The best award I have ever received ❤ thanks @TensorFlow

@GoogleDeepMind If you think this app is good for you, please UPVOTE 🙏 Also, it is an incredible experience of building this complex app in AI Studio only. Pure vibe coding rules https://t.co/OIeyzyKR2H

@GoogleDeepMind Finally turn your daily logs into a well organized blog posts including text, images, and the apps that you built for certain activities

Can it really be a coincidence? Smart Investors Hold Gold. Find out more at Goldhub

@GoogleDeepMind Get reflection of your life based on every activities. It's hard to know how which your small activity helped shaping your longterm goal.

@GoogleDeepMind About dynamic mini web app building, you can generate two types; interaction focused and learning content rich focused.

@GoogleDeepMind Log your day and make it content rich. AI assisted logging from banner image creation, dynamic mini web app building to high quality image generation

Big Pharma hijacked the natural process of aging and rebranded it as a disease. Now they can sell you a lifetime of pills to "manage" it. They don't want you to know that the key to aging well isn't a drug. It’s molecule your own body makes: NAD+. 1.… https://t.co/imXIc6gKWp

@GoogleDeepMind Design your day by having fixed schedules and flexible tasks that you wish to achieve at any given free time.

My entry for the Gemini hackathon held by @GoogleDeepMind . It shows how to manage your day in more intelligence powered by Gemini models. Below 👇 https://t.co/OIeyzyKR2H

You can now train LLMs 3× faster with no accuracy loss, via our new RoPE and MLP kernels. Our Triton kernels plus smart auto packing delivers ~3× faster training & 30% less VRAM vs optimized FA3 setups. Train Qwen3-4B 3x faster on just 3.9GB VRAM. Blog: https://t.co/j3HpsmzrCw

🚀 Paper2Slides is now open source! Transform research papers & technical reports into professional presentations with ONE click! We've generated stunning presentation slides from the latest DeepSeek V3.2 paper in diverse styles - check them out and share your feedback! 🔥 Core

Lenders can more easily work with state and local agencies to find grants and programs that may help borrowers afford a home thanks to our streamlined resources. Learn how.

This is how you get 100% accuracy in Nano Banana Pro image generation. Use JSON prompts. Here's how to you can write JSON prompts to get shockingly accurate outputs from Nano Banana Pro easily:

Our Gemini 3 Vibe Code hackathon started!, Build applications using the new Gemini 3 Pro model with a price pool of $500k. 🤯 > Top 50 winners receive $10,000 in Gemini API credits each. > Access Gemini 3 Pro Preview directly in Google AI Studio. > Leverage advanced reasoning

gemini 3 vibe coding hackathon starts now build your best app to solve a real-world problem across science, health, education or business. $500,000 in Gemini API credit prizes competition ends dec 12

.@NanoBanana is great to visualize Jupyter notebooks since their contents are well organized for education purpose. Just upload a notebook (ipynb or PDF) and ask it to generate a diagram for you! Here is an example of @UnslothAI gpt-oss GRPO.

Atlas: “I was sitting across from Anthony Fauci in August of 2020, and he looked at me and said, ‘One of the biggest problems is the American public is not afraid enough.’ This was the task force… it’s a completely unethical way to do public health.”

🥳We recently launched the Colab Data Explorer- a new feature that lets you search Kaggle datasets, models, and competitions directly from a Colab notebook! 🔍 Easily import data with a KaggleHub code snippet and use integrated filters to find exactly what you need. Learn more

We have rewritten entire CUDA programming guide! Check the new content!

We used Claude Code to train open LLMs. Check out the tutorial. basically, we plugged HF skills into claude code and it was able to train LLMs end-to-end. Best thing, this works on all major coding agents: Codex, Cursor, and Gemini CLI. - You tell the agent to fine-tune a model

Today at #NeurIPS2025, we present Titans, a new architecture that combines the speed of RNNs with the performance of Transformers. It uses deep neural memory to learn in real-time, effectively scaling to contexts larger than 2 million tokens. More at: https://t.co/pEuI5oDkkn

Trump says NEWSMAX is 'terrific!' Click below to find out why...

this is big... 50 AI researchers from Bytedance, Alibaba, Tencent, and other labs/universities just published a 300-page paper with surprising lessons about coding models and agents (data, pre and post-training, etc). key highlights: > small LLMs can beat proprietary giants RL

You can now do 500K context length fine-tuning with Unsloth! Train gpt-oss-20b to extend its context window to 530K on 80GB VRAM & 750K+ on 192GB - no accuracy loss. Unsloth's new algorithms + Tiled MLP = 72% less VRAM & 6x more context Blog + Notebook: https://t.co/Qe4hjTe6CI

Introducing the Mistral 3 family of models: Frontier intelligence at all sizes. Apache 2.0. Details in 🧵