Hey there,

Google I/O happened on Tuesday and they gave us a lot of exciting new products, models, and updates. This week we’ll cover the most important releases like Gemini 3.5 and the Gemini Spark agent. This year’s Google I/O focused on showcasing new agents and fully integrating AI features into Google’s core products like Search and Workspace. 

The two models worth knowing about

The Gemini 3.5 family of models was announced this week. 3.5 Flash is already available through the Gemini app, API, and AI Studio. 3.5 Flash is Google’s best coding and agentic model, beating out past models like Gemini 3.1 Pro. Like all Gemini Flash models, this is meant to be a balance between speed, performance, and affordability. Google highlighted that this model will complete tasks at half the cost of other frontier models. 3.5 Pro will be coming next month after the final reviews. 

Gemini Omni is a brand new model that Google is labeling a “world model”. Omni can create realistic, grounded video content from text, image, and video inputs. Think of Omni as the video version of Nano Banana. It uses its extensive knowledge to create videos that are grounded in facts. You can also edit your own videos using natural language. Omni is currently available in the Gemini app, Google Flow, and even for free on YouTube. 

Gemini Spark and personal agents

Unsurprisingly, Google has made a massive push toward agents. This includes a brand new agent called Spark that can work across your workspace, connectors, and even the web. Spark is an autonomous agent that can complete a wide range of tasks from monitoring your credit card purchases to sending you reminders about events mentioned in your emails. Spark runs in a secure sandbox and can continue working in the background even when you close your laptop. This agent will begin to roll out next week to Beta users. 

The Daily Brief is a much lower-stakes version of Spark that is already available today. This feature connects to your Google Workspace to create a unique personalized daily summary for you each morning. It tells you about meetings, upcoming events, and reminds you of tasks you need to do.

Search just got a major update

We got the first major AI-focused change to Google Search last year with the release of AI Mode and AI Overviews. Both of these have brought Gemini directly into the Search experience and receive over 1 billion monthly users. To top this success, Google is making their largest upgrade to Search ever with new Generative UIs and Search Agents. 

Generative UIs are unique interactive visuals that help bring your questions to life. Search can generate simulations, tables, graphs, and more to help you understand complex topics. These UIs are generated using Gemini 3.5 Flash. 

Similarly to Spark, the new Search agents will be active around the clock. This means you can stay up-to-date on news, topics, or tasks you assign the agent. These Search agents are meant to constantly explore the internet for any websites, posts, videos, or mentions of the topic you give it. It will then share a report with you in real time about what it finds. This feature is rolling out this summer.

Other cool updates

Google’s AI watermarking technology SynthID got a major update. Not only can you now use it in Search, but Google has partnered with other AI companies like OpenAI to have all their AI generated videos and images include watermarks. While these watermarks can’t be identified visually by humans, Google’s detection tools can identify them instantly. This is meant to help identify AI-created media as these models continue to get more capable. 

Ask YouTube is a new AI feature that mirrors Search’s AI Mode. It allows users to ask questions directly in YouTube and receive a response that includes relevant YouTube videos. Both Shorts and traditional videos will be used to answer your questions.

My Thoughts

Google I/O has brought a ton of new tools and features across the Google ecosystem that I’m excited about trying out. The Spark agent in particular is something that I’m looking forward to exploring when it releases in a few weeks. One of the things I’m a little disappointed by is the number of announcements “coming this summer”. I would have loved to get my hands on more releases today as opposed to previews of what’s coming soon. 

Despite this I think we got a great Google I/O with several standout updates. The new SynthID updates are an excellent addition to the AI ecosystem. As these models continue to get better at replicating life-like details, it's essential that we have a tool to identify what’s AI-generated.