Skip to content
How I AIHow I AI

Was anything good released at Google I/O 2026?

Today is day one of Google I/O 2026, and I walk through every major announcement live—from the new Gemini 3.5 model family to Anti-Gravity 2.0, Google AI Studio, Gemini's consumer redesign, the Omni video model, Flow, Stitch, and Pomelli. I test them in real time and tell you exactly which ones delivered. *What you’ll learn:* 1. How Gemini 3.5 Flash benchmarks against Claude and GPT models on speed and agentic coding tasks 2. How Anti-Gravity 2.0's new features (projects, scheduled tasks, subagents, slash commands) compare to Codex and Claude Code 3. Why the /grill-me slash command could be a more aggressive alternative to Claude Code's clarification flow—and how to use it 4. How Google AI Studio's new Workspace integration is designed to own the internal productivity app use case 5. How Google's new creative tools work in practice: Omni (video generation), Flow (cinematic video editing and character consistency), Stitch (streaming UI design with inline edits), and Pomelli (brand identity and asset generation) 6. Why Google's launch-to-availability gap is still a problem—and what to do when a featured product doesn't actually work yet *Brought to you by:* Magic Patterns—Prototypes that look like your product: https://magicpatterns.com/howiai Thoughtspot—Build AI-powered analytics into your product: https://go.thoughtspot.com/howIAI *In this episode, we cover:* (00:00) Google I/O 2026 day 1 overview (01:47) Gemini 3.5 flash (04:19) Antigravity updates (06:32) CLI test and agent features (07:59) Core agent features released today—May 19th, 2026 (09:43) New slash commands (11:20) Antigravity test results and takeaways (12:25) AI Studio updates (13:52) Access issues (15:20) Gemini redesign (17:24) Gemini image gen test (19:16) Omni (video generation) (22:56) Flow (cinematic editing) (24:31) Avatar creation test (26:45) Pomelli and Stitch (31:13) Recap and final thoughts *Blog:* How I AI: My Live Test of Google I/O’s New AI Tools—From Gemini 3.5 Flash to Omni Video: https://www.chatprd.ai/how-i-ai/google-io-new-ai-tools-gemini-35-flash-to-omni-video *Tools referenced:* • Gemini 3.5 Flash: https://deepmind.google/technologies/gemini/ • Antigravity: https://antigravity.google/ • Google AI Studio: https://aistudio.google.com/ • Google Gemini: https://gemini.google.com/ • Omni (video generation): https://gemini.google/overview/video-generation/ • Google Flow: https://flow.google/ • Stitch: https://stitch.withgoogle.com/ • Pomelli (Google brand tool): https://labs.google.com/pomelli/about/ *Other references:* • Google I/O 2026 announcements: https://blog.google/innovation-and-ai/sundar-pichai-io-2026/ *Where to find Claire Vo:* ChatPRD: https://www.chatprd.ai/ Website: https://clairevo.com/ LinkedIn: https://www.linkedin.com/in/clairevo/ X: https://x.com/clairevo _Production and marketing by https://penname.co/._ _For inquiries about sponsoring the podcast, email jordan@penname.co._

Claire Vohost
May 20, 202633mWatch on YouTube ↗

CHAPTERS

  1. 0:04 – 1:35

    Day 1 I/O gut check: what’s live, what’s hype, and what Claire will test

    Claire frames the episode as a quick tour through the most notable Google I/O Day 1 announcements, moving from technical model/agent updates to creative consumer features. She sets expectations: some features are live and testable today, others are announced but not broadly available yet.

    • Plan for the episode: technical → creative demos
    • Google I/O announced many products/features with mixed availability
    • Goal: assess whether new consumer creative tools match the hype
  2. 1:35 – 4:07

    Gemini 3.5 family: fast coding + strong multimodal positioning

    Google introduces the Gemini 3.5 model family, highlighting Gemini 3.5 Flash as a fast, coding-oriented model that targets agentic workflows. Claire notes Google’s emphasis on agents and multimodal strengths, positioning Gemini as especially strong for file- and video-heavy tasks.

    • Gemini 3.5 Flash pitched as “fast + smart” for coding
    • Benchmarks emphasize agentic capabilities
    • Gemini’s core advantage: multimodal/file handling (docs, video, transformations)
  3. 4:07 – 6:09

    Antigravity 2.0 IDE: projects and scheduled tasks (Codex-like catch-up)

    Claire opens Antigravity (Google’s agentic IDE) and walks through the UI changes branded as Antigravity 2.0. The additions—Projects and Scheduled Tasks—feel similar to what users have seen in other agentic coding products.

    • Projects = workspace/folder-scoped environments
    • Scheduled Tasks run prompts on a cron-like cadence
    • UI and feature set resembles Codex-style workflows
    • Gemini 3.5 Flash options available inside the IDE
  4. 6:09 – 7:40

    Antigravity CLI hands-on: turning a blog generator UI into an agent-ready API

    She tests the newly announced Antigravity CLI by prompting it to convert an internal blog generator UI into a programmatic API for an AI agent. The flow mirrors other terminal-based coding agents: scan repo, request permissions, then implement changes.

    • CLI form factor similar to Claude Code/Codex CLI
    • Prompt: build an API version of an existing admin UI workflow
    • Agent requests permissions, inspects directories, starts coding
    • Claire’s early impression: not obviously “4x faster” in feel
  5. 7:40 – 9:11

    Core agent features shipping now: subagents, hooks, worktrees, local dev

    Claire summarizes the main “agent platform” improvements that Google says are available now. The highlights are subagents (task delegation) and hooks (lifecycle events), plus project/worktree support to better manage real development workflows.

    • Subagents allow the main agent to delegate tasks (including browser use)
    • Hooks enable custom actions on tool-use/turn/session events
    • Projects + native Git worktrees for workflow isolation
    • Positioned as feature parity/catch-up with existing agentic tools
  6. 9:11 – 11:12

    New slash commands: /goal, “Grill Me,” scheduling, and browser automation

    She spotlights Antigravity’s new slash commands, focusing on long-running goal pursuit and a more forceful clarification mode. Claire is especially curious whether “Grill Me” is truly more rigorous than typical polite Q&A clarification behavior.

    • /goal for persistent, long-running objective completion
    • “Grill Me” command for aggressive requirements clarification
    • /schedule to create recurring agent runs
    • /browser to trigger autonomous browser-based actions/testing
  7. 11:12 – 12:13

    Antigravity results: API endpoint created, docs generated, overall ‘fine’

    Claire checks the output of the Antigravity run: it created an API endpoint, auth, and supporting documentation. The work is solid and quick, but she frames it as expected baseline performance rather than a standout leap.

    • Created programmatic API endpoint + API key auth
    • Supports workflow triggers and featured image URL input
    • Documentation appears via an artifact-style view
    • Conclusion: competent, feels like parity more than breakthrough
  8. 12:13 – 13:44

    AI Studio updates: Workspace-connected apps and Android app generation

    Claire transitions to Google AI Studio and highlights the promise: build apps that directly connect to Google Workspace data (Sheets, Gmail, Drive, Calendar). She also calls out a new capability to create Android apps in this lower/no-code environment.

    • Built-in Workspace integrations aim to “own” internal productivity use cases
    • Potential to replace MCP-connector-style app building in other ecosystems
    • Android app creation offered inside AI Studio
    • Powered by Gemini 3.5 Flash in the new experiences
  9. 13:44 – 17:17

    Availability friction: promised Workspace connectors don’t show up in her account

    Trying to build a calendar-based weekend planner app, Claire can’t find the connectors or access she expects—even as a paying user. She flags a recurring theme: features announced as available are sometimes gated, unclear, or not functioning yet.

    • Attempts to access Calendar/Gmail/Workspace connectors fail
    • Settings/connectors don’t reveal the new integration options
    • Frustration: unclear rollout, “later this summer” caveats
    • She plans to revisit once access is enabled
  10. 17:17 – 19:18

    Gemini consumer app redesign + NanoBanana image generation test

    Claire notes Gemini’s redesigned consumer UI and improved “prompt inspiration” templates. She tests NanoBanana image generation by uploading a photo and requesting an upscaled, studio-style portrait—finding speed and text rendering improvements, but poor face fidelity.

    • Gemini UI redesign aims for a more premium consumer feel
    • NanoBanana integrated via templates for guided creation
    • Test: upscale/beautify + change background to studio
    • Output: fast and photoreal-ish, but face accuracy is off
  11. 19:18 – 22:50

    Omni video generation: animating a kid’s drawing and editing via conversation

    Claire demos Omni, Google’s new video generation model, by animating a child’s superhero drawing into a short narrative clip. She highlights key Omni claims: longer duration, better consistency, reference-based grounding, and conversational video edits.

    • Omni positioned as “NanoBanana for video” (reasoning + video model)
    • Reference-based generation: image → animated video
    • Conversational editing: modify scenes/objects with prompts
    • Duration and consistency improvements compared to earlier video gens
  12. 22:50 – 24:21

    Flow for cinematic, production-grade editing: characters, references, and workflows

    She introduces Flow, a more structured tool embedding Omni for cinematic editing and production workflows. Flow emphasizes character definitions (create once, reuse via @mentions), reference blending (people, environments), and organized video iteration.

    • Flow targets cinematic realism and production editing
    • Character creation + reuse via @mention-style mechanics
    • Blend multimodal references to seed scenes
    • Tooling for brainstorming, organization, and iterative refinement
  13. 24:21 – 26:54

    Avatar creation attempt: identity capture succeeds, feature fails to deliver

    Claire attempts to create a personal avatar through Flow/Google Labs by scanning a QR code and recording a guided face capture. The upload completes, but the avatar fails to generate—reinforcing her critique about broken or inaccessible launches.

    • Avatar setup requires phone-based capture and spoken prompts
    • System guides head turns and uploads training data
    • Result: avatar not created despite completing the flow
    • Takeaway: “announced” doesn’t equal reliable day-one usability
  14. 26:54 – 30:58

    Pomeli + Stitch: brand books, design systems, and streaming “Figma-like” UI generation

    Closing on tools for marketing/design, Claire demos Pomeli (brand/campaign generation) and Stitch (design canvas that streams UI output). She connects these to Google’s broader push to encode design systems for AI (e.g., Design.md), while noting the predictable “AI slop” aesthetic.

    • Pomeli: generate brand identity/brand book from a website
    • Stitch: streaming design generation in an in-browser canvas
    • Inline AI edits and code sync/import-export workflows
    • Design.md mentioned as part of Google’s AI design-system direction
  15. 30:58 – 33:52

    Recap: strongest wins are video + speed, but rollout gaps and naming sprawl hurt

    Claire summarizes the day’s releases across models, coding agents, no-code app building, consumer Gemini, and creative tools. Her verdict: exciting video capabilities and fast models, but too many brands, unclear access, and features that didn’t work undermine the experience.

    • Highlights: Gemini 3.5 Flash speed; Omni/Flow video potential
    • Antigravity adds parity features: projects, tasks, slash commands, subagents
    • AI Studio Workspace integration promising but unavailable for her
    • Pain points: branding confusion, rollout gating, broken avatar feature

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.