Schedule with ChatGPT, Why o1 Thinks in Chinese and Framework for U.S. AI Regulation

PLUS: + 5 Best AI Tools for Ads Creation

AI Innovators, Gather Round


СhatGPT introduces scheduling reminders, a groundbreaking mathematical technique is shedding light on AI decision-making, OpenAI’s reasoning model, o1, has baffled experts by processing tasks in unexpected languages like Chinese, raising questions about AI training processes. Lastly, OpenAI proposes a regulatory framework for U.S. AI governance, highlighting the growing urgency for ethical oversight in the industry. Stay ahead with the latest updates!🧠

In today’s knowledge-packed edition:

  • Is Your 2024 AI Project Stuck? Let’s Fix That

  • ChatGPT Now Lets You Schedule Reminders

  • The Future of AI Agents in Enterprise

  • Mathematical Technique 'opens the black box' of AI Decision-Making

  • OpenAI’s AI Reasoning Model ‘thinks’ in Chinese Sometimes and No One Really Knows Why

  • OpenAI Proposes Framework for U.S. AI Regulation

  • 5 Best AI Tools for Ads Creation

Massive investments in AI infrastructure last year mean 2025 is the year to prove its value. Tumeryk’s AI Trust Score™ is here to help you move from experimentation to scalable success by building trust and security into your Gen AI deployments.

Why Tumeryk’s AI Trust Score™?

🔍 Quantify Risk: Gain a clear Trust Score for foundational models and custom apps.

🛡️ Ensure Security: Centralized policies safeguard against issues like prompt injection and hallucination—and prepare for tomorrow’s challenges.

💰 Cut Costs: Save 30%+ on public model token usage with role-based query management.

 Stay Compliant: Generate detailed reports and vulnerability scores for seamless due diligence.

Boost Speed: Launch production-ready systems in hours—not weeks—and empower your developers to deliver faster.

Ready to unlock your AI’s potential in 2025?

BRYCE DURBIN

OpenAI has introduced a new beta feature called “Tasks” to ChatGPT, enabling users to schedule reminders and manage recurring tasks, positioning ChatGPT as a competitor to virtual assistants like Siri and Alexa.

With “Tasks,” users can set one-time reminders, such as notifications for upcoming events, and establish recurring actions like daily weather updates. ChatGPT may also suggest tasks based on user interactions, which users can choose to accept or decline.

This feature is currently rolling out in beta to Plus, Team, and Pro users globally, starting with the web platform. Users can manage their tasks through chat threads or a dedicated Tasks section, with notifications available across web, desktop, and mobile platforms. OpenAI has set a limit of 10 active tasks per user at this time.

The introduction of “Tasks” signifies OpenAI’s commitment to enhancing ChatGPT’s functionality, aiming to provide a more versatile and user-friendly experience for managing daily activities.

Join our LinkedIn Live to explore how AI agents are reshaping enterprises! Hosted by Steve Nouri, CEO of GenAI.Works, with industry leaders Assaf Ellovic (Head of AI, Monday.com) and Rotem Alaluf (CEO, Wand AI), this session dives into the cutting-edge world of AI in enterprise.

What You’ll Learn:

 AI Agents in Action: See how AI agents are transforming job functions, not just assisting with tasks.

 Human-AI Collaboration: Discover how agents scale operations, accelerate time-to-market, and adapt in real time.

 Bridging Skill Gaps: Learn how AI agents are driving productivity and achieving ambitious goals.

 Next-Level Innovation: Explore breakthroughs shaping the future of AI in enterprise.

Why Attend?

✔️ Expert Insights: Gain knowledge from leading AI voices.

✔️ Live Q&A: Engage directly with experts.

✔️ Networking: Connect with professionals and enthusiasts.

Speakers:

🌟 Assaf Ellovic, Head of AI at Monday.com

🌟 Rotem Alaluf, CEO of Wand AI

🌟 Steve Nouri, CEO of GenAI.Works

Don’t miss this opportunity to learn how AI agents are revolutionizing industries! Register HERE

Researchers have developed a novel technique using math to understand exactly how neural networks make decisions.

This technique allows analysis of neural network operations, effectively transforming previously black box AI systems into transparent, analyzable processes.

The team validated their approach through a systematic process:

  1. Initial validation using basic geometric shapes

  2. Progressive scaling to more complex image processing tasks

  3. Successful application to natural image segmentation

  4. Extension to broader computational applications

Read more 👇️ 

OpenAI’s o1 reasoning model has surprised users by intermittently processing reasoning steps in languages like Chinese, Persian, or Hindi, even when asked questions in English. This unexpected behavior, which OpenAI has yet to explain, occurs during multi-step reasoning tasks, where o1 uses other languages before delivering an English conclusion.

Experts speculate that this phenomenon could stem from training data biases, as reasoning models often include datasets labeled by third-party services in regions like China. These labels may influence the model’s linguistic patterns. Others argue that the model’s behavior reflects its probabilistic nature, selecting languages that optimize its reasoning efficiency rather than recognizing distinct languages.

AI researchers emphasize that the lack of transparency in AI model development limits understanding of such behaviors. Some suggest that the o1 model’s linguistic switching demonstrates its ability to leverage diverse language structures, but others warn that biases in training data and tokenization could have unintended consequences.

OpenAI has released an "economic blueprint" outlining its vision for the U.S. AI regulation and infrastructure development. 

OpenAI is calling for unified federal rules instead of the current state-by-state approach. For context, states introduced almost 700 different AI-related bills in 2024 alone, creating a complex patchwork of regulations.

They're also taking a strong stance on copyright issues, arguing that AI companies should be able to use publicly available information for training, while still protecting creators' rights. 

To read more about blueprint 👇️ 

🎯 5 Best AI Tools for Ads Creation

Unlock 10% off your first plan with the code “PARTNERQUICKADS10”.

Experience the free trial and create stunning ad visuals with ease.

Enjoy a 30-day free trial using the code “KITTLAFFILIATE”.

Boost your ad campaigns with advanced AI tools designed for performance.

Generate high-converting ad copy effortlessly with this intuitive platform.

Supercharge your ad campaigns with these top tools and take your marketing to the next level! 🚀

🚨 How to Ensure Our Emails Don’t End Up in Spam

This quick setup will help ensure you never miss an update! 🚀

🚀 Contact us to reach 6M+ of tech professionals, investors, engineers, managers, and business owners worldwide.

🌟 Sign up for the first AI Hub in the world.

Reply

or to participate.