And ofc, there is a recently added Gemma 3n E4B. A SOTA, lightweight, open-source multimodal AI model, designed for efficient execution on low-resource devices.
GENERATIVE AI
-
Google AI Studio Adds Code Assist
By
–
ICYMI: Google AI Studio now has code assist in the Apps section. Now you can do vibe coding straight inside AI Studio from anywhere with diffs and all that stuff. pic.twitter.com/1ekucQyvp2
— 🚨 AI News | TestingCatalog (@testingcatalog) 25 mai 2025ICYMI: Google AI Studio now has code assist in the Apps section. Now you can do vibe coding straight inside AI Studio from anywhere with diffs and all that stuff.
-
500000 Hugging Face Spaces Showcase AI Community Innovation
By
–
I'll follow what people are building here: https://
huggingface.co/spaces?sort=tr
ending&sdk=static
…. Over 500,000 spaces have been created so far! -
AI Data Collection Scale Compared to 2010s Tech Giants
By
–
the scale of data collection in the AI labs pales in comparison to 2010s google it’s mostly web scraping and data-labeling. compare that to diligently photographing streets of every country, mapping earth via satellite, scanning every book known to man.. now *that* was ambitious
-
Veo 3 Struggles with Split-Screen Foley Artist Prompt
By
–
This was a prompt idea from @steveruizok, but I don't think it's one that Veo 3 can get right, this was the closest:
— fofr (@fofrAI) 25 mai 2025
> a video is split 50/50, on the left is a scene, on the right is a foley artist producing the sound effects for that scene pic.twitter.com/Sx4kTFbePwThis was a prompt idea from @steveruizok
, but I don't think it's one that Veo 3 can get right, this was the closest: > a video is split 50/50, on the left is a scene, on the right is a foley artist producing the sound effects for that scene -
Upcoming Podcast with Sundar Pichai on AI Progress
By
–
I'm doing a podcast with @sundarpichai soon. Let me know if you have any questions / topic suggestions. The rate of AI progress has been insane. It makes me excited for the future (even more than usual ) and excited to chat with leaders & engineers who are building that
-
Media Overstates AI Potential Despite Tremendous Promise
By
–
"The NBER paper comes on the heels of other indications suggesting that AI’s potential, while tremendous, has been vastly overstated in the media and the market." But they're doing the exact overstatement here with the phrase "while tremendous."
-
Next Word Prediction and Loss Functions for Model Understanding
By
–
This is a brilliant explanation of why next word prediction is a good loss function to achieve understanding. In addition to predicting the next bit, one can also use masking losses and diffusion to achieve understanding. The general intuition is to minimise the difference… https://t.co/tUDUQWP4NJ
— Nando de Freitas (@NandoDF) 25 mai 2025This is a brilliant explanation of why next word prediction is a good loss function to achieve understanding. In addition to predicting the next bit, one can also use masking losses and diffusion to achieve understanding. The general intuition is to minimise the difference
-
Microsoft Copilot Actions: Starter Prompts & Preview Warnings
By
–
Microsoft Copilot "Action" feature got some starter prompts:
– Find the best restaurants
– Book a holiday rental
– Rent a car
– Make a dinner reservation "Actions are in preview and can make mistakes. Please monitor their work closely." -
AI-Generated Video Labeling On Social Media Platforms
By
–
I hope there is a plan for X (as well as Facebook, Instagram, YouTube, etc.) to clearly mark AI-generated videos, like those from Veo. I assume there is a way to know through @GoogleDeepMind SythID. It seems irresponsible at this point to not publicly tag them on social media.