Sponsored by
Good morning. It’s Wednesday, May 20th.
Hey everyone! Once again, I’ll be at the Eazo AI 2026 Global Hackathon in Mountain View this weekend.
May 23–24.
$300K prize pool.
No-code app building.
Real products, not pitch decks.
If you’ve ever had an “someone should build this” idea, this is the weekend to show up!
I hope to see you there :)
-Jeff
AI Breakfast
You read. We listen. Let us know what you think by replying to this email.
Stop Paying for 6 Tools. One AI Does It All
Most e-commerce sellers are running their store across 6 to 8 separate tools — and paying hundreds of dollars a month for the privilege. StoreClaw replaces your entire stack with one autonomous AI engine that monitors competitors, optimizes listings, automates marketing, and tracks real profit across Shopify, Amazon, and beyond.
It doesn't wait for you to ask. It runs 24/7 in the background, so you wake up to a full dashboard instead of a list of things you forgot to check.
Connect your store, and StoreClaw gets to work — no prompts, no complex setup, no six-app stack.
Free to start. No credit card required.
Always-on agents arrive at Google I/O
Google I/O 2026 ushered in the always-on agent era turning idle devices into tireless multi-step workers.
The Major Leaps
Google launched Gemini 3.5 Flash, its strongest coding and agent model yet, running 4x faster than rival frontier models. Built directly for multi-step execution loops, it outpaces Gemini 3.1 Pro on core developer benchmarks, scoring 76.2% on Terminal-Bench 2.1 and 83.6% on MCP Atlas. It lands immediately as the default engine for the Gemini app and Google Search AI Mode globally, priced at $1.50 per million input tokens and $9 per million output tokens, with a heavier Gemini 3.5 Pro variant dropping next month.
Gemini Omni introduces a multimodal world model system designed with a core understanding of physics and continuity to allow conversational video editing while maintaining scene and object consistency. Its first iteration, Omni Flash, is rolling out across YouTube Shorts, the YouTube Create app, Gemini, and Google Flow, while also supporting custom AI avatars using a user’s voice and likeness.
Gemini Spark is a 24/7 cloud-based AI agent built on Gemini 3.5 Flash that stays active across Gmail, Docs, Calendar, and third-party apps via MCP. It can draft emails, organize files, monitor inboxes, and carry out workflows in the background. Instead of waiting for prompts, it continuously runs tasks within set user limits. The newly rolled out Gemini app for macOS will receive Gemini Spark support later this summer.
Google Flow is evolving into a collaborative studio with Flow Tools, which lets users describe custom editing functions and have the AI generate them automatically. It also includes specialized Flow Agents plus new section-level editing and AI-guided music video creation inside Flow Music.
Antigravity 2.0 rolls out as a standalone desktop app, CLI, and SDK built natively for orchestrating parallel coding agents that spin up subagents and compile tasks asynchronously in isolated Linux environments. To prove the platform's viability against rivals like Claude Code, Google showcased a stunt where 93 parallel subagents burned through 2.6 billion tokens to build a functioning operating system framework from scratch in just 12 hours, and then live-coded a Doom clone to run on top of it.
Watch: Google Antigravity 2.0 Beginner's Guide
Project Genie links Street View with AI world models to create editable, interactive real-world simulations and train agents in realistic environments.
Search is being rebuilt into an execution layer powered by Gemini 3.5 Flash, turning the search box into a multimodal input surface. It can generate dashboards, mini-apps, and trackers on demand, with Pro and Ultra users also getting continuous monitoring for ongoing queries.
Gemini for Science is an AI research suite designed to take over the tedious grunt work of the lab. It combines tools like Co-Scientist and NotebookLM to automate hypothesis drafting, parse massive piles of academic literature, and pull data straight from genomic databases for rapid analysis.
Refinements and Bets
Android Halo is a subtle system indicator that shows when background AI agents like Gemini Spark are active. It appears as a persistent glow at the top of the screen, turning ongoing agent activity into a lightweight, always-visible signal without interrupting normal phone use.
The Gemini app received a full Neural Expressive redesign, replacing static text responses with smoother animations, brighter colors, refreshed typography, haptic feedback, and dynamic response layouts.
Daily Brief is an opt-in feature in the Gemini app that pulls in data from connected apps and turns it into a ranked morning summary. It surfaces key updates, filters digital noise, and suggests next actions based on user goals, acting as a structured daily briefing layer inside the Gemini system
For collaborative writing inside Workspace, Docs Live turns spoken input into a structured writing tool, converting freeform verbal ideas into fully formatted, edited documents in real time for faster collaborative drafting.
Universal Cart is a cross-platform shopping agent built into Search, YouTube, and Gmail and tied to Google Wallet. The system is designed to prepare purchases with cryptographic safeguards, laying the groundwork for future autonomous checkout.
The Ask YouTube upgrade swaps standard video links for an interactive AI interface inside Search. It extracts text answers and key clips directly from video timelines, letting you talk to a video and get an immediate answer instead of scrubbing through the timeline.
CodeMender adds automated security to the developer stack, actively scanning AI-generated or external code for vulnerabilities and applying verified fixes with testing loops to challenge Anthropic Mythos.
AI Studio is bringing prompt-based vibe coding straight to mobile. The new Android app lets you type a description to generate native apps using Kotlin and Jetpack Compose, test them instantly in a mobile emulator, and ship them directly to the Play Console.
Google is taking another shot at head-worn tech. This fall, it's launching Android XR intelligent eyewear built with Samsung and Qualcomm, featuring everyday frames from Warby Parker and Gentle Monster. The audio-only glasses give you hands-free camera and Gemini access directly from your phone, while a separate tethered display pair called Project Aura uses a compute puck to float massive 3D Google Maps and multi-window apps right in your field of view.
Watch: Live Demo of Android XR at Google I/O 2026
To manage content transparency, Google is expanding its SynthID watermarking to Chrome and Search, and in a rare alignment, direct rivals like OpenAI, ElevenLabs, and Kakao are adopting the tech for their own tools.
MODELS & RESEARCH
AGENTS & AUTOMATION
INFRASTRUCTURE & HARDWARE
INDUSTRY & M&A
ROBOTICS
POLICY, LAW & REGULATION
LABOR & ECONOMY
SCIENCE, HEALTH & SPACE
FINANCE & CRYPTO
CULTURE & CONSUMER
Fere AI deploys wallet-native agents to execute automated crypto and Polymarket trading strategies based on live market signals.
PromptUnit is an AI proxy that automatically routes requests to cheaper models and bills only from savings.
CtrlOps provides a local AI terminal that generates commands and handles server deployment directly from GitHub repos without DevOps expertise.
LobeHub deploys cloud-based AI agent teams to execute complex workflows and delivers unified progress updates to your messaging apps.
ReactVision Studio provides a browser-based visual editor to build AR and VR apps from a single React Native codebase and ship them to iOS, Android, and Meta Quest devices.
Thank you for reading today’s edition.
Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.
Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!
Thinking of starting your own newsletter? AI Breakfast readers who sign up with Beehiiv receive a 14-day free trial and 20% off for 3 months.
