From Baby DNA to Robot Acrobatics: AI’s Bold, Bizarre, and Ethical Frontiers
AI just made robots jump higher, cut MIT brains some slack, and scanned your socks—yes, really. But deep under the fun, privacy, power use, and DNA ethics make the future less sci-fi and more “
Facebook is starting to feed its AI with private, unpublished photos
Tina Nguyen/ The Verge
The new food source for Meta's AI is your own, unreleased images. As the platform covertly increases its training data, the action raises new privacy concerns.
Using generative AI to help robots jump higher and land safely
Alex Shipps/ MIT News
AI simply gave robots a boost. To help machines jump higher and land more intelligently, researchers are employing generative models, which will increase agility.
Ethical questions surround the NHS sampling babies’ DNA | Letter
The Guardian
Newborns' DNA is being collected by the NHS in order to gain insights into future health, but some are concerned about consent, data security, and the potential for genetic surveillance.
Google’s emissions up 51% as AI electricity demand derails efforts to go green
Helena Horton/ The Guardian
Google’s Doppl app took off my socks
Emma Roth/ The verge
AI is not only intelligent, but also energy-hungry. Google's emissions have increased by 51%, underscoring the increasing conflict between climate goals and AI development.
M3gan 2.0 is a fun upgrade that’s a little too self-aware
Charles Pulliam-Moore/ The Verge
The Doppl app from Google is drawing attention—and reportedly taking off socks. Personalisation and strangely realistic physical interactions are combined in this experimental AR+AI tool.
Hard Fork Live, Part 1: Sam Altman and Brad Lightcap of OpenAI
New York Times
In a live interview with Hard Fork, OpenAI directors Brad Lightcap and Sam Altman discussed AI safety, future models, and the future of the game's most-watched lab.
Evaluating SOP documents: A hands-on with LLM-as-a-Judge
Arunabh Bora/ Towards AI
What occurs if you allow an AI to review your SOPs? Structure, clarity, and even tone are all graded by language models in a new experiment.
The Best Optimization Algorithm for Your Neural Network
Riccardo Andreoni/ Towards AI
Are you torn between SGD, RMSProp, and Adam? Which optimizer is best for your neural network and when to swap it up are explained in this article.
From Insight to Amnesia: How ChatGPT Cut Cognitive Load by 47% in MIT Study
R. Thompson/ Towards AI
Researchers from MIT discovered that ChatGPT reduces mental work by almost half, but there's a caveat. Overuse could result in digital forgetfulness by sacrificing understanding for convenience.