across tools.

The overview also touches on broader trends, such as Googles focus on speed and consumer applications, while adopting concepts from competitors like Anthropic and OpenAI. Despite the advancements, gaps in product maturity, inconsistent user experiences, and the complexity of AI tool discovery were flagged as areas requiring improvement. The tools examined span prototyping, design, and engineering, reflecting a push to unify workflows through agentic AI capabilities, though some features remain experimental or underdeveloped.

What If

  • What if you leveraged Antigravity IDE 2.0's scheduled tasks to automate code reviews and documentation updates?

    • Move: Integrate Antigravity's cron-based scheduled tasks with your codebase to run automated prompts for code reviews, documentation generation, or dependency checks.
    • Why now: Antigravity's updates (e.g., scheduled tasks, hooks) and Gemini 3.5 Flash's speed make this feasible for solo developers to offload repetitive tasks.
    • Expected upside: Reduce manual overhead, ensure code quality consistency, and free up time for strategic work.
  • What if you built a no-code/low-code app in Google AI Studio that integrates with Google Workspace (e.g., Gmail, Sheets) to automate client onboarding?

    • Move: Use Google AI Studio's no-code tools to create an app that pulls data from Google Workspace (e.g., client emails, project timelines) and automates onboarding workflows.
    • Why now: Gemini 3.5 Flash's performance and AI Studio's integration with Workspace services enable rapid prototyping and deployment.
    • Expected upside: Streamline client onboarding, reduce errors, and position your app as a competitive alternative to third-party tools.
  • What if you used Antigravity's /goal slash command to define and track long-term project milestones in your solo development workflow?

    • Move: Define a specific goal (e.g., "Deploy a full-stack app by Q3") using Antigravity's /goal command and let the agent autonomously break it into subtasks.
    • Why now: Antigravity's agentic capabilities (sub-agents, hooks) and Gemini 3.5 Flash's reasoning power align with solo developers' need for task automation.
    • Expected upside: Maintain focus on high-priority tasks, ensure progress tracking, and reduce burnout from micromanaging deadlines.

Takeaway

  • Leverage Gemini 3.5 Flash for faster code generation: Integrate the model into your development workflow to accelerate prototyping, especially for tasks involving multimodal data (e.g., handling files, videos) or agentic coding frameworks.
  • Adopt Antigravity IDE 2.0 features for automation: Use scheduled tasks (cron-based prompts), slash commands (e.g., /goal, /grill me), and sub-agents to streamline repetitive coding tasks and improve project management.
  • Utilize Google AI Studio for no-code/low-code app development: Build productivity tools or personal apps with Google Workspace integration (e.g., calendar, Docs) using Gemini 3.5 Flashs speed and compatibility with native Git workflows.
  • Standardize design systems with design.md: Use tools like Pameli or Stitch to generate consistent brand and marketing assets by encoding design systems in markdown, ensuring alignment across AI-driven outputs.
  • Provide feedback on early-stage tools: Test Googles AI video tools (e.g., Gemini Omni, Flow) and report usability issues (e.g., avatar reliability, incomplete features) to influence product refinement and prioritize needed improvements.

Recent Episodes of How I AI

22 Jun 2026 How Claude Mythos found a 15-year-old bug in Mozilla Firefox | Brian Grinstead

Recommended: AI finds bugs

Firefox employs AI agents as "coding archaeologists" to detect and address security vulnerabilities in its massive codebase, leveraging models like Mythos and custom validation tools to identify and systematically fix nearly 500 bugs, while balancing automation with human oversight and open-source collaboration to enhance scalability and security.

15 Jun 2026 How Braintrust uses AI agents, evals, and CI to ship better software | Ankur Goyal

AI integration in software engineering enables agents to handle complex tasks through benchmarking and optimization, shifts engineers toward higher-level work, and addresses challenges like reliability, data parsing, and balancing automation with human expertise while emphasizing outcome-focused systems over procedural methods.

9 Jun 2026 Claude Fable 5 review: what the new Mythos model gets right (and very wrong)

Anthropic's Claude Fable Five excels in long-term technical tasks with strong coding, vision, and async workflow capabilities but faces high token costs, design limitations, and restricted use in cybersecurity/biology, making it suitable for precise, extended projects rather than creative or agile workflows.

More How I AI episodes