AI Society for 3.13.25 — Telcos Embracing AI
Today: Telcos and AI, Gemini Robotics, Differential Privacy, and the letter ‘M’
To lead, a great summary by ‘McKinsey’ of the recently concluded Mobile World Congress, and here, I’ll focus on what they said about AI. Their key conclusion:
And more details:
Turning to Google, a new AI model for robotics. As reported by the ‘Mercury News,’
Get dave ginsburg’s stories in your inbox
Join Medium for free to get updates from this writer.
· Research unit Google DeepMind will release Gemini Robotics, a new branch of its flagship AI model aimed at developing robots that are more dexterous and interactive, it said Tuesday. Another model, Gemini Robotics-ER, specializes in spatial understanding, and will help robot makers build new programs using Gemini’s reasoning capabilities.
· Engineers also set out a miniature toy basketball court in the lab. Another robot, when asked to perform a dunk, pressed a small plastic ball through the hoop.
· “The team was really excited when we first saw the robot dunk the basketball,” Rao said. “It’s because the robot has never ever seen anything related to basketball. It’s getting this general concept, understanding of what a basketball net looks like and what the word ‘slam dunk’ means from Gemini and is able to connect those to actually accomplish the task in the physical world.”
A link to the Google blog.
On the security front, an interesting essay by Jonathan Yahav on what is termed ‘Differential Privacy’ and how recent tweaks can make it more realistically deployable. The approach is ‘high-budget DP,’ and worth reading if you’re in this space. I can’t help but think it has applications in other fields as well, where the search for perfect security sometimes overlooks the realistic. He references ‘Reconciling privacy and accuracy in AI for medical imaging.’
Do bots really prefer a certain letter of the alphabet? Jim the AI Whisperer at ‘The Generator’ describes his experiment, with the bot selecting the letter ‘M.’ Trust me, I just did the same, but then asked why:
He goes on to compare preferences across different models, with Sonnet, Copilot, Mistral, and Grok showing differences, and OpenAI models also showing variances. However, DeepSeek-V3 replicates GPT-4. Jim continues with a good discussion of statistics and randomness, and why we shouldn’t let AI generate our passwords.
