All Stories
The Potential of CoT for Reasoning: A Closer Look at Trace Dynamics
Chain-of-thought (CoT) prompting is a de-facto standard technique to elicit reasoning-like responses from large language models (LLMs), allowing them to spell out individual steps b...
In AI, Research, Feb 26, 2026
See the whole picture and find the look with Circle to Search
<img src=”https://storage.googleapis.com/gweb-uniblog-publish-prod/images/FTL__Try_On_Blog_header.max-600x600.format-webp.webp”>We’ve updated Circle to Search so you can now e...
In AI, Research, Feb 26, 2026
responsible-scaling-policy-v3
Policy Announcements Anthropic’s Responsible Scaling Policy: Version 3.0 Feb 24, 2026 Read the Responsible Scaling Policy We’re releasing the third version of our Responsible Scalin...
In AI, Research, Feb 26, 2026
Nano Banana 2: Combining Pro capabilities with lightning-fast speed
<img src=”https://storage.googleapis.com/gweb-uniblog-publish-prod/images/NB2_Hero.max-600x600.format-webp.webp”>Our latest image generation model offers advanced world knowle...
In AI, Research, Feb 26, 2026
detecting-and-preventing-distillation-attacks
Announcements Detecting and preventing distillation attacks Feb 23, 2026 We have identified industrial-scale campaigns by three AI laboratories—DeepSeek, Moonshot, and MiniMax—to il...
In AI, Research, Feb 26, 2026
depyf: Open the Opaque Box of PyTorch Compiler for Machine Learning Researchers
PyTorch \texttt{2.x} introduces a compiler designed to accelerate deep learning programs. However, for machine learning researchers, adapting to the PyTorch compiler to full potenti...
In AI, Research, Feb 26, 2026
Constructive Circuit Amplification: Improving Math Reasoning in LLMs via Targeted Sub-Network Updates
Prior studies investigating the internal workings of LLMs have uncovered sparse subnetworks, often referred to as circuits, that are responsible for performing specific tasks. Addit...
In AI, Research, Feb 26, 2026
Closing the Gap Between Text and Speech Understanding in LLMs
Large Language Models (LLMs) can be adapted to extend their text capabilities to speech inputs. However, these speech-adapted LLMs consistently underperform their text-based counter...
In AI, Research, Feb 26, 2026
Build with Nano Banana 2, our best image generation and editing model
<img src=”https://storage.googleapis.com/gweb-uniblog-publish-prod/images/BuildWith_Hero.max-600x600.format-webp.webp”>Nano Banana 2 (Gemini 3.1 Flash Image) delivers Pro-leve...
In AI, Research, Feb 26, 2026
Beyond a Single Extractor: Re-thinking HTML-to-Text Extraction for LLM Pretraining
One of the first pre-processing steps for constructing web-scale LLM pretraining datasets involves extracting text from HTML. Despite the immense diversity of web content, existing ...
In AI, Research, Feb 26, 2026
Featured
-
Google Launches Gemma 4: Agent-Ready Models from Mobile to Workstation
In AI, Open Source, LLM, -
The AI Coding Wars Heat Up: Claude Opus 4.6 vs GPT-5.3-Codex
In AI, Machine Learning, Coding, -
The AI Coding Wars Heat Up: Claude Opus 4.6 vs GPT-5.3-Codex
In AI, Machine Learning, Coding, -
Weekly Journals
In Journal, -
The Journey to the FIRE app
In AI, -
AI Journey
In AI, -
Data Structure and Algorithm Journey
In Algorithm, -
New Programming Languages Tour
In Java, Kotlin, Swift, -
DevOps Journey
In DevOps, -
Google Cloud Certified Professional Cloud Architect Journey
In Cloud, -
Architecture Journey
In Architecture,