Kersebom: Google I-O 25-05-20

A subjective, selective but substantial birds-view (keywords) of the Google I/O-perspectives – printable on 1 page (Quickie)

The normal WP-Textversion

01 Basics: Web address: https://www.youtube.com/watch?v=o8NiE3XMPrM. length: 1h56m+. > 10 speakers.

02 Graph: 20+ major Google AI products + features since the last IO

03 Graph2: Model Progress over the last 18 months – ELO score

04 Graph3: gemini-2.5-pro-preview-05-06: #1 (of 13) across all LMArena categories: Included: Other versions of gemini, chatGPT, deepseek…

05 Graph4: Fastest Intelligence: Output tokens per second: The same (?) 13 candidates. Google #1-3.

06 Numbers: 7M+ developers across AI Studio + Vertex AI.

07 400M+ monthly active users of the Gemini App.

08 From Project Starlink to Google Beam: A new (3D?) video communication plattform: An array of 6 cameras. 6 pictures are rendered into 1 picture, 60 frames/s. With HP they´ll deliver the device later this year.

09 10m04s+: From starline to Google Meet: Demo Videocall between US & Southamerica (.es). The voices are constructed on the basis of the individual voices (?). Rolling out more languages in the coming weeks.

10 Gemini Live: Intelligent camera + screen sharing: Rolling out on Android+iOS from today (in the US only?).

11 project Mariner: An agent eco system. More broadly available in summer. 60+ Agent2Agent Partners.

12 15m24s+: Gemini SDK now compatible with MCP (https://en.wikipedia.org/wiki/Model_Context_Protocol).

13 Agent mode. Personalization: Personal context.

14 Gmail: Automatic intelligent responses to posed questions per gmail. All (allowed) personal electronic ressources of the responder are being used. Available this summer for subscribers.

15 19m49s+: Founder Hassabis: Deepmind: 2013-2025: What we´ve already created. He will Gemini build up to a universal tool like the brain is.

16 30 things that you can build with Gemini today: Simulate: Physics (1), Nature, Light, Puzzles, turn images to code, turn your ideas into 3D-objects (6), make your drawings 3D, idea for a hairdraw, put a tuxedo on this cat, search decades of documents, search hours of video (11), build robots, code with your voice, play anything, animate anything, translate + calculate anything (17), explain anything, imagine anything, what can we do next: Here are some other 30 ideas for you…

17 23m33s+: Multispeaker support for 2 voices built on native audio output (nao). Different language styles (from speaking loudly to whispering). nao available in the Gemini API today (2.5 Flasch + Pro Preview).

18 44m30s: How Astra already now may help blind persons substantially.

19 Google Search: AI Overviews. Search Live (in realtime). The biggest revolution to come AI Mode.

20 Google Lens: 65% growth over year. year over year 100+ billion searches. 1.5B+ users every month.

21 AI Mode: For longer + more complicated queries? Rolling out starting today – in the USA. Beyond information -> intelligence. AI modus search transformed with Gemini 2.5 at its core with 5 elements: Personal context (1). Deeper research. Analysis & visualization. Live multimodality. New ways to shop (5). This summer.

22 Shopping: 1h02m30s+: 1. 50B product listings in the shopping graph. 2. A lady wants online to buy a dress and see first, how she is looking in this dress. 3. Inspire -> shop -> pay. agentic checkout features.

23 1h17m6s: Canvas – Gemini´s interactive space for cocreation.

24 AI-outputs in form of a dynamical webpage, an infographic, a helpful quizz…

25 Gemini in Chrome – your personal AI assistant für browsing. Roll out this week for US-subscribers.

26 AI image generation: Imagen 4 – a big leap forward! Starting today…

27 1h20m37s+: AI audio + video generation: Veo 3: For creators, musicians, filmmaker (generative media).

28 prices: AI Pro 20 $/m. AI Ultra 250 $/m. What is for free?

29 Gemini to Wear OS 6 in the coming months.

30 Android XR: 1. 1st Android platform built in the gemini era. 2. First unveiled in the Android show last week. 3. Space: new form factors. Glasses. TV. Dashboard. Cars. And everything in between – incl. Google Apps without end. VR-glasses (Project Moohan/Samsung with 1st device/later this year).

31 1h41m53s+: Glasses with Android XR: Including camera + microphone + AI + plays music + works with your phone + hands free + smoothly colored text messages + teleprompter. You get superpowers when you put your glasses on. My next glasses. Probably very expensive for a long time!

32 Google-IO-Arena word Count Leaderboard: Gemini 95 (1), AI 92 (2)…

33 Android Show I/O-Edition (last week/23h14s): https://www.youtube.com/watch?v=l3yDd3CmA_Y. 1. Google Android 16. 2. Material 3 expressive design. 3. Wear OS 6 – arriving later this year. 3. Gemini is coming to everything incl. the watch (often) later this year. 4. Google Pixel Smartwatch 4: Not mentioned.

(24.5./9.6.25)