Dawid Ostrowski’s Post

Gemini Show & Tell - GDEs will be showcasing what they have built with Gemini!

The new way of software development enhanced by AI and experienced by me recently: 1. I tried to configure some BLE device using the app supplied by the producer, but the app kept crashing. 2. I took the protocol specs, and fed it into Gemini 3. I asked Gemini to write a method for me to communicate with the device (even including more advanced stuff like packet splitting etc. 4. I added a few tweaks 5. Voila! It works! Success! Profit (not really :)) ! BTW. it all happened in Python, and I'm not sure if I wrote a total of 20 lines in Python in my life. I think it's really cool, it sped up my regular process significantly and I didn't have to go through forums, asking questions and having to read toxic responses like "Why are you even doing this" :) Below is an actual photo of me working on this pet project :)

In a very well-written article, Justyna Politanska-Pyszko shares her "mental health emergency kit". Worth reading! https://lnkd.in/dXXP-vND

Busy with all the AI stuff, I kind of missed that Google Home has now a very solid Developer Center! The Google Home ecosystem has matured. It now supports Matter as the communication protocol, there's a bunch of SDKs, and tools like UI Automator, VS Code extension, Matter Virtual Device, etc. Worth checking at: https://lnkd.in/dddiD-ug

As part of Google's multi-year collaboration with Liverpool FC, DeepMind just released a full AI system "Tactic AI" that can advise coaches on corner kicks :) The model is built to address three main questions: 1. For a given corner kick tactical setup, what will happen? e.g., who is most likely to receive the ball, and will there be a shot attempt? 2. Once a setup has been played, can we understand what happened? e.g., have similar tactics worked well in the past? 3. How can we adjust the tactics to make a particular outcome happen? e.g., how should the defending players be repositioned to decrease the probability of shot attempts? TacticAI is a full AI system that could give coaches instant, extensive, and accurate tactical insights – that are also practical on the field. You can read the full paper at: https://lnkd.in/djiYpT3h

In “Global prediction of extreme floods in ungauged watersheds”, published in Nature, Google demonstrates how machine learning (ML) technologies can significantly improve global-scale flood forecasting. https://lnkd.in/dRkEhXHk Moreover, Google's Flood Hub provides real-time river forecasts up to seven days in advance, covering the river reaches across over 80 countries. Read more at: https://lnkd.in/dkqsScqH

An interesting TPgM role at the coolest part of Google these days! :)

Great opportunity to lead policy efforts where it matters the most!

What's easier - being a planetary explorer or being a goat? For AI it's still easier to go to the spaceship than to jump a fence... It looks like SIMA (The Scalable, Instructable, Multiworld Agent from Google DeepMind) had a slightly better average performance in No Man's Sky than it had in the Goat Simulator 3 :) The full paper on SIMA is super interesting: https://lnkd.in/ghKENJea

An interesting learning opportunity in Switzerland in the area of ML and Data Science. High standards guaranteed by one of AI/ML GDEs Umberto Michelucci (Parts of the program are in German)