Highlights
Top Insights
AI can dramatically reduce the chaos caused by complex, fast-changing tariffs by creating dynamic digital twins of supply chains, automating compliance classification, and optimizing pricing and inventory strategies in real time.
AI doesn’t just aid decision-making—it enables organizations to simulate future scenarios and adapt as conditions evolve, making scenario planning and strategic agility a competitive edge.
Practical success will depend less on avoiding tariffs and more on building a strong human + AI collaboration model, supported by quality data, empowered cross-functional teams, and emotionally resilient leadership.
Source: How AI can help address tariff challenges (Board of Innovation)
Top News
1. Meta introduced the new Llama 4 models—Scout, Maverick, and the teacher-model Behemoth—marking a leap forward in multimodal AI with industry-leading performance.
2. OpenAI has upgraded ChatGPT with long-term memory, enabling it to recall and reference past conversations—even those not manually saved.
3. ElevenLabs and Google adopt Anthropic’s open Model Context Protocol.
4. Google Cloud has launched the Agent2Agent (A2A) protocol, an open standard supported by over 50 partners.
5. Articulate Medical Intelligence Explorer (AMIE), a diagnostic large language model, significantly improves clinicians’ diagnostic accuracy and comprehensiveness in complex cases.
Additional Insights
1. How People Are Really Using Gen AI in 2025 (Harvard Business Review)
In 2025, generative AI has seen a significant evolution, with personalized tools like Custom GPTs, increased competition from new players like Grok and DeepSeek, and broader, cheaper access driving rapid adoption. The most prominent trend is a shift from technical to emotional and personal use cases, with therapy, life organization, and finding purpose topping the charts, reflecting AI’s growing role in self-actualization. Users are becoming more sophisticated, using AI not just to solve tasks but to improve thinking, learning, and well-being—while also expressing concerns about overdependence, data privacy, and political bias. Forums like Reddit reveal a diverse range of real-world use cases, from appealing fines and planning travel to managing health and engaging in deep conversations. Despite polarized opinions, the prevailing insight is that AI is now embedded in both the practical and emotional dimensions of daily life, and its development continues to shape how we live, think, and relate.
2. How AI can help supercharge creativity (MIT Technology Review)
Artists and researchers are exploring how AI can enhance creativity—not by automating it, but by adding surprise, friction, and dialogue into the creative process. Rather than relying on one-click generative tools that risk dulling originality, they’re building AI collaborators that challenge assumptions, provoke reflection, and inspire unexpected outcomes. This co-creative approach treats AI as a medium rather than a magic tool, encouraging deeper engagement and skill-building through iteration, failure, and experimentation. However, many current tools prioritize polished results over creative control, leaving a gap between commercial usability and experimental flexibility. The future of AI-powered creativity lies in developing tools that balance surprise with agency, empowering more people—not just coders—to engage in truly original artistic expression.
3. Three Takeaways On The Future Of AI From The Nvidia Conference (Forbes)
At Nvidia’s GPU Technology Conference, key insights underscored that AI has transitioned from experimental pilots to becoming central to business strategy and operations. Smaller, specialized AI models are now transforming economic viability by significantly reducing costs, enabling broader adoption, and reshaping competitive dynamics through innovation. Agentic AI, particularly semiautonomous systems, is gaining momentum, demanding structured design and careful integration to ensure trust, transparency, and effectiveness. Multimodal generative AI tools are accelerating creative processes, enabling rapid, personalized content generation at scale, substantially reducing marketing timelines and content production costs.
Innovation Radar
1. AI Model Releases and Advancements
Meta introduced the new Llama 4 models—Scout, Maverick, and the teacher-model Behemoth—marking a leap forward in multimodal AI with industry-leading performance, advanced mixture-of-experts architectures, unprecedented 10-million-token context lengths, enhanced visual-text integration, efficient inference, and strengthened safeguards, all openly available to drive innovation across diverse AI applications (Meta).
Midjourney has released V7, its first new AI image model in nearly a year, featuring a new architecture with enhanced image quality, built-in personalization, faster rendering via a Draft Mode, and early support for smarter text prompts—though some features like upscaling are still pending (TechCrunch).
Amazon has released Nova Reel 1.1, a powerful update to its text-to-video AI model that enables the creation of high-quality, multi-shot videos up to two minutes long with improved consistency, creative control, and support for both automated and manual storyboard-style prompting via Amazon Bedrock (Amazon).
ByteDance introduced Seed-Thinking-v1.5, a compact Mixture-of-Experts model trained with reinforcement learning that achieves state-of-the-art reasoning performance across STEM, coding, and general tasks, outperforming competitors while maintaining broad generalization and efficiency (GitHub).
Anthropic has launched the Max plan for Claude, offering up to 20x higher usage limits and priority access to new features and models for users who rely heavily on Claude for frequent, in-depth collaboration across complex tasks (Anthropic).
2. AI Tools and Features
OpenAI has upgraded ChatGPT with long-term memory, enabling it to recall and reference past conversations—even those not manually saved—to deliver more personalized responses, with the feature currently rolling out to paid users outside the EU and other highly regulated regions (The Verge).
Microsoft is ushering in a new era of personal computing with Copilot, an AI companion designed to deeply understand and assist each user through personalized memory, real-time vision, advanced actions, and dynamic tools like Pages, Podcasts, and Deep Research—empowering individuals while ensuring they remain in control (Microsoft).
At Google Cloud Next 2025, Google unveiled powerful new AI infrastructure like the Ironwood TPU and Cloud WAN, alongside advanced models like Gemini 2.5 and cost-efficient Gemini 2.5 Flash, all designed to accelerate enterprise innovation, enhance AI reasoning, and make cutting-edge AI capabilities widely accessible (Google).
ElevenLabs has launched its Model Context Protocol (MCP) server, a powerful and accessible platform for advanced audio processing—including text-to-speech, voice cloning, transcription, and soundscape creation—designed for seamless integration with popular clients and offering a free tier to support a wide range of creative and professional applications (Geeky Gadgets). Google will adopt Anthropic’s open Model Context Protocol (MCP) for its Gemini models and SDK (TechCrunch).
Google Cloud has launched the Agent2Agent (A2A) protocol, an open standard supported by over 50 partners that enables AI agents—regardless of vendor or framework—to securely communicate, collaborate, and complete complex tasks across enterprise systems, marking a major step toward seamless multi-agent interoperability in business environments (Google).
Elon Musk’s AI company xAI has launched an API for its Grok 3 model and its smaller variant, offering advanced image analysis and reasoning capabilities at premium prices, despite criticisms over benchmark transparency, limited context window, and political bias (TechCrunch).
Writer has launched AI HQ, a new enterprise platform featuring autonomous AI agents and self-evolving models designed to automate complex workflows, enhance productivity, and transform businesses into AI-native organizations by bridging the gap between generative AI’s potential and real-world impact (VentureBeat).
Canva has launched Visual Suite 2.0, its most significant update yet, introducing a unified design platform with AI-powered tools like Canva AI, Sheets, Magic Charts, and Canva Code to revolutionize how teams create, collaborate, and scale content across documents, data, and interactive experiences (Business Wire).
3. AI Research and Social Impact
A study introducing the Articulate Medical Intelligence Explorer (AMIE), a diagnostic large language model, found that it significantly improves clinicians’ diagnostic accuracy and comprehensiveness in complex cases compared to standard tools alone (Nature). AMIE outperformed primary care physicians in diagnostic accuracy and multiple clinical communication metrics during simulated text-based consultations, marking a significant advancement toward conversational diagnostic artificial intelligence (Nature).
4. Other
Researchers have created the most detailed brain map to date by reconstructing a cubic millimetre of a mouse’s visual cortex, revealing the structure and activity of over 200,000 brain cells—including 82,000 neurons and 500 million synapses—offering unprecedented insights into how brain circuits process information (Nature).
Researchers at Washington University have developed a new biosensor that can detect airborne bird flu in just five minutes, potentially enabling early outbreak prevention and helping protect food supplies and public health (MIT Technology Review).
Researchers have developed a “double neural bypass” brain-computer interface that not only translates brain signals into movement but also restores tactile feedback, enabling a paralyzed patient to move and feel with his hand, significantly enhancing motor function and sensory experience (IEEE Spectrum).
Leave a Reply