OpenAI cuts off Cursor after SpaceX acquisition

Hi there, this is your daily ☕️ Techpresso.

OpenAI cuts off Cursor after SpaceX acquisition LINK

• OpenAI said late Friday it will end its partnership with Cursor and cut the coding tool's access to its models on November 12, roughly three months after SpaceX bought Cursor's parent company Anysphere.
• OpenAI blamed the split on trust rather than technology, writing that it cannot be sure SpaceX will follow its terms of service given past contract violations by X and other Musk-owned companies.
• The cutoff pushes Cursor to rely more on Anthropic's Claude and its own in-house Composer model, leaving open the question of whether Anthropic will make the same move to drop the tool.

🇰🇷 South Korea gives every citizen free AI LINK

• South Korea has picked three groups led by SK Telecom, KT and Kakao to build free, unlimited basic AI services for every citizen, with a full launch planned by the end of 2026.
• The government will hand the consortia a combined 512 Nvidia B200 chips during 2026, and starting in 2027 will cover part of the running costs while the companies must add their own money and find revenue.
• Rules set in July require at least half of each service's model use to come from independent Korean foundation models, with 30% from other domestic developers, though the total cost and real-world capacity remain unsettled.

Meta tests robots for data center jobs LINK

• Meta is trying out robots to help run the data centers behind its AI, with machines that can swap networking cables, reboot servers, move racks, and check equipment, according to a WIRED report.
• The company is testing hardware from Watney Robotics, Kinova, and ABB, but the machines still need people watching over them and struggle with tangled cabling, battery life, obstacles, and visual inspections.
• One worker estimated a working cable-swapping robot could take over up to 80% of some jobs, though that was the employee's guess, not a Meta forecast, and the robots remain slower than humans.

States can regulate prediction markets LINK

• A federal appeals court ruled Friday that states can treat prediction markets like Kalshi as gambling, a setback for an industry that has argued its bets are federally regulated financial contracts rather than wagers.
• The 3-0 decision from the Ninth Circuit, involving Nevada's push to shut down Kalshi, now binds federal judges in California, Arizona and six other states, opening the door for them to regulate or close the site.
• The ruling clashes with a Third Circuit decision this year that sided with prediction sites against New Jersey, creating a split that experts say likely sends the fight over sports event contracts to the Supreme Court.

X finds 200000 fake energy accounts LINK

• X says it found a network of 200,000 fake accounts tied to China, with 200 of them posting cartoons and text designed to sway American debate over AI data centers and energy policy.
• The 200 accounts claimed data centers push up household electricity prices and strain local grids, sharing AI-made cartoons that showed operators as greedy businessmen while ordinary families paid the bill.
• X said the accounts used OpenAI tools to create the images and talking points, matching a June OpenAI report about China-linked ChatGPT users generating anti-data center content and other political material.

Other news you might like

• New US export controls reportedly target Chinese access to remote AI servers LINK

• Google Scraps Search Ranking Penalties for Publishers After EU Pressure LINK

• Brave launches Email Aliases to hide users’ real email addresses LINK

• AI-generated videos are already displacing actors and livestreamers across China's entertainment industry LINK

• SpaceX recovers Starship from Indian Ocean a month after launch, shipping it to Texas LINK

• Xbox Project Helix is a 'family of devices,' CEO reveals LINK

• Microsoft-backed AI data center has been accused of violating federal law LINK

• An Anthropic researcher just gave us a peek at self-improving AI LINK

• Chinese automakers are following Tesla’s bet that robots are the next big profit machine LINK

• Tencent released open-source Hy4 preview model LINK

• Early leak of NVIDIA's DLSS 5 has an uncanny valley problem LINK

Trending tools

1752vc Pitch Deck Analyzer: analyzes startup pitch decks slide-by-slide against 25,000+ real decks, flagging weak narratives and contradictory claims investors will catch before you send. LINK

Hy4 preview: Tencent's multimodal AI model family handling text, image, video, and 3D generation, letting developers build content tools and multimodal apps. LINK

Cohere Parse 5: builds customizable, high-performance language models for enterprises, deployable across public, private, or hybrid clouds to keep data secure. LINK

Neo: a distraction-free writing app that formats your manuscript to look like a book while you draft, managing multiple works in one place. LINK

Cursor Craft v2: browse, install, and switch custom Mac cursor sets in one click, with free and premium packs across styles like minimal, neon, and anime. LINK

Staats: cookieless, agent-native site analytics where your coding agent tracks deploys, flags changes, and suggests next steps through chat instead of dashboards. LINK

Trending papers & reports

Factory-controller attacks get their first real-hardware test, showing that autonomous AI agents turned loose on industrial controllers sustained real physical damage in ~31% of 240 attempts, proving the threat is concrete not theoretical. LINK

Hidden agent skills can be stolen just by submitting normal tasks and studying the outputs, letting rivals rebuild ~87% of a paid specialized service's capability in about 32 requests, even with disclosure protections on. LINK

Autonomous agent safety shows that current guardrails, which reset with each task, are blind to attacks spread across many steps, and offers a persistent monitor that caps unauthorized irreversible actions at a fixed limit. LINK

Automated attack agent stress-tests AI assistants that take real actions like using tools, learning reusable attack tricks that break them more reliably than fixed methods, so companies can find dangerous vulnerabilities before deployment. LINK

AI model security scanners vary wildly in how often they actually deliver a verdict, with one tool judging all 135 test files while another judged only half, showing accuracy scores hide big coverage gaps. LINK

See you tomorrow for a new dose of ☕️ Techpresso!