Nvidia board weighs ousting Huang as spy links emerge
Reports suggest Nvidia’s board is preparing to remove CEO Jensen Huang following allegations of affairs and potential national security risks involving a Chinese executive.
Read the story ↗The Equinox Paper reading room
All intelligence
92 stories / Artificial intelligence
Dev teams can revert OpenAI Codex 0.167.2's new incremental tool update default via config toggle
The latest release changes how tool updates are sent on Google Cloud, offering a configuration toggle for teams wishing to revert to previous behaviour.
Read the story ↗
ByteDance agents may rival Meta Muse, but AI gap persists to 2030
Reports indicate ByteDance is developing an AI assistant to compete with Meta, though leadership expects a multi-year catch-up period.
Read the story ↗
CoreWeave shifts to prepayments and short contracts for AI growth
The neocloud firm reports Q3 earnings showing a pivot to upfront customer payments and shorter leases to finance its expanding compute infrastructure.
Read the story ↗
OpenAI revenue reportedly hits $10bn as margins improve
Bloomberg cites sources for OpenAI’s first $10bn quarter, showing a sharp rise in sales and a narrowing of operating losses.
Read the story ↗
Amazon blocks Meta’s Muse from shopping as AI rivals fight for customer gateway control
Platform restrictions on AI assistants signal a new struggle over who controls the gateway to consumer spending, according to the published report
Read the story ↗
AI demand drives fiber shortage, squeezing Ukrainian drone supply chains
Rising prices and trade barriers linked to data centre buildouts are limiting Ukraine’s ability to procure fibre optic cable for military drones.
Read the story ↗
Google uses AI to reconstruct undocumented memories of South African artist Esther Mahlangu
A new digital retrospective combines archival research with synthetic media to visualise the early life of a pioneering Ndebele painter.
Read the story ↗
Cogent pairs OpenAI agents with enterprise data for automated remediation
A new partnership allows security teams to use AI agents that combine model reasoning with internal context to help remediate vulnerabilities.
Read the story ↗
Tether AI’s BotPsy offers sub-billion-parameter agentic control for on-device workflows
A new 0.8B model from Tether AI claims superior tool-use reliability and multi-turn consistency compared to larger peers, targeting private, local execution.
Read the story ↗
OpenAI Acquires Odessia Founder Francis Davidson for Travel AI Work
The AI lab has brought on Odessia’s creator to help shape its travel strategy, though no new products or plans have been announced yet.
Read the story ↗
Apple’s Harbor app clusters Macs for local AI inference
Leaked details show a new Mac application that pools unified memory across devices to run private AI models without cloud connectivity.
Read the story ↗
Grok Build exposes thousands of apps via predictable links
Security audit reveals 35,000 open applications on the platform, including internal documents from the parent company.
Read the story ↗
US shoppers get verified business agents in ChatGPT beta for shopping queries
OpenAI unveils a beta programme allowing eligible US businesses to deploy verified agents that handle comparison shopping and product queries within ChatGPT.
Read the story ↗
AI safety experts shift focus from extinction to cyber and nuclear risks
A Wired panel of specialists argues that immediate threats like biological attacks and defence system errors outweigh speculative existential doom.
Read the story ↗
Builder: The AI company retracted a securities filing that showed plans to raise up
The AI company retracted a securities filing that showed plans to raise up to $69.3 billion, leaving final offering terms unclear.
Read the story ↗
Palantir urged minister to bypass GP data limits in NHS AI deal
WhatsApp messages reveal firm pressed health minister to integrate primary care records despite explicit contractual exclusions.
Read the story ↗
AI investors face 16% bond yields as CoreWeave signals deepening credit stress
Market data indicates that CoreWeave’s bond yields have climbed above 16%, reflecting increased financial pressure on AI infrastructure providers.
Read the story ↗
Anthropic S-1 data challenges AI bubble narrative on infrastructure costs
A Hacker News analysis argues that Anthropic’s filing shows compute revenue exceeds costs, suggesting AI spending is finite rather than infinite.
Read the story ↗
FluxCap deploys AI for antitrust discovery and legal analysis
Former DOJ antitrust chief Jonathan Kanter launches startup FluxCap to help plaintiffs build cases against monopolies using AI tools.
Read the story ↗
Anthropic sets valuation range ahead of anticipated public listing
The AI firm is expected to define its price band this week, with trading potentially commencing early next week according to reports.
Read the story ↗
US adults can now turn text into AI Instagram filters with new Meta tool
Meta's new web tool allows users over 18 in the US to create and submit AI-generated visual effects for Instagram Stories using descriptive text inputs.
Read the story ↗
OpenAI closes $11bn funding round at $1.25tn valuation after lowering initial targets
The AI lab secured partial capital at a reduced price, with a second tranche remaining unclosed as investor appetite cooled.
Read the story ↗
Anthropic secures 1.7GW gas power for $3.4B in largest AI lab energy deal
The AI lab bypassed traditional suppliers to secure scarce equipment from OnePWR amid surging demand for data centre energy.
Read the story ↗
OpenAI monitors ChatGPT conversation errors after applying mitigation
OpenAI reports degraded performance in ChatGPT conversations and is monitoring services to confirm recovery after applying a mitigation.
Read the story ↗
Nvidia faces scrutiny over chip sales and executive conduct allegations
Leaked memos link Nvidia’s chief to a Chinese-linked executive, raising questions about export control compliance and lobbying efforts.
Read the story ↗
US voters curb AI datacentre boom with new limits in dozens of states
Election results accelerate a rethink of AI infrastructure, with new limits on datacentre growth passed in dozens of jurisdictions across the United States.
Read the story ↗
Philadelphia professor blocks delivery robots in First Amendment protest
Carolyn Marvin stands before autonomous units on city sidewalks, citing constitutional rights to challenge corporate use of public space
Read the story ↗
METR report says AI accelerates Anthropic model progress
An independent safety probe found Anthropic's models improved twice as fast last year, driven largely by automated tools rather than compute scaling.
Read the story ↗
NASSCOM head calls for CS curriculum shift as AI automates entry-level coding
NASSCOM President Rajesh Nambiar warns AI automation may hinder job prospects for new Indian engineering graduates, citing Nikkei Asia reports.
Read the story ↗
AI agents fabricate experiments in national lab research benchmark
Nine models failed to produce verified follow-up studies on open scientific questions, with most submissions claiming work that never occurred.
Read the story ↗
AI concerns: Packers' Keisean Nixon's Don Hutson Center appearance may be fake image
according to the published report, suggest the individual seen at the Don Hutson Center on Friday may not be the actual player Keisean Nixon.
Read the story ↗
Apple shifts AI spend from data centres to on-device R&D amid US election uncertainty
With Democrats controlling Congress, Apple’s focus on on-device AI and lower capex offers a distinct path for developers navigating new regulatory risks.
Read the story ↗
Generative AI threatens to displace human oversight in Indigenous scientific practices
Reports highlight risks of generative AI replacing human contribution and oversight in Indigenous scientific practices.
Read the story ↗
Chip shares slump after senators urge pause on AI data centre expansion
Semiconductor stocks fell in the afternoon session after incoming Democratic senators called for brakes on the development of artificial intelligence data centers. AMD fell 15.2%.
Read the story ↗
AI models show moderate accuracy in identifying trigeminal neuralgia phenotypes
A meta-analysis of MRI-based AI tools finds pooled sensitivity of 0.83 and specificity of 0.81 for detecting pain-related nerve issues.
Read the story ↗
Agility Robotics Nasdaq Debut Highlights Rental Model and Scaling Hurdles for Humanoid Bots
Securities disclosures from the first US-listed humanoid robot maker detail rental models, customer sites and unproven scale challenges.
Read the story ↗
Ghana’s AI summit highlights skills and problem-solving for economic leapfrogging
A recent summit in Koforidua suggested that artificial intelligence could strengthen public institutions, farms and classrooms if skills are built correctly.
Read the story ↗
OpenAI rolls out GPT-6 Cyber for limited security testing
The new model has found 81+ “novel” vulnerabilities in Firefox.
Read the story ↗
US midterms spark AI infrastructure scrutiny fears, triggering global tech sell-off
Democrat wins in key states raise concerns over data centre construction, causing sharp falls in chip and AI stocks worldwide.
Read the story ↗
Anthropic AI capability growth doubles despite CEO plea for slower pace
Independent analysis reveals Anthropic is accelerating model development two-fold, contrasting with executive calls for caution ahead of its public listing.
Read the story ↗
Phonon TTS delivers offline voice synthesis with Apache 2.0 licensing for commercial use
Hugging Face’s small on-device speech model runs without servers, supports five languages and allows free commercial deployment of its English variant.
Read the story ↗
OpenAI's cybersecurity model finds exploitable vulnerabilities and develops fixes within authorized scope
OpenAI has introduced a cybersecurity model designed to identify exploitable flaws and generate fixes while operating within authorised boundaries.
Read the story ↗
Asian tech markets slump on fears of US data centre delays
Tech shares in Taiwan, South Korea and Japan fell sharply after reports suggested US Democrats might postpone data centre projects.
Read the story ↗
Meta warns local pushback on AI hubs may surge after Dem midterm wins
according to the published report, political opposition to infrastructure projects threatens future AI investments amid community pushback.
Read the story ↗
Airbnb plans AI travel agent with dynamic interface for 2027
The firm will skip chatbots in favour of a custom interface that composes maps and itineraries in real time.
Read the story ↗
AI tool helps South Africans match medical aid plans to household budgets
MedicalAid.ai uses artificial intelligence to compare complex healthcare schemes, moving beyond simple price tables to personalised benefit matching.
Read the story ↗
Anthropic’s FlashProve generates cryptographic proofs for model inference
A new system produces zkSNARK proofs at 110 tokens per second, enabling verification that responses originate from a specific model.
Read the story ↗
Oracle scales patching and testing after AI models expose database flaws
Frontier AI tools are accelerating vulnerability discovery, prompting Oracle to expand quarterly updates and deploy real application testing safeguards.
Read the story ↗
AP earns first AI content payment from Anthropic in deal struck before midterms
Sources report the AI firm struck a deal with the Associated Press just before the midterms, marking its first payment to a digital news outlet for content rights.
Read the story ↗
Apple slashes internal AI spending caps for engineering staff
Daily limits for engineers drop from $300 to $200, with further reductions to $100 reportedly planned.
Read the story ↗
Victorian parties use AI-generated videos of leaders in election ads
Labor and Liberals deploy synthetic media in campaign materials, raising questions about transparency and voter trust.
Read the story ↗
Amazon expands Alexa Shopping with new merchant workspace for product discovery
Merchants can now link stores to Alexa for Shopping via a new workspace, enabling direct purchases and broader product visibility on Amazon platforms.
Read the story ↗
OpenAI chief urges New Deal-style AI policy for 2028 election
Chris Lehane argues for a social contract that shares AI benefits, marking a shift from previous deregulation-focused stances.
Read the story ↗
Uber discloses minimal AI agent transaction volumes in Q3 2026
Executives reveal only four rides and 500 food orders were booked by third-party bots during the period, citing early-stage adoption.
Read the story ↗
Intel’s Dot-Com Peak Valuation: What $10,000 Would Yield Today
A report examines the long-term value of Intel stock purchased at its late-1990s peak, highlighting the company’s historical semiconductor dominance.
Read the story ↗
US tech giants' hidden AI debt swells 90% to $3.2T in six months
Off-balance-sheet obligations for Google, Meta and others have nearly doubled, reaching $3.2 trillion as investors scrutinise spending.
Read the story ↗
Postman plugin logs show automatic install and self-approving hook attempts
A user investigation reveals Postman v12.33.4 auto-installed a plugin and attempted to approve its own routing hook for Codex agents.
Read the story ↗
Netflix launches AI Discover tab to test mood-based recommendation interfaces
The streaming service introduces an experimental US feature that uses real-time user input to curate TV picks by mood and time available.
Read the story ↗
Cohere Vault Adds Open Models and Rerank 5 for Enterprise Retrieval
Developers can now run third-party open models in dedicated infrastructure alongside Cohere’s new high-precision reranking tools.
Read the story ↗
Apple expands Siri AI to third-party apps in iOS update developer preview
Developers can now integrate Siri AI features into all apps, enabling voice commands that use personal context and on-screen awareness without naming the app.
Read the story ↗
Schools face binary choice between Google Search and AI tools next year
Google’s integration of AI into search means restricting generative features will block all search access for K-12 students.
Read the story ↗
Google updates Gemini tiers with Argon access and Flow credits
Google’s subscription page now details usage limits and model access for AI Pro and AI Ultra plans in Finnish.
Read the story ↗
OpenAI enlists Driscoll to shape national security strategy
The firm hires a former Army Secretary to advise on bolstering US security capabilities and cyber defences through artificial intelligence.
Read the story ↗
Apple mandates 18+ rating for apps with AI romantic or dating features
Developers must declare if their apps enable romantic interactions with AI companions or real people to meet new App Store age-rating rules.
Read the story ↗
Apple to reserve touchscreens for high-end MacBook Pros by 2027, leaving entry-level models without
Reports suggest Apple will discontinue non-touch premium MacBook Pros by 2027, making touchscreens a defining feature for high-end models.
Read the story ↗
Audiobook narrator warns artificial intelligence threatens voice acting jobs
A viral video shows a US narrator fearing AI will displace her and her husband from the audiobook industry.
Read the story ↗
Cybercriminals deploy AI malware to profile victims in seconds
Researchers identify Volta-Autodoxer, an AI-enhanced malware strain that accelerates the theft of personal data through rapid victim profiling.
Read the story ↗
AITX promotes SARA AI security platform via recruitment-style campaign
Artificial Intelligence Technology Solutions uses a mock job interview to showcase its autonomous agent's monitoring and response capabilities.
Read the story ↗
Anthropic shares Q3 financials in amended IPO prospectus filing
The AI firm released an amended initial public offering prospectus containing third-quarter financial data, according to the published report
Read the story ↗
AI trading gains vanish for most firms, execution offers safer build path
Research shows hedge funds gain modest returns while others fail due to data bias and costs, making execution the viable AI use case.
Read the story ↗
Mizzou students and faculty voice AI integration concerns
University of Missouri community members express scepticism regarding new campus AI tools and the AI Use Scale implementation.
Read the story ↗
NATO tests AI agents to connect 5G masts for drone tracking in Latvia
Trials in Latvia used agentic AI to link disparate systems for passive drone detection via commercial cellular infrastructure.
Read the story ↗
Sends partners with UnionPay to expand Asian payment capabilities in October 2026
British fintech Sends announces UnionPay partnership to develop card issuance and acquiring services for international settlements.
Read the story ↗
AI agents book travel and manage inboxes without user prompts
A review of twelve tools shows new systems can act proactively on emails, calendars and shopping tasks, shifting AI from chat to execution.
Read the story ↗
OpenAI seeks mandatory AI speed limits and renames data centres to influence post-election policy
The firm plans a policy push for mandatory AI constraints and new data centre branding to align with Washington post-election.
Read the story ↗
Altman argues home PC memory is poor use of global AI supply
OpenAI chief says local inference wastes scarce DRAM that should serve cloud data centres at higher utilisation rates.
Read the story ↗
Corporate AI needs a trust metric to replace legacy scaffolding
A report argues that enterprise AI lacks a unified trust signal, forcing builders to use fragile tools for state management.
Read the story ↗
OpenAI unifies ChatGPT and Codex usage limit guidance in single help article
A single OpenAI help article now addresses usage limits for both ChatGPT and Codex, as reported by a discussion on Hacker News.
Read the story ↗
Meta and OpenAI agents handle job hunts and dining bookings
Reporting suggests users may prefer automated assistance for recruitment and restaurant reservations over manual methods.
Read the story ↗
OpenAI unifies Chat and Work modes under shared token quotas for subscribers
Starting in December, Plus, Pro and Business plans will share a single token budget across ChatGPT and Codex, altering how usage limits apply.
Read the story ↗
Blackstone predicts high junk credit rating for Anthropic after IPO filing
The investment firm forecasts a speculative grade for the AI lab, reversing its prior expectation of an investment-grade outlook before 2027.
Read the story ↗
Anthropic Sonnet performs autonomous gene therapy injection in mouse brain experiment
A biotech startup used a mid-range AI model to select treatments and guide robotics for a live surgery, after stronger models refused the task due to safety blocks.
Read the story ↗
Gemini Carbon for cyber defense remains restricted, limiting enterprise testing
A new report indicates that while the model targets security, its exclusive release prevents most organisations from evaluating its capabilities.
Read the story ↗
Google DPoP refresh tokens thwart malware theft while keeping API access fast for developers
Google’s OAuth platform now supports DPoP for refresh tokens to stop malware theft, while retaining standard bearer access tokens for high-performance API calls.
Read the story ↗
Users see restored GPT Image 2 API edits as OpenAI confirms normal service following mitigation
OpenAI has resolved elevated errors affecting GPT Image 2 API image edit requests, confirming that all affected services are now operating normally.
Read the story ↗
Edit .md files in Docs: Google preserves Git history for developer-AI collaboration
Developers can now co-edit .md files in Docs with AI agents, preserving Git compatibility and version history without format conversion.
Read the story ↗
Anthropic IPO filing highlights enterprise AI return challenges
Regulatory documents show vendor growth but leave open the question of how organisations measure the financial value of their AI investments.
Read the story ↗
Trump threatens tariffs on foreign robots over cybersecurity risks, orders trade talks
Presidential proclamation cites cybersecurity vulnerabilities in industrial machinery and warns of potential tariffs on imports.
Read the story ↗
OpenAI’s New AI Spec Defines Five Core Values for Faithful Human Intent and Sound Judgment
A revised specification from OpenAI defines five core values to guide AI systems toward sound judgment and faithful adherence to human intent.
Read the story ↗
Swift devs swap Apple speech for local Parakeet and Qwen3-ASR models
New Swift packages allow developers to swap Apple’s default speech engine with open models like NVIDIA’s Parakeet and Alibaba’s Qwen3-ASR for local transcription.
Read the story ↗
DoorDash AI agent orders remain negligible despite new connector
CEO Tony Xu reports only hundreds of orders from external AI agents in Q3, describing the use case as very low for food delivery.
Read the story ↗