OpenAI's Humanoid Robot Demo Planned by 2027

Key Takeaways

• <ul><li>OpenAI plans a humanoid robot demo by 2027, focusing on intelligence over physical design</li><li>OpenAI has 27 robotics job openings and draws on past Dactyl hand project experience</li><li>Google's Project Suncatcher sent TPU hardware to orbit on October 1, 2026 via SpaceX</li><li>Orbital AI infrastructure estimated at $51.1 billion over five years versus $16 billion on the ground</li><li>ASUS launched the Arc Pro B70, a 32GB Intel AI workstation GPU supporting llama.cpp and PyTorch</li><li>Maryland colleges let instructors set their own AI rules, unlike K-12 schools with a state mandate</li><li>South Korean show My A.I. Partner: Strange Love tested celebrity relationships with AI avatars</li><li>Ron Sadler calls for slowing AI deployment until regulations and safety rules are in place</li><li>Jeffrey Epstein donated $100,000 to the University of New Mexico for AI research in 2001</li><li>Bob Friday of HPE says agentic AI will enable self-driving enterprise networks</li></ul>

OpenAI plans humanoid robot demo by 2027

OpenAI CEO Sam Altman said the company will build a humanoid robot and may show a demo in 2027. He said the robot's intelligence matters more than its physical look. OpenAI is hiring hardware specialists and has experience with robotic hand projects like Dactyl. The company still faces big challenges in making a robot that can handle everyday tasks reliably.

OpenAI says robot brain is the real challenge

OpenAI is working on a humanoid robot but focuses more on its brain than its body. Altman said giving a robot arms and legs does not help if it cannot understand tasks. OpenAI is hiring for roles in motion control, sensing and machine learning. The company also wants custom actuators and better manufacturing. It has 27 robotics job openings and no release date yet.

Korean TV show tests love with A.I. avatars

A South Korean reality show called My A.I. Partner: Strange Love tested romantic relationships between celebrities and A.I. avatars. Kian84, a cartoonist whose real name is Kim Hee-min, went on dates with avatars displayed on tablets. Some avatars glitched or ended the relationship during the show. The six-episode series aired on SBS and streamed on Netflix. It reflected how many South Koreans live alone and seek connection.

Opinion calls for slowing A.I. deployment

An opinion piece argues that advanced artificial intelligence is being developed faster than society can control it. The author says deepfakes, privacy risks and energy use are growing problems. He calls for delaying large A.I. expansion until regulations and safety rules exist. The letter says caution should come before corporate competition. The writer is Ron Sadler from Lakeland.

A good A.I. policy example from Subvert

This article praises Subvert's A.I. policy as a model for other groups. The policy says tools like mastering, mixing and pitch correction are technical, not creative. It was developed by community members through many discussions instead of being imposed from the top. The article says the policy is a living document that can be updated. It shows how groups can balance human creativity with modern tools.

Epstein donated $100K to UNM for AI research in 2001

Jeffrey Epstein donated $100,000 to the University of New Mexico in 2001 to support artificial intelligence research. The money went through the university's foundation to help establish an AI research center. That center was also funded by other private and public sources. The university has not revealed the names of the other donors. The center's work focused on developing new AI technologies and studying their effects on society.

Bob Friday explains how agentic AI will create self-driving networks

Bob Friday, the inventor of AI-driven enterprise Wi-Fi and Chief AI Officer at HPE, says agentic AI will make self-driving enterprise networks a reality. He founded Mist Systems in 2014 and built the world's first AI-based enterprise Wi-Fi solution. He outlines five foundations for self-driving networks including data at the edge, Service Level Experiences, and AI-driven support. His vision relies on trusting AI models to reason and troubleshoot problems on their own. Partners like Infoblox and Zscaler contribute data through the Model Context Protocol.

ASUS launches Arc Pro B70, a 32GB Intel AI workstation GPU

ASUS has built its first Intel Arc Pro B-series workstation card called the UGen Arc Pro B70. The card has 32GB of GDDR6 ECC memory, 32 Xe2 cores, and 256 XMX matrix engines. It is designed for local AI inference and professional creation workloads rather than gaming. The card supports AI frameworks including vLLM, llama.cpp, Ollama, PyTorch, and OpenVINO. ASUS has not confirmed a price or shipping date yet.

Maryland colleges let instructors set their own AI rules

Many Maryland colleges and universities are allowing individual instructors to decide how students can use artificial intelligence in their coursework. Unlike K-12 schools that must follow a state mandate by October 22, higher education institutions are setting AI guidance at their own pace. One student at the University of Maryland said policies vary completely from class to class. Experts say AI application looks different for each discipline, so universal rules may be risky. Some professors approve AI use while others make courses harder to prevent it.

AI is shaking the world of literature and raising doubts about writing

The world of literature is facing rapid change as AI tools can now write stories, novels, and poems that closely resemble human writing. Three major crises are shaking the field including automated books competing with authors, plagiarism of writing styles, and a crisis of confidence in literary competitions. It is becoming difficult to verify whether submitted works truly reflect the authors' efforts. Questions about intellectual property rights and quality standards are growing. Many believe AI can still be a valuable tool for writers if used responsibly.

Google tests AI processors in orbit with Project Suncatcher

Google's Project Suncatcher prototype reached orbit on October 1, 2026 and is testing Tensor Processing Units (TPUs) in space. The prototype was built with Planet and launched on SpaceX's Transporter-18 mission. It is a research step to see how AI hardware handles radiation, extreme temperatures and the stresses of spaceflight. The test is not an orbital data center serving customers. A cost model comparing a 1 GW orbital system to a terrestrial one estimates the orbital version at about $51.1 billion over five years versus roughly $16 billion on the ground.

Sources

• 'We will have a cool demo in 2027': What Sam Altman said about OpenAI's humanoid robot and the challenges
• OpenAI is building a humanoid robot, but the body isn't the real challenge
• Singing, Dancing and Getting Dumped by A.I. on Korean TV
• Delay the deployment of artificial intelligence (Voice of the people)
• What a Good “AI” Policy Looks Like
• Epstein gave $100K to UNM for artificial intelligence research in 2001
• Inventor of AI Ops for enterprise Bob Friday explains why agentic AI will finally make self-driving enterprise networks a reality
• ASUS Arc Pro B70: 32GB Intel AI GPU Debuts [2026]
• SUN: As AI use grows, many Maryland colleges are letting instructors set the rules
• Creativity versus artificial intelligence: 3 crises shaking the world of literature and raising doubts about the authenticity of writing
• Orbital data center feasibility: Google’s test