New Gemini Spark Update Is INSANE!
This video covers Google's new Gemini Spark feature, an agentic AI tool available to Google AI Ultra subscribers in the US that can autonomously execute tasks in the background without requiring the user's screen to be active. Spark operates on three core components — tasks, schedules, and skills — and integrates directly with Google apps like Gmail, Calendar, Drive, and Docs. The presenter positions Spark as a major shift from AI that merely responds to AI that independently takes action.
Summary
The video introduces Gemini Spark, a new agentic AI feature from Google that runs on Google's servers and continues working even when a user's device is off or locked. Unlike traditional AI assistants that only respond to prompts, Spark is designed to autonomously complete tasks — browsing the web, managing emails, building documents, and interacting with apps — without requiring the user to remain active. It is currently available in beta for Google AI Ultra subscribers in the United States, in English only, with broader rollout to businesses and additional platforms planned.
Spark is built around three foundational components: Tasks (what you want done, described in plain language), Schedules (when you want it done, either at a set time or triggered by an event), and Skills (saved step-by-step recipes that Spark can reuse or invoke automatically). Together, these allow users to set up automated workflows that run repeatedly without further input.
The tool integrates with core Google services including Gmail, Google Calendar, Drive, Docs, Sheets, Slides, Maps, and YouTube. These integrations are opt-in via settings. Practical use cases demonstrated include sorting inboxes, drafting email replies, building slide decks from a single sentence, managing calendar scheduling, summarizing newsletters, and identifying forgotten subscriptions in financial statements.
Beyond Google apps, Spark can operate a real web browser — navigating websites, clicking, and filling in forms — essentially acting as a human operator at a computer. A transparency panel shows every planned, active, and completed step, and users can intervene and take over the browser at any point. Spark can run up to 15 tasks simultaneously.
Technically, Spark is powered by the Gemini 3.5 Flash model (the 'brain') combined with what the presenter calls the 'anti-gravity harness,' which is the layer that enables the model to interact with apps and interfaces. The presenter notes Spark is currently available on web and in the Gmail Android/iOS app, with a Mac desktop app coming that will extend its reach to local files.
The video closes with practical onboarding advice — start with one task, add a schedule, then build a skill — and promotes two paid/free communities: the AI Profit Boardroom and the AI Success Lab.
Key Insights
- The presenter argues that Spark represents a fundamental shift in AI utility — moving from AI that 'talks' (answers questions) to AI that 'does' (takes autonomous action on the user's behalf without the screen being on).
- Spark's ability to execute tasks is explained technically as the combination of the Gemini 3.5 Flash model (the decision-making brain) and an 'anti-gravity harness' — the layer that gives the model the ability to click, type, and navigate apps, turning a chatbot into an autonomous worker.
- The presenter notes that Spark can operate a real web browser — navigating to websites, clicking elements, and filling in forms — and includes a step-by-step transparency panel that allows users to monitor progress and manually take over at any point, which Google itself recommends doing.
- The presenter claims that Spark can run up to 15 tasks simultaneously in the background, and that completing a task frees up a slot for a new one — enabling parallel automation across multiple workflows at once.
- Google has announced that Spark will expand to a Mac desktop app, which the presenter highlights as significant because it will allow Spark to work with local files stored on the computer — not just cloud-based Google apps — broadening its automation reach to the entire machine.
Topics
Transcript
[0:00] New Gemini Spark update is insane. Google just turned Gemini into something new. It's called Gemini Spark. And as of right now, it's live for Google AI Ultra users in the United States. Here's the part that got my attention. Spark keeps working when you close your laptop. Keeps working when you lock your phone. Keeps working while you sleep. Lives on Google's own servers, so it doesn't need your screen on to do its job. Let me say that again in plain words. You give it a task, you walk away. Does the work for you. For years, AI just talked. You asked a question, it typed an answer. That was it. You still had to go do…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
NEW Nvidia Autonomous AI is WILD!! 🤯
Nvidia announced Nemo Clo, a new autonomous AI agent system that operates independently without continuous prompting. Powered by Nemotron 3 Ultra (a 550 billion parameter model), the system is five times faster and cheaper than previous versions, with OpenShell providing secure sandboxed execution.
Laguna XS 2.1: New FREE + Opensource Local AI!
Julian reviews Laguna XS 2.1, a new free open-source local AI coding model from Poolside that performs comparably to Qwen 3.6 and outperforms Claude Haiku on benchmarks. He demonstrates its practical capabilities by building landing pages and functional apps, highlighting its speed, offline functionality, and multiple deployment options through local setup, Claude Code, or OpenRouter's free API.
How to Run Hermes FREE Forever!
The video demonstrates how to run the Hermes AI agent for free using Gemma 4, a local open-source model from Google, with significant speed improvements through MLX optimization. The setup works on Apple Silicon Macs or via free APIs on Open Router, enabling autonomous agents to work offline and privately without subscription costs.
This NEW Chinese AI is INSANE! (FREE + Open Source!)
Long Cap 2.0 is a new open-source Chinese AI model from a food delivery app company that offers 1 million tokens of free context memory, beats GPT-4.5 on SWE bench pro benchmarks, and uses efficient parameter activation to reduce computational overhead while maintaining high performance.
Claude Code is now FREE: Here’s how…
Google's new Gemma 4 model running on Ollama is 90% faster on Apple Silicon, enabling free Claude Code usage locally without token costs. The setup requires three simple steps: downloading Ollama, Gemma 4, and installing into Claude Code, with alternatives available via OpenRouter API for non-Mac users.