Modernize your marketing with AdQuick

AdQuick unlocks the benefits of Out Of Home (OOH) advertising in a way no one else has. Approaching the problem with eyes to performance, created for marketers with the engineering excellence you’ve come to expect for the internet.

Marketers agree OOH is one of the best ways for building brand awareness, reaching new customers, and reinforcing your brand message. It’s just been difficult to scale. But with AdQuick, you can easily plan, deploy and measure campaigns just as easily as digital ads, making them a no-brainer to add to your team’s toolbox.

Learn more now.

Today:

• OpenAI Removes ‘io’ Branding References

• China Launches Dynamic AI Benchmark

• Salesforce Launches Agentforce 3 Platform

• AI Beats Math Tests With Games

• Disney Explores Partnership With OpenAI

OpenAI Removes ‘io’ Branding References

OpenAI erased references to “io,” the hardware startup founded by ex-Apple designer Jony Ive, from its website and feeds after a legal dispute from hearing-aid company IyO, which claims the name is too similar. OpenAI says its $6.5 billion plan to buy Ive’s team and build new AI devices is still on track today. The blog post and video with CEO Sam Altman and Ive were temporarily removed by court order.

WHY IT MATTERS

1. Hardware push – OpenAI’s commitment to specialized devices shows big labs are moving beyond software to control the full stack of AI experience, from chips to gadgets.
   

2. Brand battles – The naming dispute reveals how intellectual-property fights can stall fast-moving AI projects and shape which products reach users first.
   

3. Design meets research – Merging Jony Ive’s hardware talent with OpenAI’s models could set new standards for how people physically interact with AI, influencing future products across the industry.

China Launches Dynamic AI Benchmark

HongShan Capital’s new test suite, Xbench, rates AI systems on textbook knowledge and real-world work. It mixes classic quiz sections with job-style tasks like recruiting engineers or picking marketing partners, then posts the scores. Part of the question pool is now open-source, free for anyone, and the list will refresh every quarter. The moving target makes cheating hard and shows how well models can research, reason, and create useful value.

WHY IT MATTERS

• Rolling, ever-changing tests stop models from memorizing answers, giving a truer read on reasoning progress.
  

• Real-world tasks reveal the money-making skills investors and companies need, linking lab scores to business value.
  

• Open access leaderboard invites global scrutiny and drives healthy competition among Western and Chinese models alike.

Salesforce Launches Agentforce 3 Platform

Salesforce’s Agentforce 3 brings real-time dashboards, faster models and open standard links to make AI agents reliable at large companies. A new Command Center tracks each agent’s work, giving alerts and easy handoff to humans. Built-in support for the Model Context Protocol lets agents plug into business apps without custom code. Lower latency, stronger security, and 200 ready-made actions help firms like PepsiCo cut costs, speed decisions and scale staff.

WHY IT MATTERS

• Gives companies clear visibility and control over autonomous agents, making large-scale use safer.
  

• Embraces the open MCP standard, nudging the AI field toward plug-and-play agent ecosystems.
  

• Shows agents already saving money and time for big brands, proving they’re moving from demos to core operations.

🧠RESEARCH

MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents

MEM1 is a new AI system that helps agents remember and reason better over long conversations. Instead of storing everything, it smartly keeps only important info, cutting memory use and speeding up performance. It beats larger models in complex tasks while using less memory, making AI agents more efficient and scalable.

Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights

Drag-and-Drop LLMs (DnD) let large language models adapt instantly to new tasks without retraining. Instead of fine-tuning for each task, DnD uses a prompt to generate custom model weights in seconds. This method is faster, more efficient, and performs better on unseen challenges like reasoning, coding, and multimodal tasks.

Towards AI Search Paradigm

The AI Search Paradigm is a new design for smarter search engines that act more like humans. It uses four AI agents—Master, Planner, Executor, and Writer—that work together to understand queries, plan tasks, use tools, and generate answers. This system handles both simple and complex searches, making search faster, more accurate, and more adaptable.

🛠️TOP TOOLS

Llama Coder - Open-source web application that generate entire applications from simple natural language prompts.

Kubee - AI tool designed to enable users to create, interact, and customize their own digital avatars.

Humbot AI - Designed to transform AI-generated text into human-like writing, effectively evading AI detection systems.

Durable AI - AI that builds a website for you

Cursor AI - Cutting-edge technology company that has revolutionized the way software developers write and manage code.

📲SOCIAL MEDIA

🗞️MORE NEWS

• Researchers trained AI to reason mathematically by playing games like Snake and Tetris instead of using math data. This “visual game learning” approach beat traditional models in geometry and reasoning tasks, showing games can teach thinking.

• Disney is suing AI startup Midjourney for using its characters without permission, while also exploring partnerships with firms like OpenAI. This dual strategy aims to protect its intellectual property as AI technologies rapidly advance.

• Goldman Sachs has rolled out its AI assistant to all employees, helping with tasks like document summaries and data analysis. It supports multiple language models and aims to eventually act and think like a Goldman employee.

• Leaked code reveals xAI is building a Grok-powered file editor with spreadsheet support. This could rival Google and Microsoft tools and move Elon Musk’s X platform closer to becoming an all-in-one productivity and social app.

• Mark Zuckerberg is personally contacting top AI experts to join Meta’s new Superintelligence lab, offering $100 million pay packages. The aggressive recruitment push aims to close Meta’s gap in the AI race.

• Harvey AI, which automates legal work, raised $300M at a $5B valuation just four months after its previous round. Now expanding into areas like tax accounting, it’s rapidly hiring and growing revenue at a fast pace.

• Meta explored acquiring AI video startup Runway but talks fell through. This came before its $14.3B investment in Scale AI and reflects Mark Zuckerberg’s broader push to secure top AI talent and tools in 2025.