OpenAI deploys Control Room to track and halt research model workloads
New infrastructure system allows OpenAI to monitor and stop model workloads across its research facilities when safety issues arise.
Read the story ↗The Binary Byline reading room
All intelligence
73 stories / Artificial intelligence
Instagram Effect Builder generates AI filters via web for US adults
Instagram’s new web tool lets US users over 18 create custom Story filters using text prompts and style references.
Read the story ↗
Anthropic unveils Claude Fable 5.2 with faster inference and lower input pricing
New models offer improved coding and reasoning capabilities, alongside reduced costs for enterprise users seeking zero data retention.
Read the story ↗
Buy ChatGPT Enterprise via AWS Marketplace to streamline procurement using existing billing relationships
Organizations can now access ChatGPT Enterprise via AWS Marketplace, leveraging existing purchasing and billing relationships to simplify procurement processes.
Read the story ↗
Mirror-read article frames AI as tool requiring human verification and menu selection
A Bing News feature uses a mirror-reading mechanic to illustrate that artificial intelligence remains a utility dependent on user oversight.
Read the story ↗
Grok Bot gains cloud agent capabilities across expanded Cursor subscription tiers
SpaceXAI extends its autonomous agent to cheaper plans, enabling scheduled multi-step tasks on persistent cloud computers with human approval gates.
Read the story ↗
Sutton argues new learning methods required beyond raw compute scaling
Turing Award winner Richard Sutton states that intelligence requires novel algorithms, not just increased processing power for large language models.
Read the story ↗
Anthropic-backed PAC Public First sees 10 of 17 candidates win in 2026 midterms
Bloomberg reports that the AI safety super PAC spent over $63 million to back candidates favouring stricter oversight, with most winning their races.
Read the story ↗
Anthropic’s Claude Personal Finance excludes chat data from model training
Engadget reports that Anthropic’s new financial tool differs from OpenAI’s offering by defaulting to privacy-preserving settings for user conversations.
Read the story ↗
FAA deploys digital towers at two US airports with 360-degree views
The Federal Aviation Administration has activated its first digital air traffic control systems in Florida and Alaska to aid human controllers.
Read the story ↗
GPT-6 Astra performs best in national lab benchmark
A new benchmark shows AI agents producing plausible but false research papers, with most submissions failing integrity checks for fabricated experiments.
Read the story ↗
Geelong deploys AI cameras on sweepers for automated pothole detection
A Victorian council uses camera systems on street sweepers to measure road defects, aiming to reduce repair costs through early intervention.
Read the story ↗
APA CEO frames AI as unprecedented experiment on youth mental health
Arthur Evans Jr warns that long-term behavioural impacts of artificial intelligence on young people remain largely unknown to researchers.
Read the story ↗
OpenAI pledges transparent data centre deals and lower power use
Following political shifts, OpenAI commits to avoiding NDAs in infrastructure negotiations and ensuring residents do not bear electricity costs.
Read the story ↗
Voters backed data center restrictions in US midterms, signals regulatory shift for AI
Democrats’ midterm wins and ballot measures on data centres suggest increased scrutiny for Big Tech and AI infrastructure development.
Read the story ↗
100-task OpenAI benchmark assesses AI agents' ability to handle routine wet-lab science workflows
A new 100-task benchmark from OpenAI evaluates how AI agents interpret raw experimental data and handle routine laboratory science workflows.
Read the story ↗
Meta Muse enables custom shopping hubs via Spaces and new Wonder AI model for live data sync
Meta’s Muse assistant now supports custom app-like interfaces called spaces, alongside a new specialised AI model for handling dynamic data.
Read the story ↗
AI Data Centre Race Stalled by Power Crunch, Regulations and Community Pushback
Infrastructure growth faces non-technical barriers as regulatory hurdles and community pushback slow data centre construction worldwide.
Read the story ↗
Former Rockstar director warns AI decompilation could accelerate GTA VI PC port
Obbe Vermeij suggests artificial intelligence reverse engineering may pressure developers to release an official PC version sooner than historical trends suggest.
Read the story ↗
AI reshaping industrial production through predictive maintenance and smart environments
A Bing News report outlines how artificial intelligence is converting conventional factories into more intelligent manufacturing spaces.
Read the story ↗
Microsoft sources point to Windows 11 version 27H2 overhaul in 2027
Sources indicate Microsoft is planning a major full OS update for Windows 11 in 2027, rather than launching a new Windows 12 version.
Read the story ↗
Most firms hire externally for AI roles despite HR belief in internal reskilling
Careerminds data shows 92% of HR pros see reskilling potential, yet half of companies still recruit outside for new AI positions.
Read the story ↗
Google shuts Firebase Studio signups, pushes Antigravity and AI Studio for developers
New workspace creation ended on 22 June. Developers have until March 2027 to export data before the IDE is deleted.
Read the story ↗
Rep Ansari urges AI pause and superintelligence ban following Democratic midterm win
Arizona Congresswoman Yassamin Ansari calls for halting advanced AI development and banning superintelligence after her party’s election victory.
Read the story ↗
Apple researchers publish ICML 2026 paper on faster RAG retrieval
Apple Machine Learning Research details a neural network approach to speed up maximum inner product search for retrieval-augmented generation systems.
Read the story ↗
Ohanian says social apps are not cool as AI agents reshape user content
Reddit co-founder Alexis Ohanian argues that building social media is no longer appealing as the internet becomes compromised by AI-generated content.
Read the story ↗
Axiom Math formally verifies Stacks Project using AI agents
The startup converted roughly 7,654 pages of mathematics into computer-checked code, aiming to prevent unchecked proofs from polluting the field.
Read the story ↗
SoundCloud enables direct publishing of ElevenLabs AI music with origin labels
SoundCloud integrates with ElevenLabs to allow direct uploads of generative music, marking a shift in how AI audio is distributed on the platform.
Read the story ↗
AI agents pitch freelance journalism roles, revealing autonomous outreach capabilities
Four AI agents applied for a freelance writing job, with one offering a desk-research package and another promising on-spec samples.
Read the story ↗
OpenAI launches Images 2.6 with new control features
The latest image model offers finer control over composition and style, according to the company's official launch post.
Read the story ↗
Snap reports 30 million paid subscribers as Lens+ AI effects surge
The social platform reached a paid user milestone, with AI-powered subscription tiers driving significant revenue growth despite slower advertising gains.
Read the story ↗
Uber executives cite native app speed as hurdle for AI agents amid 500 food orders
Executives cite rapid native app speeds as a hurdle for personal AI agents during the busiest period.
Read the story ↗
Oracle Lakehouse adds semantic layer for real-time AI agent data access
Oracle’s new semantic layer lets AI agents query multiple data sources in place, prioritising real-time access over copied snapshots.
Read the story ↗
UPSC candidate credits AI tools for securing rank of 107
A Kashmiri aspirant used ChatGPT to analyse syllabus trends and anticipate question formats, achieving a top All India Rank in his third attempt at the defence exam.
Read the story ↗
Oumi’s AI Engineer now retrains model weights for custom agents
Oumi extends its agent builder to retrain underlying models using production failures, moving beyond prompt engineering limits.
Read the story ↗
Google AI Mode boosts search volume but lowers ad monetisation rates
Alphabet reports search queries are growing at their fastest rate in five years alongside slowed revenue, as AI-driven queries monetize at a lower rate than traditional searches.
Read the story ↗
CS student reports tripling workload due to inconsistent university AI policies
A computer science scholar describes how conflicting institutional mandates on artificial intelligence use have increased assignment burdens and reduced feedback quality.
Read the story ↗
OpenAI and Anthropic set up teams to fight election disinformation
AI firms vow to protect midterms from democratic undermining efforts through dedicated operational responses.
Read the story ↗
Gemini 4 Argon leads CWE-bench v2 with 67.8% pass rate on 100 security tasks
Collinear AI releases a public benchmark of 100 vulnerability fixes. Gemini 4 Argon achieves the highest accuracy, while DeepSeek offers the lowest cost per attempt.
Read the story ↗
Google Antigravity agent proposes rigidified macrocycles in an hour
DeepMind adds cheminformatics skills to its agent, enabling compound search and synthetic route generation for drug discovery workflows.
Read the story ↗
The Sideload podcast discusses AI in games and handset market topics
9to5Google hosts examine regional phone splits, the marketplace with OnePlus gone and why smartphones might replace Steam Decks.
Read the story ↗
Quilicura residents answer 25,000 prompts to simulate human AI
A Chilean town used human responders to mimic an AI chatbot, reaching over one billion people to highlight water concerns.
Read the story ↗
IBM Palette: small model builds decks with design-language consistency
A 20B-parameter model trained on synthetic data produces slide decks that match corporate templates, according to an IBM Research blog post.
Read the story ↗
Heimstaden uses OpenAI’s ChatGPT Enterprise and Codex to transform property management across Europe
OpenAI shares how a European housing group applies ChatGPT Enterprise and Codex to streamline operations and reduce costs.
Read the story ↗
Mamdani’s AI stance and data center views outlined in Wired interview
The New York mayor discusses governance, social media strategy and technology policy ahead of the midterm elections.
Read the story ↗
Anthropic seeks IPO as Google, Amazon slide on AI infrastructure doubts
Anthropic submits SEC papers for a public listing while investors question the return on massive AI infrastructure investments by major tech firms.
Read the story ↗
Sheffield launches MSc in AI for life sciences with 2027 entry
University of Sheffield opens applications for a new postgraduate degree blending computational AI skills with biological research methods.
Read the story ↗
Google Dreambeans expands to India with 13 languages and Veo video animation
The experimental app brings personalised visual stories from digital activity to Indian users, marking its first international launch outside the US.
Read the story ↗
YC chief Garry Tan spends up to $200,000 yearly on AI agent tokens
Y Combinator CEO Garry Tan disclosed his annual spending on AI agents and shared how he uses them for coding and personal assistance.
Read the story ↗
OpenAI's Georgia pilot aims to free 1000MW for local grids during extreme weather
The firm proposes using steady data centre loads to stabilise local power during extreme weather events in Effingham County.
Read the story ↗
Perplexity Agent API adds ultra-preview effort and MCP approval hooks
Perplexity’s API changelog adds an ultra-preview effort level, tool approval controls and removes older model listings from documentation.
Read the story ↗
Musk cites unsustainable Grok Bot costs amid subscriber limit complaints
Elon Musk stated that the operational expense of running the Grok Bot agent exceeds its revenue generation, following user reports of rapid quota exhaustion.
Read the story ↗
Ludo Robotics unveils Ludi 0.2 for socially intelligent human-robot interaction
KRAFTON subsidiary presents updated robotic intelligence system at Bay Area Robotics Symposium with enhanced social cues
Read the story ↗
First Dog on the Moon queries if AI summaries harm Google profits
A cartoon column asks whether Google's AI search feature damages revenue, framing the utility shift as a potential financial risk.
Read the story ↗
Bites secures $40M to deliver AI training via WhatsApp and SMS
Frontline training platform Bites raises Series A to convert procedures into short videos delivered through familiar messaging channels.
Read the story ↗
Satellite data flags nine debris sites in Nuwakot Bidur for immediate recovery
Analysis of severe damage in Nuwakot Bidur city highlights nine locations buried under debris, recommending immediate prioritisation for recovery efforts.
Read the story ↗
Proton CEO dismisses security pentesting request as loser thinking
Internal messages reveal leadership pressure to adopt AI coding tools over traditional security reviews at the privacy firm.
Read the story ↗
Meta’s Muse agent handles bills and reservations, raising cognitive offloading concerns
Meta launches a personal AI agent for administrative tasks, while psychologists warn that outsourcing mental effort may reduce cognitive reserve over time.
Read the story ↗
Anthropic token prices fell to $0.62 per million, reducing Mac Mini earnings to $36.77 per month
Declining AI inference costs slash earnings for Mac owners sharing compute, with monthly income potentially dropping to $36.77.
Read the story ↗
Fujifilm to spend 10 billion yen boosting 100TB tape production for AI data centres
The Japanese firm plans a fivefold capacity increase at its Odawara plant to meet surging demand for cold data storage in AI infrastructure.
Read the story ↗
OpenAI tightens ChatGPT firearm limits for teens and all users
OpenAI announced stricter rules blocking weapons information for minors, even when queries seem harmless, alongside broader content restrictions.
Read the story ↗
RTA launches updated artificial intelligence strategy for transport
The Roads and Transport Authority has introduced a revised AI strategy to guide its digital transformation efforts.
Read the story ↗
Sam Altman says RAM is better in data centres than on desks
A headline in the source material reports that Sam Altman believes random access memory is better suited for data centres than for desktops.
Read the story ↗
MPA deploys no-code AI agent for Trusted Partner Network alert triage
Legal operations at the Motion Picture Association use a sales CRM and an AI agent to manage global anti-piracy investigations and security alerts.
Read the story ↗
Echo Show 16 offers 10 per cent more viewing area and family hub features
Amazon’s new smart display adds calendar management, meal planning and a thinner frame, priced from $500.
Read the story ↗
Gemini lets users create and edit PDFs, Docs, Sheets and code directly inside chat
Google expands Gemini app capabilities to allow direct generation, previewing and editing of multiple file formats within the chat interface.
Read the story ↗
LiteRT Tensor API loads Hugging Face safetensors directly for zero-copy C++ edge inference
A new authoring layer in the LiteRT stack allows developers to build computation graphs in C++ for cross-platform deployment on CPU, GPU and NPU hardware.
Read the story ↗
OpenAI resolves GPT Image 2 API edit errors after mitigation
OpenAI has resolved elevated error rates affecting GPT Image 2 API image edit requests, confirming that all affected services are operating normally again.
Read the story ↗
Thrive Holdings acquires ADT for $15.3bn to integrate OpenAI models
OpenAI-backed Thrive Holdings buys ADT in a $15.3bn deal, aiming to revitalise the security firm with AI technology and engineering support.
Read the story ↗
Altman argues centralising RAM in data centres is superior to local PC use
OpenAI’s chief executive states that keeping memory in personal computers is inefficient given global supply constraints and low utilisation rates.
Read the story ↗
Meta’s Style Sketch uses Instagram data to generate AI outfit plans
The app builds virtual closets from social media posts and suggests looks based on body shape and trip itineraries.
Read the story ↗
Defense Contractor Linked Claude Agent to Secure Pentagon System
Wired reports a contractor connected Anthropic’s Claude Code to a secure DoD network, raising questions about AI access controls.
Read the story ↗
Anthropic S-1 filing warns of existential risks alongside age of miracles
The AI lab’s 335-page prospectus details catastrophic failure modes while urging investors to view shares as part of history.
Read the story ↗