TechCrunch AI
·
20 hours ago
OpenAI released a $230 light-up keyboard called the Codex Micro, designed to control AI coding agents through customizable keys, status lights, and a dial for adjusting computational reasoning levels. The keyboard is a limited-run collaboration with specialty keyboard designer Work Louder and connects to ChatGPT's desktop app. OpenAI is simultaneously developing a separate screenless smart speaker with moving mechanical parts, designed by former Apple engineers, which is still in development and unfinished.
Wired AI
·
22 hours ago
Thinking Machines Lab, founded by OpenAI exiles including former CTO Mira Murati, released Inkling, an open-weight AI model trained on audio, video, and text inputs. Inkling contains 975 billion parameters and performs comparably to leading Chinese open-weight models while being cheaper to run than closed alternatives. The release positions Thinking Machines as a competitor in the AI market and supports its vision for decentralized AI development outside the control of a few dominant companies.
MIT Technology Review AI
·
23 hours ago
OpenAI built GPT-Red, an AI system that automatically finds vulnerabilities in its models by simulating cyberattacks, particularly prompt injection attacks where hackers embed hidden instructions to manipulate LLM behavior. In testing against GPT-5.6, fewer than 23% of GPT-Red's strongest attacks succeeded compared to over 90% against the earlier GPT-5 released in August last year. The system supplements human red-teaming efforts and enables OpenAI to identify new attack types before deployment, though it remains limited in handling multi-turn conversations and image-based attacks.
The New Stack
·
1 day ago
OpenAI launched the Codex Micro, a $230 programmable macropad developed with Work Louder that includes six agent keys that light up to show the status of AI agents. The device features customizable keys, a joystick, dial, and push-to-talk functionality, with keycaps that can be remapped for use with other applications. The product targets Codex users, which OpenAI says will reach 9 million, and allows them to control agentic coding tasks through dedicated hardware.
Ars Technica
·
1 day ago
OpenAI released its first branded hardware product, the $230 Codex Micro, a custom mini-keyboard with color-coded keys for monitoring AI agents. The device features six frosted keys that display status updates—white for idle, blue for processing, green for completion, amber for requiring feedback, and red for errors—across up to six concurrent Codex threads. Users can tap the illuminated keys to quickly switch between different AI agent tasks without manually searching for windows on-screen.
The New Stack
·
1 day ago
OpenAI's Codex and ChatGPT Work combined reached 8 million active users following the July 9 launch of GPT-5.6, growing from 5 million in early June. The rapid user surge from 7 million to 8 million occurred between July 12 and July 15, exposing scaling issues that forced OpenAI to reduce the context window from 372,000 to 272,000 tokens and temporarily lift usage caps. The consolidation of Codex into a unified desktop app with integrated tools like Slack and Google Drive marks a shift toward positioning AI as a workspace layer rather than a standalone product.
OpenAI Blog
·
1 day ago
OpenAI has proposed a "reverse federalism" model where state-level AI regulations would inform and shape federal policy rather than the other way around. The approach suggests allowing states to pioneer AI safety rules that could eventually form the basis of a coherent national standard. This strategy aims to balance localized governance experimentation with the development of consistent safeguards across the country.
TLDR Dev
·
1 day ago
OpenAI's ChatGPT web app serves approximately 1 billion users through a carefully optimized architecture that migrated from Next.js to Remix and now uses React Router 7 with server-side rendering. The initial HTML payload is 84 KB compressed with a time to first byte of 50-65ms, delivering a 548-node DOM that prioritizes the chat input as the primary interaction point. The engineering approach uses off-the-shelf libraries rather than custom frameworks, with the server controlling client behavior through prefetch decisions and streaming data to enable rapid interaction without requiring user authentication.
TLDR
·
1 day ago
Apple sued OpenAI alleging that former employees, including senior hardware engineer Chang Liu and vice president Tang Tan, misappropriated trade secrets and recruited others to do the same. The lawsuit focuses on over 400 Apple employees who have moved to OpenAI and names specific allegations including use of Apple codenames to solicit confidential information and a proprietary "trade secret metal-finishing technique." The case could delay OpenAI's hardware launch timeline and force it to redesign products to avoid Apple-controlled intellectual property, while discovery proceedings may expose details about OpenAI's unreleased hardware plans.
TLDR
·
1 day ago
OpenAI is developing its first consumer device, a mobile smart speaker without a screen designed to function as an AI companion for the home. The device will perform tasks including smart-home control, media playback, and answering questions through voice interaction. Users will be able to carry the speaker around their homes and interact with it as a conversational companion.
OpenAI Blog
·
1 day ago
OpenAI developed GPT-Red, an automated system that uses self-play to identify vulnerabilities in AI models and improve their robustness against adversarial attacks. The system iteratively generates adversarial prompts and defensive improvements in a feedback loop, with evaluations showing measurable increases in resistance to prompt injection attempts. This approach allows AI developers to systematically strengthen defenses without manual red teaming, reducing the time required to identify and patch safety vulnerabilities.
The Neuron
·
1 day ago
OpenAI restored access to ChatGPT on WhatsApp in the European Economic Area after Meta had removed it in January 2026, following an EU Commission order in June 2026 requiring Meta to reinstate competing AI assistants. Users in the 30-country region have been able to reach ChatGPT via the number +1-800-242-8478 since July 13, 2026, with support for text, images, and voice messages across multiple languages. The service allows optional account linking to sync chat history and context, though usage limits for unlinked accounts remain unclear.
The Neuron
·
1 day ago
OpenAI's path to going public is encountering obstacles from multiple lawsuits, strained relations with Microsoft, and competitive pressure from Anthropic's growing valuation. The company's valuation has been set at $80 billion in recent funding rounds, while Anthropic has reached $60 billion, narrowing the gap between the two AI labs. These combined pressures may force OpenAI to reconsider its timeline for an IPO or the terms on which it eventually goes public.
Wired AI
·
1 day ago
OpenAI employees have donated over $215,000 to Guardrails Alliance, a super PAC advocating for stricter AI regulations, in opposition to Leading the Future, a pro-industry PAC backed by $100 million from OpenAI president Greg Brockman. The largest individual donation came from research engineer Juan Felipe Cerón Uribe at $200,000, with Guardrails Alliance aiming to raise $15 million this election cycle compared to the $50 million Brockman and his wife committed to the opposing PAC. The donations highlight internal company tensions over AI policy direction, with some employees using personal funds to counter what they view as efforts by tech leaders to prevent AI regulation.
TechCrunch AI
·
1 day ago
Miles Wang, an OpenAI researcher specializing in AI for scientific discovery, is leaving to start a drug discovery startup with several other OpenAI researchers. The new company is in funding discussions with Lightspeed to raise $200 million at a $2 billion valuation. The startup would focus on developing AI models to identify new uses for existing drugs, potentially reducing time to market compared to discovering entirely new treatments.
