OpenAI Hardware Debut, New Models from Thinking Machines, and Industry Legal Updates
OpenAI unveils its first hardware products, including a smart speaker and a coding keyboard. Thinking Machines releases its first open-weight model, Inkling, while Weco demonstrates recursive self-improvement in AI agents. Legal and business developments include updates on the Apple-OpenAI lawsuit, a new enterprise AI firm from Anthropic, and a data-related lawsuit against Meta.
Podcast В· 3 min
OpenAI's First Hardware: A Screenless Smart Speaker
OpenAI is reportedly developing its first consumer hardware device: a portable, screenless smart speaker. The device is designed to function as a physical companion for ChatGPT, featuring cameras, sensors, and the GPT-Live voice interface. It is intended to be movable around the home, capable of understanding context and managing smart-home tasks. While the product is expected to arrive around 2027, it signals a strategic shift for OpenAI toward physical, voice-first interfaces rather than screen-based software.
OpenAI Launches 'Codex Micro' Keyboard
OpenAI has released 'Codex Micro,' a $230 mechanical keyboard developed in collaboration with Work Louder. The device is specifically designed for steering AI coding agents, featuring status lights, command keys, a joystick, and a reasoning-effort dial. The keyboard aims to move AI controls from software menus to physical hardware, allowing developers to manage coding agents more intuitively.
Thinking Machines Releases 'Inkling' Model
Thinking Machines Lab has introduced 'Inkling,' its first open-weight, multimodal model. The model is built with a focus on customization and controllable reasoning effort, allowing users to trade thinking depth for cost. Developers can access the model via Hugging Face or fine-tune it through the company's 'Tinker' service. Inkling is positioned as a foundational tool for agentic web design and instruction following.
Weco Demonstrates Recursive Self-Improvement
AI research team Weco has reported evidence of recursive self-improvement using its automated agent, AIDE^2. Over an eight-day period, the agent redesigned its own research process, testing 100 rewrites and implementing seven upgrades. The resulting versions outperformed a hand-built agent on benchmarks ranging from physics-based weather forecasting to engineering, suggesting a potential breakthrough in autonomous system evolution.
OpenAI Publishes GPT-Red Research
OpenAI has released research on 'GPT-Red,' an automated red-teaming system. The tool utilizes self-play to identify vulnerabilities and harden frontier models against prompt injections. This system was reportedly used to train GPT-5.6, highlighting the company's focus on automated safety testing.
Update: Apple vs. OpenAI Lawsuit
The ongoing legal dispute between Apple and OpenAI has seen a minor development. Reports indicate that Apple's legal team mistakenly identified two OpenAI employees while pressing the case regarding alleged trade secret theft. While this error does not invalidate the core allegations, it suggests potential gaps in Apple's initial evidence gathering. The case remains focused on the alleged poaching of employees and the unauthorized use of hardware secrets.
Anthropic Launches Enterprise AI Firm 'Ode'
Anthropic has partnered with Blackstone and Hellman & Friedman to launch 'Ode,' a new enterprise AI services firm. The company aims to embed AI engineers directly into enterprises to build and implement systems based on Claude. This move represents a significant expansion of Anthropic's enterprise strategy.
Meta Faces Lawsuit Over Layoffs
Meta is facing a lawsuit from 26 employees alleging that AI-driven decisions skewed recent layoffs toward staff on medical leave. Despite Meta's assertion that layoff decisions were made by people rather than AI, the plaintiffs argue that the process was biased. This case highlights growing concerns regarding the use of AI in HR and workforce management.
Walden Robotics Emerges from Stealth
Walden Robotics has launched with a $1.1 billion valuation and approximately $300 million in funding. Backed by Toyota, the startup is focused on developing industrial humanoid robots. The company's emergence underscores the continued high-value investment in physical AI and robotics.