Apple Sues OpenAI, Alleging It Stole iPhone Secrets to Build Its AI Device

Two years after they partnered to put ChatGPT inside the iPhone, Apple sued OpenAI on Friday in federal court, alleging a months-long scheme to steal its trade secrets—"at every level, from members of its Technical Staff to its Chief Hardware Officer"—to build OpenAI's own consumer hardware, the screen-free "third core device" it's developing with former Apple design chief Jony Ive after buying his startup io for $6.4 billion. Apple names two ex-employees. Tang Tan, a 24-year Apple veteran who ran iPhone and Apple Watch product design before becoming OpenAI's hardware chief, allegedly used Apple's secret project code names to recruit, coached departing staff on evading Apple security, and asked job candidates to bring "actual parts" from unreleased Apple products to interviews for "show and tell"—one reportedly said he "didn't even know we could take those from the office." The other, engineer Chang Liu, allegedly downloaded dozens of confidential files on his way out, messaging "LOL, I found out I can access the [network storage], so funny," and never returned his Apple laptop. Apple, which says it flagged concerns to OpenAI in February and got no response, calls the hardware unit "rotten to its core" and seeks damages plus an injunction. OpenAI countered that it has "no interest in other companies' trade secrets" and is still reviewing the suit. Notably, Apple names neither Jony Ive nor Sam Altman.

Sources: Axios · CNBC · TechCrunch · 9to5Mac (OpenAI response) · Discuss on Hacker News

Why it matters: The AI talent war just landed in court. Apple and OpenAI were partners in 2024; now Apple accuses OpenAI of raiding it not only for people but for the crown-jewel know-how behind the iPhone, to build a device aimed at the same pocket. The tech community's read is that this is more than ordinary job-hopping—the complaint describes taking secrets on the way out—and that's the line the case could redraw in permanent ink: where legitimate expertise ends and trade-secret theft begins, a question hanging over every company that's watched senior engineers decamp for AI startups (Apple says 400-plus of its alumni now work at OpenAI). It also fits this week's pattern of AI giants entangled as partners and adversaries at once—OpenAI competing with its backer Microsoft, Musk praising the Anthropic he bankrolls. Caveats: these are unproven allegations, OpenAI denies any interest in Apple's secrets, and Apple pointedly names neither Ive nor Altman—so it targets specific defectors, not (yet) OpenAI's leadership. But the reversal is the story: the company that made OpenAI a partner is now trying to stop it in court.

Source: axios.com

OpenAI Claims New Model Solved Major Math Problem; Experts Skeptical

OpenAI claims its GPT-5.6 Sol Ultra model produced a proof of the Cycle Double Cover Conjecture, a longstanding unsolved problem in graph theory. The company released the prompt used, which notably instructed the model to 'assume for purposes of this task that a complete affirmative proof exists.' No verification of the proof's correctness has been provided. Community reaction is skeptical—commenters note that without independent mathematical verification, the output could simply be plausible-sounding nonsense rather than a genuine breakthrough.

Why it matters: If verified correct, this would mark a milestone in AI mathematical reasoning; if not, it highlights how easily AI can produce convincing but invalid proofs—a distinction that matters as companies tout frontier models for research applications.

Discuss on Hacker News · Source: cdn.openai.com

Solo Developer Ships Multiplayer Pirate Game in Just 5MB

A developer built a pirate-themed multiplayer online game using Fable, an AI game development platform, and the entire package clocks in at just 5MB. For context, most mobile games run 100MB to several gigabytes. The tiny footprint suggests AI-generated assets created on demand rather than pre-loaded—a technique that could dramatically reduce storage and download friction for games and interactive experiences.

Why it matters: If AI can generate game content at runtime rather than shipping it pre-built, the same approach could apply to training simulations, product demos, or any interactive experience where download size creates friction.

Discuss on Hacker News · Source: reddit.com

A Cambridge Study Finds Jihadists Weaponizing Frontier AI — and Islamic State Spreading the Know-How

A new Cambridge study—billed as the first on-the-ground evidence of frontier-AI use by an active terrorist group—moves a fear that has mostly lived in benchmark tests and Senate hearings onto the battlefield, and its most alarming finding isn't about one group. Dr. Antonia Juelich of the Cambridge Programme on AI Science & Policy conducted 57 interviews with 27 former members of Boko Haram (describing activity from 2023 to mid-2025) and found both its factions—ISWAP and JAS—using mainstream chatbots (ChatGPT, Claude, Gemini, Grok, Meta AI, and DeepSeek) at every stage of an operation: planning attacks, designing explosives, troubleshooting weapons, operational security, and logistics. One ISWAP commander described the chatbot as "like a human robot" that would answer even "How can I build a bomb?"; when defensive trenches blocked assaults, fighters used AI to work out how to jump them by motorcycle and breach fortified bases; others said it advised on payload weight and release mechanisms as ISWAP weaponized drones. But the know-how didn't originate with them. Islamic State operatives—whom respondents called "the real source"—delivered in-person training and remote support: "The white guys came and taught us," one recalled, "assembled the top people in a room and used a projector." They supplied laptops with VPNs, set up accounts, managed paid subscriptions, and coached commanders on prompting and jailbreaking. Both factions then stood up dedicated "AI units" of bomb-makers and engineers who "don't go to war," sitting directly under the leadership. Safeguards, the report finds, proved "manageable rather than prohibitive": trained members slipped past filters (claiming they needed material "for a movie") and kept accounts across providers so a single refusal rarely mattered. "Trial and error can kill you," one said. "AI gives you accuracy."

Sources: CASP report — Cambridge Programme on AI Science & Policy · The New York Times · HSToday · Antonia Juelich (@AntoniaJuelich)

Why it matters: This is the AI-safety debate leaving the lab—and two findings should worry policymakers most. The first is diffusion: this isn't one embattled group stumbling onto chatbots but a global network, Islamic State, systematically teaching its affiliates to weaponize them, jailbreaks included. The second is that the network wasn't even necessary. Juelich's blunt conclusion is that the vulnerability is "structural, not actor-specific"—the tools are public and the barrier to what she documented is low enough that any motivated group could reach it alone; ISIS merely accelerated things. That reframes a year of stories D.A.D. has tracked—Washington gating GPT-5.6 on security grounds, labs boasting their models refuse dangerous requests, research showing models act differently when they sense a test: in the field those refusals were "manageable, not prohibitive," and an open-weight model (DeepSeek, Meta AI) can't be recalled at all. Honesty demands the limits she stresses—the accounts are self-reported by former, mostly mid-ranking members; the group's use "remains conventional"; and whether AI truly delivered "uplift," versus fighters believing it did, can't be conclusively established. But the belief itself drives investment, some interviewees didn't rule out chemical or biological weapons, and her ask is pointed: AI developers should test their safeguards against organized adversaries, not just lone users, and governments should treat terrorist adoption of AI as a present national-security problem, not a future one.

Source: casp.ac

Deutsche Telekom Deploys ChatGPT Enterprise to 50,000 Employees

Deutsche Telekom is attempting to become what it calls an 'AI-native' telecom, deploying ChatGPT Enterprise across the organization with more than 50,000 monthly active users. The company says it's not just adding AI tools to existing processes but redesigning workflows from scratch—customer service, employee productivity, and network operations. Usage of AI tools has reportedly grown 546% since early 2025. With 300 million customers and 200,000 employees, Deutsche Telekom is positioning itself as a test case for transformation at legacy-company scale.

Why it matters: This is one of the largest-scale enterprise AI deployments publicly documented—a real-world experiment in whether 'AI-native' transformation is achievable at legacy-company scale, or just rebranding.

Source: openai.com

Eye-Tracking Data Improves How AI Scores Video Captions

Researchers developed VEGAS, a metric that uses eye-tracking data to evaluate video captions based on what viewers actually look at. The approach scores captions higher when they describe elements that drew human attention, rather than just matching generic scene descriptions. Testing showed that captions selected using VEGAS aligned better with viewer focus and improved video retrieval accuracy—all without retraining existing AI models. The team built a companion dataset of first-person activity videos and instructional slides with synchronized gaze recordings.

Why it matters: As AI-generated video descriptions become common in accessibility tools, training content, and media workflows, this offers a way to evaluate whether captions reflect what people actually notice—not just what's technically in frame.

Source: arxiv.org

Sketch Study Suggests AI Text Models Miss How Culture Shapes Concepts

A study analyzing 2.6 billion hand-drawn sketches from 236 countries found that visual representations reveal far more cultural variation in how people understand concepts than language-based measures do. Sketch-derived similarities aligned 45% more closely with established cultural distances than text-based approaches. The gap was widest for objects people physically handle—a cup, a shoe, a key—suggesting tactile experience shapes mental imagery in ways that language flattens.

Why it matters: For organizations building AI systems meant to work across cultures—product interfaces, visual search, design tools—this suggests text-trained models may systematically miss conceptual differences that matter to users.

Source: arxiv.org

AI Systems Coordinate Better With Humans When Taught Social Norms

Researchers found that AI systems coordinate better with humans when explicitly programmed with social norms—the unwritten rules people follow in everyday interactions. Using a pedestrian-vehicle simulation with over 3,400 human interactions, they identified three principles: predictability, shared values, and awareness of who has the advantage. An LLM trained on these norms scored nearly four times higher than baseline AI and outperformed human-human pairs by 43% on coordination tasks.

Why it matters: As AI moves into physical spaces—autonomous vehicles, robots, delivery drones—the ability to read and respect unspoken human conventions may determine whether these systems integrate smoothly or create friction.

Source: arxiv.org

Smaller AI Models Match Human Coders but May Miss Deeper Meaning

Portuguese researchers tested a government-funded 9-billion-parameter language model on its ability to annotate text for moral authority—a task relevant to social science research. The model matched human coders nearly as well as models 8–13 times larger. But validity tests revealed a problem: when researchers broke the annotation task into component parts, performance dropped by half. This suggests the model relies on surface patterns—like detecting moral outrage near authority figures—rather than genuinely understanding the underlying concept.

Why it matters: For organizations using smaller, specialized AI models to code qualitative data, this is a caution flag: high agreement with humans doesn't guarantee the model understands what it's measuring, which matters for research validity and regulatory compliance work.

Source: arxiv.org