AI News Daily - 2026-10-31
2026-10-31
Today’s Highlights
- North Korean hackers stole a record $6.3 billion from Binance in a swift 20‑minute cyberattack, marking the largest crypto heist in history.
- Cracks are appearing in the AI spending boom as companies reassess investments and shift toward more practical applications.
- Google’s first data‑center satellite went online, successfully transmitting its first AI‑generated response to Earth.
- Concerns grow over GLP‑1 weight‑loss drugs potentially worsening eating disorders, prompting calls for better screening.
- A supply‑chain attack placed malicious Claude extensions on thousands of devices, exposing vulnerabilities in AI ecosystems.
TechCrunch
Last-minute Halloween costumes for tech nerds
A roundup of last‑minute Halloween costume ideas tailored for tech enthusiasts, featuring DIY options like dressing as a living keyboard, a bug in your code, or an AI chatbot. The article suggests quick fixes using household items and includes tips for creating recognizable tech‑themed outfits without spending much time or money.
Cracks are emerging in the AI spending boom
Despite the massive surge in AI‑related spending over the past few years, emerging data suggests that some companies are pulling back or reallocating budgets, indicating that the AI investment boom may be hitting its first signs of fatigue. Analysts point to tightening capital markets, lower returns on certain AI projects, and shifting priorities toward more practical applications as key factors.
AI law firm Crosby bids on disintegrating, famed firm Weil Gotshal
Crosby, an AI‑focused law firm, has submitted a bid to acquire Weil Gotshal & Manges, a historic but financially strained partnership that has been reeling from talent departures and client losses. The move marks one of the first major attempts by an AI‑native firm to absorb a legacy practice, though the deal faces uncertainty as the target firm continues to fragment.
Satlyt takes another step toward turning satellites into space data centers
Satlyt announced progress in its vision of transforming retired satellites into orbital data‑processing nodes, aiming to reduce latency and bandwidth costs for space‑based AI workloads. The company demonstrated a prototype that repurposes decommissioned communication satellites with added compute modules, bringing the concept of “space data centers” closer to commercial reality.
Zest’s new AI agent finds restaurants and books reservations over text
Zest launched a conversational AI agent that uses text messages to locate open tables at restaurants and complete reservations on behalf of users. The agent integrates with major booking platforms and learns user preferences to suggest venues, aiming to eliminate the friction of scrolling through apps and making multiple calls to secure a dinner spot.
Google’s first data center satellite is online and writing poetry
Google’s inaugural data‑center satellite, part of its Project Suncatcher initiative, is now operational in orbit and capable of running AI models. The satellite’s first task was a creative test: generating original poetry, which Google released as a proof‑of‑concept for its ability to perform lightweight AI inference in space without relying on Earth‑bound servers.
Theranos whistleblower Tyler Shultz warns that the AI boom is ‘a great opportunity to go deceive people’
Tyler Shultz, known for exposing the Theranos blood‑testing scandal, cautioned that the rapid AI boom creates fertile ground for deception, with bad actors able to fabricate demos, fake metrics, and misleading results. He urged regulators, investors, and the public to apply rigorous scrutiny to AI claims, comparing the current climate to the pre‑Theranos era of unchecked health‑tech hype.
North Korean hackers blamed for stealing record $6.3B in Binance breach
North Korean state‑linked hackers are believed to have executed a massive heist on cryptocurrency exchange Binance, stealing approximately $6.3 billion in digital assets—a record for a single crypto attack. Security firms traced the breach to known DPRK hacking groups, highlighting the growing sophistication of nation‑state actors targeting financial platforms.
Skip the small talk: Make game-changing connections at TechCrunch Founder Summit with Braindate
TechCrunch is partnering with AI‑powered networking platform Braindate to help attendees at its Founder Summit forge meaningful connections without the usual small talk. Braindate uses matchmaking algorithms to pair founders, investors, and experts based on shared interests and goals, streaming the networking process and increasing the likelihood of high‑value conversations.
DoorDash said it’s only received ‘several hundred’ orders from AI agents
DoorDash reported that AI agents have placed only a few hundred orders through its platform so far, suggesting that while the technology exists, consumer adoption of autonomous food‑delivery bots remains in its infancy. The low volume underscores the practical hurdles—reliability, payment security, and user trust—that still limit AI agents from handling routine purchases at scale.
The Verge
Exclusive details about Akai’s MPC Sample Key
Akai has unveiled the MPC Sample Key, a compact hardware controller designed for sampling and beat‑making. The device features a velocity‑sensitive keyboard, pad controls, and direct USB connectivity to DAWs, targeting producers who want a portable, tactile alternative to software‑only workflows. Exclusive specs reveal expanded memory and improved latency.
Sam Altman says it’s better for RAM to be in a data center than on your desk
OpenAI CEO Sam Altman argued that RAM resources should be centralized in data centers rather than distributed across personal devices, citing efficiency, scalability, and the need to serve massive AI models. The remark reinforces the industry’s shift toward cloud‑first AI, where compute and memory are provisioned on‑demand instead of being embedded in end‑user hardware.
The biggest names in AI are backing a new way to opt out of scraping
OpenAI, Google, Microsoft, and several other AI leaders have endorsed a new technical standard that allows website owners to explicitly opt out of having their content scraped for training. The “Opt‑Out for AI” protocol sends a signal via HTTP headers, giving publishers clearer control over how their data is used by automated systems.
Google’s satellite AI server transmits its first response to Earth
Google’s orbital AI server successfully sent its first computed response back to ground stations, marking a milestone for space‑based machine learning. The satellite, equipped with specialized processors, performed a test inference task and relayed the result through a dedicated downlink, demonstrating that AI workloads can be executed off‑Earth and still communicate with users.
Starbucks wants to use AI to show you a ‘personalized digital menu’ at the drive-thru
Starbucks is piloting an AI‑driven digital menu system for its drive‑thru lanes that adjusts suggestions based on customer history, time of day, weather, and even local events. The goal is to increase average order value by presenting personalized drink and food recommendations in real time, reducing decision fatigue and speeding up service.
The delightfully messy state of Apple gadgets
A recent episode of The Verge’s podcast examines the current ecosystem of Apple devices, noting both the seamless integration that fans love and the occasional software glitches, feature fragmentation, and pricing oddities that make the lineup feel delightfully messy. The hosts discuss how Apple balances consistency with incremental innovation across iPhones, Macs, Watches, and Services.
Rivian’s autonomy gamble faces its first real test
Rivian’s ambitious self‑driving truck project is undergoing its first serious field tests, as the electric‑vehicle maker races to prove that its Level 4 autonomy system can operate reliably on public roads. The trial focuses on long‑haul freight routes, with safety drivers monitoring the system’s ability to handle complex highway scenarios without human intervention.
Vizio’s discounted Mini LED TV lets you skip the sign-ups and smarts
Vizio is offering a 65‑inch Mini LED television at a reduced price, emphasizing a no‑frills approach that omits costly smart‑TV subscriptions and streaming‑service integrations. The panel delivers strong brightness and contrast, appealing to viewers who prefer to use their own external streaming devices and avoid recurring monthly fees.
Google’s latest Nest Learning Thermostat is $60 off just in time for sweater weather
Google has dropped the price of its fourth‑generation Nest Learning Thermostat by $60, coinciding with the onset of cooler autumn weather. The updated thermostat features improved sensors, faster learning, and better energy‑usage insights, making the discount an attractive opportunity for homeowners looking to upgrade their HVAC control before winter.
Pope Leo calls for the disarmament of AI: ‘Machines are tools, not persons’
Pope Leo issued a public appeal for the global community to treat artificial intelligence as a tool rather than a person, calling for ethical guidelines and safeguards to prevent the misuse of autonomous systems. The statement, delivered during a Vatican address, urged policymakers to focus on human‑centered AI development and to resist anthropomorphizing machines.
Ars Technica
Dartmouth’s provost warned us about AI cheating—but his own AI use cost him his job
Dartmouth University’s former provost, who had publicly cautioned faculty and students about the dangers of AI‑assisted cheating, lost his position after it was revealed that he had used an AI tool to draft a controversial administrative email. The incident has sparked debate over double standards in AI policy and the difficulty of policing use among those who set the rules.
Over 25% of people lie to get antibiotics—and their lies aren’t even good
A new survey found that more than a quarter of respondents admitted to fabricating symptoms or exaggerating illnesses to obtain antibiotics, often providing poorly constructed excuses that doctors could easily spot. Researchers warn that this pattern of deceptive behavior contributes to antibiotic resistance and undermines efforts to promote judicious prescribing.
Basically no one is using AI agents for their DoorDash orders
Reiterating DoorDash’s disclosure that AI agents have generated only a handful of food‑delivery orders, the article notes that consumer adoption of fully autonomous purchasing bots remains negligible. Industry observers suggest that trust issues, payment complexity, and the novelty factor are keeping volumes low.
Rare meteorite from Mars gives us a view into its ancient crust
Scientists have analyzed a rare meteorite believed to have originated from Mars, offering unprecedented insights into the planet’s ancient crust and early geological activity. The sample contains minerals that formed under conditions inconsistent with Mars’ current dry surface, suggesting a wetter, more dynamic past that could inform future searches for microbial life.
Fujifilm to boost tape production 5x as AI fuels archiving demand
Fujifilm announced a fivefold increase in the production of magnetic tape storage media, driven by surging demand from data‑center operators and AI developers who need cold‑storage solutions for massive datasets. The company states that tape remains the most cost‑effective and energy‑efficient medium for long‑term archival, despite the popularity of cloud‑based storage.
Thousands of “ghost students” use AI to steal millions in student loans, DOJ says
The Department of Justice has exposed a scheme in which thousands of fraudulent “ghost students”—appearing to be real learners—used AI to complete online courses and siphon federal student loans. The AI‑generated identities and coursework have cost taxpayers millions and prompted calls for stricter verification measures in distance‑education programs.
Analysts find looming launch crisis is “more severe” than anticipated
Industry analysts warn that a projected shortage of space launch vehicles for the coming decade is deeper than previously thought, threatening the deployment of satellites, lunar missions, and other critical payloads. The deficit stems from bottlenecks in rocket manufacturing, regulatory delays, and an overreliance on a limited number of providers.
GM to pay $220 million after selling customer Smart Driver data
General Motors agreed to pay $220 million to settle a lawsuit accusing the automaker of illegally selling anonymized driving data from its Smart Driver app to third‑party brokers. The settlement highlights growing concerns about vehicle‑generated data privacy and the lack of explicit consumer consent for commercial data transactions.
Hackers steal $6.3 billion from Binance in record crypto heist
A team of North Korean hackers executed a sophisticated cyberattack on cryptocurrency exchange Binance, absconding with $6.3 billion in digital assets—the largest theft in crypto history. Security researchers traced the breach to known DPRK affiliate groups, underscoring the escalation of nation‑state‑sponsored financial crime.
2027 Nissan Rogue Hybrid: Drives like an EV, but there’s no plug involved
The 2027 Nissan Rogue Hybrid delivers a smooth, quiet ride that mimics an electric vehicle while relying on a conventional gasoline‑electric hybrid powertrain. Early tests praise its efficient city driving, capable highway cruising, and competitive fuel economy, positioning it as a practical alternative for buyers who want EV‑like refinement without range anxiety.
Product Hunt
StereoLift Live
StereoLift Live is a real‑time audio processing application that enhances stereo imaging and spatial depth for music production and live sound reinforcement. The tool offers intuitive controls for widening, depth, and center‑image stabilization, catering to producers and engineers seeking a polished stereo field.
iSurge
iSurge appears to be a mobile app focused on surge‑capacity management for healthcare or event staffing, using predictive analytics to forecast demand spikes and automate scheduling. The platform aims to reduce overtime costs and improve resource allocation in time‑sensitive industries.
Teable Mobile
Teable Mobile is a lightweight, on‑the‑go database client that syncs with the Teable web platform, allowing users to view, edit, and collaborate on spreadsheets and relational data from iOS and Android devices. The app emphasizes offline access and real‑time conflict resolution.
Knocket AI
Knocket AI is an artificial‑intelligence assistant that joins video meetings, records transcripts, extracts action items, and sends follow‑up summaries automatically. The tool integrates with popular conferencing platforms and aims to eliminate manual note‑taking for teams.
Vaani Developer Platform
Vaani Developer Platform provides a suite of APIs and SDKs for building voice‑enabled applications, with features like natural‑language understanding, text‑to‑speech, and multi‑language support. The platform targets indie developers and startups looking to add conversational interfaces without building AI models from scratch.
SparkBots AI (Powered by Jev)
SparkBots AI, built on the Jev engine, enables users to create custom chatbots for websites, customer support, and internal knowledge bases. The service offers drag‑and‑drop training, sentiment analysis, and seamless integration with CRM systems, lowering the barrier to AI‑powered automation.
Cluinn
Cluinn is a data‑visualization and clustering tool that helps analysts group large datasets into meaningful segments without extensive coding. The platform uses machine‑learning algorithms to auto‑detect patterns and present results in interactive dashboards.
SpeakON Agent
SpeakON Agent is a voice‑based AI assistant designed for phone‑based scheduling and reservation services. It can handle inbound calls, verify identities, and book appointments across multiple calendars, targeting small businesses that need an affordable alternative to human receptionists.
Canva Brush
Canva Brush is an AI‑powered design element that generates custom graphics, icons, and illustrations from text prompts within the Canva editor. The feature expands Canva’s template library by allowing users to create unique visual assets on the fly, streamlining the design process for non‑designers.
MCP Server by XWiki
MCP Server by XWiki is a lightweight server component that extends XWiki’s wiki platform with multi‑channel publishing and content‑distribution capabilities. It enables wikis to push updates to RSS feeds, email newsletters, and third‑party aggregators, improving content discoverability.
GitHub Trending
thedotmack / claude-mem
claude-mem is a Python library that provides persistent memory storage for Claude AI sessions, allowing users to retain context across multiple interactions. The package implements a simple key‑value store that can be integrated with Claude’s API to avoid re‑entering repetitive information.
milind-soni / OpenMausBot
OpenMausBot is an open‑source hardware project that builds a desktop robot controlled by an Arduino and servo motors, mimicking a mouse movement to automate repetitive computer tasks. The design is documented for makers who want to experiment with robotic automation on a budget.
pbakaus / impeccable
impeccable is a lightweight CSS framework focused on accessibility and flawless rendering across browsers. It provides a set of utility classes and components that adhere to WCAG guidelines, aiming to reduce the effort required to build inclusive web interfaces.
CopilotKit / OpenBot
OpenBot is a modular, open‑source chatbot kit that lets developers assemble conversational AI agents using plug‑and‑play components for NLU, dialogue management, and integration with third‑party services. The project lowers the barrier to building custom bots for support, sales, or education.
Renhuai123 / ziwei-doushu
ziwei-doushu is a software implementation of the Chinese Zi Wei Dou Shu (Purple Star Astrology) system, calculating horoscopes and destiny charts based on birth date and time. The project offers both a command‑line interface and a web frontend for enthusiasts of traditional Chinese metaphysics.
expo / skills
expo/skills is an extension library for the Expo framework that adds reusable “skills” (pre‑built modules) for common mobile‑app functionalities such as biometric authentication, push notifications, and in‑app purchases. The package simplifies cross‑platform development by providing tested, platform‑specific implementations.
macro-inc / macro
macro is a desktop application that records and replays sequences of mouse clicks, keystrokes, and API calls, enabling users to automate repetitive tasks without scripting. The tool includes a visual editor and supports export to scripts for advanced customization.
mattpocock / skills
mattpocock/skills appears to be a TypeScript library that provides a collection of composable skills for building AI‑augmented applications, with a focus on type‑safety and integration with popular LLM frameworks. The author emphasizes a modular architecture that allows developers to mix and match capabilities.
pingdotgg / t3code
t3code is a CLI tool that scaffolds full‑stack web projects using the T3 Stack (TypeScript, Tailwind CSS, tRPC, Next.js) with pre‑configured boilerplate for authentication, database, and deployment. It aims to accelerate development by generating production‑ready codebases from a single command.
HKUDS / DeepTutor
DeepTutor is an educational AI system developed by HKUDS that adapts to individual student learning patterns, providing personalized explanations and practice problems in real time. The platform uses reinforcement learning to adjust difficulty and pacing, targeting K‑12 and higher‑education settings.
MIT Technology Review
The Download: GLP-1s and eating disorders, self-aware cities, and forever chemicals
This week’s Download episode explores the link between GLP‑1 weight‑loss drugs and rising eating‑disorder rates, the emergence of “self‑aware” cities that use sensor networks to optimize resources, and the persistent presence of forever chemicals (PFAS) in the environment. Host Thomas Macaulay breaks down the latest research and policy debates.
Eating disorders are on the rise. GLP-1s could make things worse.
A new analysis warns that the widespread use of GLP‑1 agonists like Ozempic and Wegovy may be exacerbating eating‑disorder symptoms, particularly among young women. The drugs’ appetite‑suppressing effects can trigger restrictive behaviors, and clinicians are calling for better screening and mental‑health support for patients.
The Download: AI agent tracking and climate feedback loops
The Download examines the growing need to track AI agents as they act autonomously in critical infrastructure, and the discovery of dangerous climate feedback loops that could accelerate warming. The episode features interviews with experts on monitoring AI systems and modeling Earth’s changing climate.
We urgently need to monitor AI agents’ actions in real time
As AI agents begin to operate independently in sectors like healthcare, finance, and transportation, experts argue that real‑time monitoring systems are essential to detect harmful decisions before they cause damage. The article proposes a framework for logging agent actions, setting intervention thresholds, and ensuring human oversight.
Our climate models have a big, dangerous blind spot
Researchers have identified a significant gap in current climate‑modeling efforts: the underrepresentation of cloud‑aerosol interactions and their impact on precipitation patterns. This blind spot could lead to underestimates of regional climate risks, urging scientists to incorporate higher‑resolution atmospheric simulations.
Creating AI value with enterprise context
This MIT Sloan Management Review insight piece argues that enterprises can unlock greater value from AI by embedding domain‑specific context into models rather than relying on generic foundational AI. Case studies show that contextualized AI reduces errors, improves decision‑making, and accelerates ROI in industries like manufacturing and logistics.
The Download: dead startups fueling AI and interspecies communication
The Download delves into how the collapse of certain AI startups is affecting the broader ecosystem, and highlights breakthroughs in interspecies communication research that use AI to decode animal vocalizations. The episode also touches on regulatory shifts and the future of AI safety.
Pairing industrial AI autonomy and human control
A new study explores optimal ways to balance AI autonomy with human oversight in industrial settings, such as factories and power grids. The findings suggest that “guided autonomy,” where humans can intervene at critical junctures, yields better safety and efficiency outcomes than fully autonomous systems.
The Download: Anthropic’s IPO, smarter power grids, and exercise pills
This episode covers Anthropic’s planned IPO and its implications for AI governance, advances in smart‑grid technology that integrate renewable energy, and the development of exercise pills that mimic the metabolic benefits of physical activity. Host Thomas Macaulay synthesizes the week’s most impactful stories.
Industrial Intelligence: The next growth opportunity for manufacturing
MIT Technology Review Insights reports that industrial intelligence—combining IoT sensors, AI analytics, and edge computing—is poised to drive the next wave of manufacturing growth. Companies that adopt these technologies report higher uptime, predictive maintenance, and reduced waste, signaling a shift from automation to cognitive operations.
WIRED
How a Contractor’s Claude Agent Got Into the Pentagon
An investigation reveals that a contractor’s Claude‑based AI agent was inadvertently granted access to Pentagon systems, raising alarms about the security of military infrastructure when third‑party AI tools are introduced. The incident has prompted a review of vendor AI policies and cybersecurity protocols.
A Supply Chain Attack Put Malicious Anthropic Claude Extensions on Thousands of Devices
A supply‑chain compromise injected malicious extensions into legitimate Claude plugins, affecting thousands of developers and organizations. The attack exploited a trusted update channel to distribute code that exfiltrated sensitive data, highlighting the vulnerabilities inherent in open‑source AI ecosystems.
[Read more →](https://www.wired.com/story/trojan-malware-targets-thousands-of-an