LaunchGoogle Gemma 4 release
Google launches Gemma 4, a new open-source model: How to try it
Google just released its newest AI model, Gemma 4, which is now both open-weight and open source. mashable.com
Friday, 3 April 2026
42 articles
Executive summary of events for the last 24 hours
Google launched Gemma 4, a new open-source large language model family built on Gemini 3 technology, while Microsoft unveiled three new foundational models to compete in the rapidly evolving AI landscape. Additionally, enterprise adoption of agentic AI accelerated with Domo's new AI Agent Builder and increased investments from US card networks in autonomous AI systems.
List of sourced links used in the brief
LaunchMicrosoft multimodal foundation models
Microsoft takes on AI rivals with three new foundational models
MAI released models that can transcribe voice into text as well as generate audio and images after the group's formation six months ago. techcrunch.com
LaunchGoogle Gemma 4 release
Google releases Gemma 4, a family of open models built off of Gemini 3
When Google released Gemini 3 Pro at the end of last year, it was a significant step forward for the company's proprietary large language models. engadget.com
ResearchLLM behavior and interpretability
Emotion concepts and their function in a large language model
All modern language models sometimes act like they have emotions. They may say they're happy to help you, or sorry when they make a mistake. anthropic.com
Launchopen-source reasoning model
Arcee AI Releases Trinity Large Thinking: An Apache 2.0 Open Reasoning Model for Long-Horizon Agents and Tool Use
The landscape of open-source artificial intelligence has shifted from purely generative models toward systems capable of complex, multi-step reasoning. marktechpost.com
ResearchLLM application in materials science
Predicting new research directions in materials science using large language models and concept graphs
Due to an exponential increase in published research articles, it is impossible for individual scientists to read all publications, even within their own... nature.com
LaunchAlibaba Qwen model release
Alibaba Releases Qwen 3.6-Plus AI Model With Enhanced Coding Capabilities
Alibaba Group Holding Ltd. on Thursday released Qwen 3.6-Plus, a new large language model it says markedly improves automated coding and AI-agent... caixinglobal.com
Researchuncertainty quantification in LLMs
Better Uncertainty Quantification for Large Language Models
Improving uncertainty quantification in LLMs, this method combines epistemic and aleatoric uncertainty, leading to better error detection and model... azorobotics.com
Launchproduct launch
Domo Launches AI Agent Builder, MCP Server to Connect Enterprise Data to AI Ecosystem
Domo's AI Agent Builder and MCP Server connect enterprise data to AI ecosystems, enabling businesses to orchestrate AI-driven workflows. demandgenreport.com
Launchmodel release
Bring state-of-the-art agentic skills to the edge with Gemma 4
Google DeepMind introduces Gemma 4, a family of state-of-the-art open models designed for on-device agentic workflows. Learn how to leverage multi-step... developers.googleblog.com
Newsmajor adoption
US card networks accelerate bets on agentic AI
Key insights: Visa and Mastercard have debuted new products that use agentic artificial intelligence. What's at stake: The card networks' moves will make... americanbanker.com
Newssafety and access control
Control which domains your AI agents can access
AI agents that can browse the web open powerful possibilities—from research automation to real-time data gathering. However, giving an AI agent unrestricted... aws.amazon.com
Opiniongovernance and alignment
Why Agentic AI Needs Boundaries Before Freedom
As AI agents evolve from chat tools to actionable systems, the core challenge is how they can be contained, governed and aligned with human judgment. unu.edu
Newsmonitoring and governance
Exabeam expands platform to monitor AI agent activity | brief | SC Media
The expanded platform transforms AI agent services into sources of behavior telemetry, feeding directly into Exabeam's threat detection workflows. scworld.com
Newsenterprise adoption
Scaling Agentic AI: Why Transactional Messaging and a Converged Database Matter
Agentic AI is moving quickly from experimentation to real enterprise adoption. Many organizations are exploring agents that may plan, reason, use tools,... blogs.oracle.com
Researchinterpretability
Emotion concepts and their function in a large language model
Interpretability research from Anthropic on emotion concepts. anthropic.com
Newsgovernance and policy
Trump administration appeals ruling that blocked Pentagon action against Anthropic over AI dispute
The Trump administration is appealing a judge's order blocking the federal government from taking punitive measures against artificial intelligence company... tribdem.com
Newsgovernance and risk
Agentic AI: Greater Capabilities and Enhanced Risks
An examination of the legal risks and challenges arising from the rapid adoption of agentic AI and the shifting global regulatory landscape,... reuters.com
NewsCopilot Sales Strategy
Microsoft Hit ‘Audacious’ Copilot Goals After Analyst Input
Microsoft Corp., responding to Wall Street feedback, has pivoted its AI sales strategy to focus on selling Copilot rather than offering it for free as part... bloomberg.com
NewsAI Training Data & Privacy
GitHub Will Use Copilot Interaction Data from Free, Pro, and Pro+ Users to Train AI Models
GitHub will use Copilot interaction data from Free, Pro, and Pro+ users to train AI models starting April 24, opting in by default. infoq.com
NewsEnterprise AI Security
As Microsoft expands Copilot, CIOs face a new AI security gap
Earlier this week, Microsoft expanded its Copilot capabilities with new features designed to provide a persistent AI co-worker across enterprise workflows. informationweek.com
NewsAI-Native Development
Inside Shopify’s AI-first engineering playbook
Shopify has generated a lot of buzz about AI-native development, much of which has been fueled by viral posts from CEO Tobi Lütke and other leaders... bvp.com
NewsCopilot Adoption
Microsoft executive touts Copilot sales traction as AI anxiety weighs on stock
Microsoft adjusted its approach to selling the Microsoft 365 Copilot artificial intelligence productivity add-on to corporate workers after feedback from... cnbc.com
NewsCopilot Sales
Microsoft Defends Copilot Sales as AI Doubts Hit Stock
Microsoft executive publicly defends Copilot sales momentum amid mounting analyst skepticism over the $30/month AI assistant's adoption rate. techbuzz.ai
OpinionAI Market Segmentation
Copilots vs. Agents
Been thinking about market segmentation and patterns of investment in AI at the application layer. I'm sure someone deeper in the weeds in that area has... healthapiguy.substack.com
LaunchFinance AI Tool
TradingView AI Chart Copilot: public beta is available now
Charts don't speak, but now they can answer your questions. We introduce the TradingView AI Chart Copilot, an AI-powered assistant that lives in your... tradingview.com
NewsGoogle Vids avatar and Veo 3.1 integration
Google now lets you direct avatars through prompts in its Vids app
Google on Thursday added new features to its video editor app Vids, including directing and customizing avatars through text prompts, Veo 3.1 support,... techcrunch.com
NewsMicrosoft MAI foundational models
Microsoft's MAI unveils trio of foundational AI models
Microsoft's MAI division launches three foundational models: voice-to-text transcription, audio generation, and image synthesis, per TechCrunch. techbuzz.ai
NewsMicrosoft voice and image AI models
Microsoft's New AI Models Go Beyond Just Text
Microsoft is flexing its muscles with new voice, transcription and image AI models. aol.com
NewsGoogle Vids Veo 3.1 integration
Google is turning video creation into a one-click experience with latest Vids updates
Google Vids just integrated Veo 3.1, letting you create polished video clips from a simple text or image prompt. Every Google account gets 10 free... androidauthority.com
LaunchMicrosoft AI models for image, speech, transcription
Microsoft Releases AI Models for Transcription, Speech, Image Generation
Microsoft on Thursday released three new AI models, building on the small but growing catalog of models it has trained internally as it aims to become more... theinformation.com
NewsCapCut Dreamina Seedance 2.0 integration
AI Video Creation Tools
Dreamina Seedance 2.0 - CapCut's integration of Dreamina Seedance 2.0 signals a shift toward fully automated content creation workflows, where video... trendhunter.com
LaunchPixExact platform launch
PixExact Launches AI Platform Enabling Pixel-Perfect Image and Video Generation for Modern Content Workflows
United States, April 2, 2026 -- PixExact today announced the launch of its AI-powered creative platform designed to deliver image and video generation with. aijourn.com
Opinionhumanoid robot development strategy
Gill Pratt Says Humanoid Robots’ Moment Is Finally Here
AI for humanoid robots reveals why data, not hardware, now decides which robots safely handle real-world tasks without overhyped promises. spectrum.ieee.org
Newsrobot manipulation and learning
Sanctuary AI’s robotic hand demonstrates zero-shot in-hand manipulation
Sanctuary AI said the robotic hand and AI system achieves the target orientation 10 times in a row without dropping the cube. therobotreport.com
ResearchML applications
FuXi-Air: air quality forecasting based on emission-meteorology-pollutant multimodal machine learning
Air pollution has emerged as a major public health challenge worldwide. Numerical simulations and single-site machine-learning approaches in air quality... nature.com
ResearchML applications
Machine learning uncovers thousands of unknown bacterial immune defense systems
Bacterial immune systems protect against viral invaders called phages by precisely targeting specific phage genetic sequences. news-medical.net
Researchbiological computing
Living Brain Cells Enable Machine Learning Computations
A collaborative research group has shown that biological neurons can be trained to perform a temporal pattern learning task that was previously carried out... asiaresearchnews.com
PolicyUS chipmaking equipment export restrictions
US Lawmakers Propose Crackdown on Chip Tool Sales to China
US lawmakers unveiled legislation that seeks to crack down on exports of chipmaking tools to China, especially from allies including the Netherlands and... bloomberg.com
LaunchAMD Gemma 4 support
Day 0 Support for Gemma 4 on AMD Processors and GPUs
AMD is proud to provide Day Zero support for the full set of Gemma 4 models across our portfolio of AI-enabled hardware. This includes AMD Instinct™ GPUs... amd.com
PolicyUS chip export controls to China
Bill to ban sale of key AI chipmaking equipment to China introduced in House
A bipartisan group of senators introduced a bill that would severely limit the sale of advanced semiconductor machinery to China used for AI systems. nbcnews.com
LaunchNVIDIA Gemma 4 acceleration
From RTX to Spark: NVIDIA Accelerates Gemma 4 for Local Agentic AI
Gemma 4 brings powerful reasoning, coding and multimodal AI directly to NVIDIA RTX PCs, DGX Spark and edge devices. blogs.nvidia.com
Support the project
AIskimIQ is an independent project. If you find it useful, you can support its development with a coffee.
