OpenAI’s AI Device Plans 🤖, DeepSeek Bans In Australia 🌏, U.S. Crackdown 🚫

In partnership with

Create, Publish & Earn with Synthflow AI Voice Agents Marketplace

• Discover templates for routine/repetitive tasks like lead qualification and managing appointments.

• Publish your own Voice AI solutions to help businesses thrive—and earn commissions.

• Access custom actions that automate CRM updates, appointment scheduling, and more.

Build Your AI Agent Now

💬 OpenAI Developments

OpenAI’s new logoOpenAI’s big rebranding effort brings a new logo and a new typeface, OpenAI sans.

OpenAI has unveiled a comprehensive rebranding, introducing a redesigned logo and a new custom typeface called OpenAI Sans. The updated logo features a slightly larger central space and cleaner lines, while the new typeface blends geometric precision with a rounded, approachable character to convey a more humanized identity.

OpenAI Plans to ‘Replace Smartphones’ With New AI Device

OpenAI CEO Sam Altman has announced plans to develop an AI-focused hardware device in collaboration with former Apple chief design officer Jony Ive. This partnership aims to create a product that utilizes generative AI to handle complex user interactions more efficiently than traditional software, potentially offering a less socially disruptive alternative to current smartphones.

OpenAI's Deep Research smashes records for the world's hardest AI exam, with ChatGPT o3-mini and DeepSeek left in its wake

OpenAI's Deep Research, an AI agent integrated into ChatGPT, has achieved a 26.6% accuracy score on the "Humanity's Last Exam" benchmark, surpassing previous leaders such as DeepSeek's R1 and ChatGPT o3-mini. This significant improvement highlights rapid advancements in AI reasoning capabilities within a short period.

🌐 AI in Robotics

Belt and robot: Chinese start-up to open first robotics facility in Central Asia

Shanghai-based start-up AgiBot has partnered with Kazakhstan's Ministry of Digital Development, Innovation and Aerospace Industry to establish the first Chinese robotics manufacturing and research facilities in Central Asia. This collaboration aims to diversify Kazakhstan's economy by focusing on artificial intelligence, manufacturing, and green energy sectors.

AgiBot Establishes Robotics and AI Solutions Enterprise in Kazakhstan

AgiBot, a Shanghai-based leader in AI and robotics integration, has signed a memorandum of understanding with Kazakhstan's Ministry of Digital Development and Aerospace Industry to establish a joint venture for producing and implementing robotics and AI solutions in Kazakhstan. This partnership includes plans to launch a "data factory" for training robots, set up a joint R&D center, and collaborate with local universities to advance robotics expertise in the region.

🚀 Tech Industry Moves

Nvidia: Researchers from Shenzhen MSU-BIT University have developed the PD-General framework, an algorithm that enhances peridynamics simulations by optimizing algorithm design and memory management using Nvidia's CUDA technology. This advancement enables large-scale material simulations to be conducted on consumer-grade GPUs like the Nvidia RTX 4070, achieving up to 800 times faster performance compared to traditional serial programs, thereby reducing reliance on high-performance computing clusters.

Nvidia: CoreWeave has become the first cloud service provider to offer NVIDIA's GB200 NVL72-based instances, marking the general availability of the NVIDIA Blackwell platform. These instances feature a rack-scale solution with 72 interconnected GPUs and 36 Grace CPUs, delivering the performance necessary for advanced AI reasoning models and agents.

DeepSeek: Australia has prohibited the use of the Chinese artificial intelligence app DeepSeek on all government devices and systems, citing security concerns. This move aligns with similar actions taken by other countries worried about potential data security risks associated with the app.

DeepSeek: Taiwan has officially banned all government departments from using DeepSeek, a Chinese AI service, due to security concerns, including potential data transmission to China and issues related to censorship. This decision intensifies previous advisories and aligns with actions taken by other countries scrutinizing DeepSeek's data practices.

DeepSeek: U.S. Senator Josh Hawley has introduced the "Decoupling America’s Artificial Intelligence Capabilities from China Act," aiming to prohibit the import and export of AI technology between the United States and China due to national security concerns. If enacted, the legislation would criminalize the use of Chinese-developed AI applications like DeepSeek, imposing penalties of up to 20 years in prison and fines of $1 million for individuals and $100 million for companies.

Anthropic has introduced "constitutional classifiers," a system designed to prevent its AI models from generating harmful content by monitoring both inputs and outputs. To validate this system's robustness, Anthropic conducted over 3,000 hours of internal testing and is now inviting the public to attempt to "jailbreak" its latest AI model, Claude 3.5 Sonnet, during a week-long open test.

Microsoft's annual Build developer conference is scheduled from May 19th to May 22nd, 2025, at the Seattle Convention Center. The event is anticipated to feature numerous AI-related announcements, including updates on Microsoft's Copilot AI for consumers and businesses, as well as integrated AI features in Microsoft 365 subscriptions. Additionally, there are rumors of new Surface hardware launching around the same time as the conference.

Meta has announced that it may halt the development of AI systems it deems too risky, emphasizing its commitment to safety and ethical considerations. The company outlined specific scenarios in which AI models could be restricted, particularly if they pose threats related to misinformation, privacy, or security risks, aligning with increasing regulatory scrutiny on AI development.

⚖️ AI Ethics and Regulation

The U.S. Copyright Office has ruled that AI-generated art cannot be copyrighted if there is no significant human involvement in its creation. This decision emphasizes that copyright protection is reserved for works that reflect human creativity and originality, excluding fully autonomous AI-generated content. The ruling aligns with previous legal interpretations that only human authorship qualifies for intellectual property rights, reinforcing concerns about the legal status of AI-generated works in creative industries.

💡 Emerging AI Technologies and Companies

CodeSignal wants you to practice soft skills with its AI-powered guide

CodeSignal has launched AI-powered soft skills courses, expanding beyond technical training to help professionals improve leadership, communication, and negotiation skills. With over 100 courses, users can practice real-life scenarios using AI simulations, receive personalized feedback, and progress through competency levels. The platform offers free and paid plans ($24.99 for unlimited access) and enterprise options tailored to company culture. While AI provides valuable practice, experts warn about cultural sensitivity challenges and lack of emotional intelligence in feedback. CodeSignal plans to launch mobile apps later this year to broaden accessibility.

AMD pulls up the release of its next-gen data center GPUs

AMD is accelerating the launch of its next-gen data center GPUs, the Instinct MI350 series, moving production up to mid-2025 due to strong demand. CEO Lisa Su announced the change during AMD’s Q4 2024 earnings call, citing better-than-expected progress. While AMD’s data center revenue lags behind Nvidia’s, it saw $5 billion in AI chip sales in 2024 and expects double-digit growth in 2025. With major customers like Meta, Microsoft, and IBM, AMD aims to expand its data center business to tens of billions in the coming years. The news boosted AMD’s stock by 4.58%.

🎨 AI in Creative Industries

Swiss startup LogicStar has raised $3 million in pre-seed funding to develop AI agents focused on autonomous app maintenance, rather than just code co-development. Led by CEO Boris Paskalev, the company leverages large language models (LLMs) like OpenAI’s GPT to detect and fix software bugs efficiently, using a test-driven development approach. Targeting enterprises, LogicStar aims to complement human developers by handling routine maintenance tasks at scale. Currently supporting Python, the platform plans to expand to TypeScript, JavaScript, and Java. The funding round was led by Northzone, with backing from DeepMind, Sequoia scouts, and other industry investors.

If you want more updates related to AI, subscribe to our Newsletter