📰Trace Lab Newsletter #14 - 2025 Google I/O Edition

🌍 Welcome back Tracers!🌟

This week, we’re taking you straight into the heart of Google I/O 2025, which just happened overnight — the annual event where Google unveils the innovations set to transform the way we create, produce, and collaborate. And this year? No doubt about it: AI is everywhere.

🎥 4K video generation with native audio, 🎨 one-click design tools, 🧠 autonomous agents to simplify your workday, 🗣️ real-time meeting translations, and smart assistants that summarize, organize, and plan…
Welcome to the era of AI embedded at every level.

That’s why we’ve crafted a special two-part newsletter — because choosing between the creative revolution and the productivity revolution just wasn’t an option. Buckle up: the updates are coming fast and furious — but every single one was handpicked for its real impact on how we work at Trace.

PART 1 : AI Creative Studio

Google I/O 2025
🎬 Veo 3 & Flow: AI filmmaking goes 4K — with native sound

The era of AI-generated video just entered a whole new dimension with Veo 3 and Flow.
Google unveiled a model capable of producing 4K clips complete with dialogue, sound effects, and ambient audio — all from a simple prompt. With Flow, you can chain scenes together, direct camera movements, and even import your own visuals.

Key Points :

• Veo 3 generates videos with native audio — including dialogue, music, and sound design.

• You can control camera movements, characters, and scene direction.

• Flow is a creative AI video studio for assembling, cutting, editing, and exporting sequences.

• Available now for Google AI Ultra subscribers.

🔗 Full Article

Why it matters for Trace:

This is a game changer for Trace Studios — imagine prototyping a clip, teaser, or full animation in minutes, with coherent visuals and synced audio. Massive time-saver for creatives, and a whole new way to approach audiovisual storytelling.

Google I/O 2025
🖼️ Imagen 4 & Canvas: The instant generative design power duo

The new version of Imagen takes image generation to the next level — producing richer, sharper visuals with perfectly integrated text and typography.
Now built into Canvas, it lets you transform any content — article, report, or project brief — into a web page, infographic, quiz, or podcast. All with one click.

Key Points:

• Imagen 4 finally nails typography inside images.

• Generates visuals 10x faster than the previous version.

• Canvas lets you turn any written content into creative or interactive formats.

• Available in the Gemini app for all subscribers.

🔗 Full Article

Why it matters for Trace:

For brand, production, and social teams: this means rapid creation of sleek visuals, educational formats, or interactive assets — ideal to refresh our content or generate custom materials on the fly.

Google I/O 2025
🧾 NotebookLM + AI Podcast: Summarize, present, broadcast

NotebookLM is quickly becoming a secret weapon for turning dense documents into visual presentations and audio content.
It now generates video summaries, 5- to 20-minute podcasts, and infographics — ideal for briefs, debriefs, or client presentations.

Key Points:

• Generates audio and video summaries from notes or documents.

• Three podcast formats available: 5, 10, or 20 minutes.

• Mobile version announced for Android and iOS.

• Included in the AI Ultra plan, with preview access already available.

🔗 Full Article

Why it matters for Trace:

Internal briefs or client decks turned into podcasts? Dense docs converted into visual explainers? That’s a major boost in clarity, time savings, and a great way to repurpose content both internally and externally.

Google I/O 2025
⚡ Gemini Diffusion & Deep Think: Lightning-fast generation and reasoning

Gemini is accelerating the creation of complex content with two powerful new features:

• Diffusion, an ultra-fast model for text and code generation

• Deep Think, an advanced reasoning mode designed to tackle highly complex problems
  

Key Points:

• Diffusion generates output 5x faster while maintaining precision and structure

• Deep Think leads benchmark scores in math and code reasoning

• Includes “Thinking Budgets” to balance cost, latency, and quality

• Available in Gemini API and Vertex AI, with preview access already live

🔗 Full Article

Why it matters for Trace:

These tools boost automated content generation — scripts, FAQs, articles — and enable smarter features in Trace+ products and internal apps. Speed and intelligence combined: this is the future of dev and content at Trace.

PART 2 : Productivity + Agents Revolution

Google I/O 2025
🤖 Project Astra + Gemini Live: A truly “alive” AI assistant

Gemini Live brings together vision, voice, conversation, and Google apps into one intelligent assistant.
Thanks to Project Astra, Gemini can now see what you show via camera, hear you, understand your context, respond in real time — and even make a call or add an event to your calendar. Welcome to the age of the contextual AI companion.

Key Points:

• Interacts via camera, microphone, screen sharing, and natural voice

• Direct integration with Maps, Gmail, Keep, Calendar, and Tasks

• Understands objects, locations, handwriting, and supports decision-making

• Rolling out progressively on Android and iOS via the Gemini app

🔗 Full Article

Why it matters for Trace:

Picture a producer showing a shoot plan on their phone — and Gemini generates the gear checklist. Or a manager filming a whiteboard and instantly getting a summary. Less friction, more action.

Google I/O 2025
🌐 Agent Mode (ex-Project Mariner) : L’IA fait le boulot sur le web pour vous

Google passe à l’étape supérieure avec des agents capables de naviguer et interagir en autonomie.
Agent Mode peut comparer des offres, chercher un appart, planifier un voyage ou exécuter une tâche répétée. Il apprend avec le “teach & repeat” et peut gérer plusieurs actions en parallèle.

Ce qu’il faut retenir:

• Multitâche : jusqu’à 10 tâches simultanées.

• Mode “Teach & Repeat” : montre-lui une tâche, il saura la refaire.

• Bientôt dans Gemini App, Chrome, et Search.

• API disponible pour créer ses propres agents.

🔗 Lire l’article complet

Pourquoi c’est utile pour Trace :
Ces agents pourraient être intégrés à nos propres interfaces internes pour automatiser des flux répétitifs : sourcing de prestataires, suivi de diffusion, booking de talents… L’IA devient votre assistant perso ultra-efficace.

Google I/O 2025
🧠 Agent Mode (formerly Project Mariner): Let AI handle the web for you

Google is leveling up with AI agents that can browse and interact on the web autonomously.
Agent Mode can compare deals, search for apartments, plan trips, or carry out repetitive tasks. It learns via “Teach & Repeat” and can manage multiple tasks at once.

Key Points:

• Multitasking support: up to 10 tasks at the same time

• “Teach & Repeat” mode: show it once, it’ll do it again next time

• Coming soon to Gemini App, Chrome, and Search

• API available to build your own custom agents

🔗 Full Article

Why it matters for Trace:

These agents could be embedded into our internal tools to automate repetitive workflows — from vendor sourcing to broadcast tracking to talent booking. AI becomes your ultra-efficient personal assistant.

Google I/O 2025
📬 Gmail Smart Reply with Context: Replies that sound like you

Smart Replies are getting smarter — and more personal.
Powered by Gemini, Gmail now analyzes the content of your emails and connected Drive files to suggest automatic responses that reflect your tone, priorities, and context.

Key Points:

• Personalized style based on your past email exchanges

• Optional integration with Gmail and Drive for deeper context

• Rolling out this summer for AI Pro and Ultra subscribers

• Built into the new AI-powered version of Gmail

🔗 Full Article

Why it matters for Trace:

Replying to a client email with precision — without rereading 10 previous messages? Totally doable. Your answers are faster, more coherent, and more professional — even when you’re in a rush.

Google I/O 2025
🌍 Google Meet Real-Time Translation: No more language barriers in meetings

International meetings just got a whole lot smoother thanks to AI-powered voice translation.
With Gemini, Meet now offers real-time translations — while preserving the speaker’s original voice, tone, and style

Key Points:

• Currently available in English ↔ Spanish, more languages coming soon

• Real-time translation with faithful voice reproduction

• Available for Pro/Ultra subscribers, then rolling out to enterprises

• Built with technology from Project Starline

🔗 Full Article

Why it matters for Trace:

Our teams are spread across the globe. This feature improves communication flow, prevents misunderstandings, and boosts inclusion in international meetings.

UNTIL NEXT TIME…

Let’s keep exploring — together!

As you’ve seen, Google isn’t just dreaming about AI anymore — it’s building it into everything.
Whether it’s unlocking new creative potential or streamlining our workflows, the announcements from this I/O 2025 mark a clear new chapter.
👉 And at TRACE, we’re front row for turning this into real impact.

🎯 These tools and agents aren’t here to replace your talent — they’re here to give you more time, more clarity, and better flow.

📩 Got an idea, a pilot project, or just curious about how to use these tools at Trace?
→ Reach out to TRACE Lab! We’re here to help you experiment, integrate, and scale.
Reply directly to this newsletter or drop by and chat with us!

🎥 Want to dive deeper?

• 🔴 Watch the full keynote (1h40): Google I/O 2025 – Full Keynote

• ⚡ Quick 10-minute recap: Official YouTube Recap

Subscribe to Trace Lab Newsletter