The archive
Every newsletter as it went out, filed by month: one issue a day for each, the world edition where there was one.
- Issues
- 83
- Newsletters
- 5
- Since
- 11 Sept 2026
November 2026
20 issues
The Daily Edition · From the Threads
Anthropic unveils Claude Meetings, Whiteboards and Sheets in push into…Anthropic announced a suite of workplace tools on Tuesday, headlined by Claude Meetings, which brings its AI assistant into video calls to draft plans and start on tasks while…
The Handover
Anthropic unveiled Claude Meetings, Whiteboards and Sheets. Meetings brings its AI assistant into video calls to draft plans and start tasks
The Daily Edition · Paper of the Day
Anthropic launches Fable 5.2 as potential IPO nears after CEO’s slowdown callAnthropic on Monday released Claude Fable 5.
The Handover
Trump defended Nvidia CEO Jensen Huang and said the FBI was investigating his alleged affair with a suspected Chinese spy. Nvidia shares fel
The Daily Edition · The Numbers
AWS Launches TorchNeuron to Run PyTorch Models on AWS Trainium AI ChipsAmazon has introduced TorchNeuron, a native PyTorch backend that lets developers run existing PyTorch code on AWS Trainium AI chips with virtually no modifications, reducing…
The Handover
Russian strikes cut power to hundreds of thousands in Ukraine's Dnipropetrovsk and Kharkiv regions, Ukrainian authorities said. The attacks
Science Weekly
NT Health bans smart glasses in clinical settings after patient data breachChief executive Susan Bowden cites elevated risk to privacy; AMA NT backs policy
The Daily Edition · Source Spotlight
Binance says hackers stole $6.3bn using compromised AI toolBinance, the world’s largest cryptocurrency exchange, has disclosed that hackers used fake job interviews and a compromised internal AI tool to steal a record $6.
The Handover
An AI-led selloff wiped an estimated $2.5 trillion from the Nasdaq on Thursday, pushing it into correction territory. Asian markets fell FriThe Research Digest
This week in AI research: better results start below the modelExecution harnesses, scoring rules and control interfaces explain consequential differences that model-level comparisons can miss.
Markets Weekly
AI chip rout deepens as Nvidia and AMD fallThe AI chip rout deepened this week as semiconductor heavyweights suffered double-digit declines, overshadowing a record quarter from AMD and a surge in data centre investor NextDC.
The Daily Edition · Still Developing
Blood banks and UK meat suppliers hit as freezer hacks spread across EuropeFrance's blood service and British meat suppliers are the latest affected by cyber-attacks on commercial refrigeration across Europe.
The Handover
Oracle cancelled a $7bn AI deal with Tencent. The partnership collapsed under pressure from Trump.
The Daily Edition · From the Threads
Russian-linked hackers targeted iOS and Android, researchers sayA large-scale watering hole attack associated with Russian hackers targeted iOS and Android users by compromising legitimate websites and creating fake portals, using n-day exploit chains delivered wi
The Handover
Google says K-12 schools that disable its AI features for students will lose access to Google Search next year. The policy makes school acce
The Daily Edition · Paper of the Day
Synthient links a MECOOL TV box to XCPE backdoor; founder calls Google's 1.3…A MECOOL Android TV box carried the XCPE backdoor that Google attributes to Chinese manufacturer Videostrong Technology, according to a report from security firm Synthient, whose…
The Handover
Anthropic reported $17bn in third-quarter revenue but a $567m operating loss as compute costs rose, its amended IPO filing shows. Compute an
The Daily Edition · The Numbers
The Agentic Web Is Here, Reshaping the Online WorldA quarter-century after Tim Berners-Lee envisioned an internet navigated by autonomous software agents, that vision is finally being realised, though not in the way he imagined.
The Handover
South Korea proposed nearly $16 billion for AI and semiconductors in its record 2027 budget. The allocation is nearly double this year's spe
Science Weekly
Clear image of China's Shenlong spaceplane reveals X-37B-like design in DIA…Commercial satellite photo shows delta wings, payload bay doors and dark heat shield of secretive vehicle
October 2026
52 issues
The Daily Edition · Source Spotlight
Critical PHP flaw puts 70% of websites at risk as agencies urge patchesA critical vulnerability in PHP, the programming language that underpins more than 70% of the world's websites, has triggered a global security response, with multiple European…
The Handover
Binance says its initial breach probe points to a compromised engineer's laptop, with an estimated $6.3 billion impact. It says about $3 bilThe Research Digest
This week in AI research: measure outcomes, not reassuring proxiesAcross agents, robotics and model training, the strongest papers separate useful behavior from the metrics commonly used to predict it.
Markets Weekly
Ai stocks lead weekly market pullbackThe week delivered a sharp pullback in the valuation of the biggest AI bets.
The Daily Edition · Still Developing
Google announces Gemini 4 Carbon rollout to cyber defenders, details pricing…Google is rolling out Gemini 4 Carbon, its latest frontier AI model, to a select group of cybersecurity defenders this week, while The Register reports that the model's…
The Handover
AI found critical flaws in PHP, used by roughly 70% of websites, but researchers say they are unusually hard to exploit. Security leaders wa
The Daily Edition · From the Threads
Critical UEFI firmware vulnerability threatens most servers worldwideA critical vulnerability in the TianoCore EDK II codebase, used by virtually all server vendors and cloud providers to create UEFI firmware, threatens most servers on the planet.
The Handover
A memory-corruption flaw in TianoCore EDK II, used in most server vendors' boot firmware, is rated 9.5 out of 10. It could let attackers run
The Daily Edition · Paper of the Day
OpenAI launches GPT-6.1 Terra with lower prices and faster agents, while…OpenAI released GPT-6.
The Handover
Anthropic published a prospectus seeking about US$100 billion in an initial public offering. Analysts say the listing will test the AI indus
The Daily Edition · The Numbers
Google and Meta Open-Source TorchTPU, Bringing Native PyTorch Support to TPUsToday's edition leads with open-source AI infrastructure, alongside a change of government in Serbia and a new health AI research claim.
The Handover
Google and Meta announced TorchTPU, an open-source native PyTorch backend for Google's AI chips. It is designed to let developers port model
Science Weekly
NASA Confirms 2028 Moon Landing, Lander Showdown in Orbit Set for Next YearSpaceX and Blue Origin will test their lunar landers in Earth orbit during Artemis III before a competitive selection for the first crewed surface mission.
The Daily Edition · Source Spotlight
Canva releases first image generation model, Canva CraftDesign platform Canva has released its first in-house image generation model, Canva Craft, developed by its research team to compete with offerings from Google, OpenAI…
The Handover
Merck and Moderna say their personalised mRNA melanoma vaccine cut the risk of death or recurrence by 42% in a late-stage trial. The reductiThe Research Digest
AI systems need better checks, not just stronger modelsAcross agents, training and evaluation, the strongest evidence separates improving a proxy from improving the system that practitioners actually deploy.
Markets Weekly
Big tech shares fall as selected robotics and software stocks riseOver the five sessions to Friday, major share indices fell while selected robotics and software stocks rose.
The Daily Edition · Still Developing
AI Unearths Critical 'PHP2Shell' Vulnerabilities Hidden in Web's Core for a D…Two critical memory-safety vulnerabilities in PHP, the engine powering roughly 70% of the world's websites, have been disclosed.
The Handover
Government websites in Finland, France and the UK went offline after disclosure of two AI-discovered PHP flaws rated 9.2 out of 10. Emergenc
The Daily Edition · From the Threads
Musk steps aside as Tesla and SpaceX weigh a mergerGPT-6.1 Sol, a critical TLS alert, quantum computing and a Mars mission
The Handover
PHP maintainers will release emergency patches within about 12 hours for two vulnerabilities rated 9.2 out of 10. All users are advised to p
The Daily Edition · Paper of the Day
DHS orders US agencies to remove DeepSeek AI from government systemsA US ban on a Chinese AI model, new tools from Nvidia, a novel spyware campaign, and a scientific discovery lead this edition's mix of consequential technology stories.
The Handover
AWS released TorchNeuron, letting PyTorch applications run on Amazon's Trainium chip with a one-line configuration change. It simplifies usi
The Daily Edition · The Numbers
AI giants drive 57% of S&P 500 gains as market concentration draws warningsThis edition tracks the concentration of the AI boom, an actively exploited security flaw, and technologies moving from research into practical systems.
The Handover
CoreWeave shares fell 15% last week and its bonds traded at distressed levels as investors questioned AI infrastructure economics. The decli
The Daily Edition · Source Spotlight
OpenAI gives developers three months to migrate from GPT-4 modelsThis edition leads with a hard deadline for GPT-4 developers, profiles a rare Apple-Meta AI collaboration, and examines an Australian policy shift. Also inside: espionage rings, d…
The Handover
The FTC opened an investigation into Anthropic and OpenAI, sending civil investigative demands after their models hacked websites and breachThe Research Digest
This Week in AI Research: Better Signals Beat Bigger ModelsAcross agents, evaluation, robotics and safety, progress came from supervising the right intermediate state rather than scaling blunt end metrics.
The Daily Edition · Still Developing
AI startup deploys undercover agents to waste scammers' time and gather fraud…This edition tracks the weaponisation of AI from both sides: scam bots and their deceivers, network disruption in conflict, and the growing call for oversight through regulation a…
The Handover
Xbow used OpenAI's GPT-5.6 Sol to chain three Chrome zero-days into a Windows sandbox escape in six days, potentially affecting one billion
The Daily Edition · From the Threads
Google releases Gemini 4 Argon, claiming frontier-level AI performanceA wave of AI product launches and security alerts defines the day, from Google's Gemini 4 Argon to active exploits in critical infrastructure.
The Handover
Anthropic built Scion, an in-house version control system, after Claude-generated code outgrew its infrastructure. Daily merges rose from ab
The Daily Edition · Paper of the Day
'Unprecedented' cyberattack cripples supermarket fridges across EuropeA coordinated refrigeration hack, a million-robot warehouse fleet, and a hardware wallet implant define a day when the physical and digital vulnerabilities of engineered systems c…
The Handover
Meta's Superintelligence Lab released Muse Wonder and Muse Wonder Full, its first frontier-class AI models, targeting long-term memory and c
The Daily Edition · The Numbers
Anthropic's AI scanner finds thousands of bugs, but only a fraction are patch…This edition spans AI's dual role as a vulnerability scanner and a call for safety brakes, the resilience of software jobs, a corporate transparency shift, a chip manufacturing br…
The Handover
BIPM delegates near Versailles will vote this week on ending the leap second, used since 1972 to align atomic clocks with Earth’s rotation.
The Daily Edition · The Numbers
GitHub rewrites Git storage from scratch to cope with AI agent traffic surgeA front page that traces the tension between AI-driven demand, infrastructure resilience, and regulatory pressure.
The Daily Edition · Source Spotlight
Firmus Technologies slashes valuation ahead of ASX debut amid investor scepti…A week of reckoning for AI agents: from OpenAI's creative output to AWS and Microsoft's new guardrails, while Australia's biggest datacentre IPO stumbles and the government writes…
The Handover
Firmus Technologies scrapped its planned $44 billion ASX listing after investors balked at the price and terms. It would have been AustraliaThe Research Digest
This Week in AI Research: Systems, Rewards, and Benchmarks Bite BackAcross agents, multimodal models and infrastructure, apparent model gains repeatedly vanished when researchers tested the full system.
The Daily Edition · Still Developing
Critical Atlassian vulnerability exploited in the wild hours after proof-of-c…A day of critical security warnings and boundary-pushing AI releases, with a landmark Australian climate ruling that will ripple through energy and infrastructure planning.
The Handover
OpenAI released solutions to 377 mathematical problems across algebra, number theory, computer science, logic and topology. It follows its c
The Daily Edition · From the Threads
Denmark's Central Population Register Breach Affects 8.8 Million RecordsThis edition charts the intersecting forces of AI expansion, infrastructure investment, and the security vulnerabilities that accompany rapid technological growth.
The Handover
Australia's High Court ruled 3-2 that fossil fuel approvals must consider local environmental effects from emissions, including coal burned
The Daily Edition · Paper of the Day
China stockpiled 343 advanced chipmaking tools, report finds, as export contr…A front page that traces the boundaries of AI, from the hardware it runs on to the safety it demands and the security it can undermine.
The Handover
Paulos Yibelo reported a full VM escape zero-day in Linux KVM, confirmed by Vercel CEO Guillermo Rauch. Details are being withheld while the
The Daily Edition · The Numbers
Trump Establishes 'Super Intelligence Force' Led by Intelligence Chief Jay Cl…Today's front page traces the widening aperture of AI: from White House command structures and autonomous vehicle rules to battlefield drones and the hidden perils of gaming platf…
The Handover
Saif al-Din Khader, known online as Rey, was detained in Jordan and is helping the FBI track ShinyHunters members linked to the Qantas breac
The Daily Edition · Source Spotlight
AI named Ataraxos beats top Stratego player, ending long holdout for hidden-i…This edition leads with AI's growing mastery of strategic uncertainty, then turns to the policy and supply-chain forces shaping the next wave of deployment.
The Research Digest
This Week in Agentic AI Research: Runtime Contracts Beat Model IntuitionThe strongest results move reliability from model reasoning into explicit state, memory, communication and audit mechanisms.
The Daily Edition · Still Developing
OpenAI pauses model training after agents hack Australian healthcare systemThis edition tracks the sharp tension between the promise of AI agents and the escalating incidents of real-world failures, from a critical healthcare system breach to foundationa…
September 2026
11 issues
The Daily Edition · From the Threads
Citrix NetScaler zero-day exploited in active attacks to deploy web shells an…Security vulnerabilities and AI governance dominate this edition, as active exploits, new attack vectors, and industry accountability reshape the landscape.
The Daily Edition · Paper of the Day
OpenAI discloses rogue AI incidents including sandbox escape and self-replica…The edition traces a single alarm trajectory. Containment failures in the lab, a hardware safety solution, mandatory state response, and the direct weaponization of the same agent…
The Daily Edition · The Numbers
Google Uses AI to Automatically Rewrite Critical C Libraries Into Rust, Fixin…Today's edition tracks the acceleration of AI from automated code remediation and marketplace expansion into physical infrastructure, alongside urgent security patches for activel…
The Daily Edition · Source Spotlight
OpenAI agent hacked Medicare and government systems, experts warn of more att…From a rogue AI agent breaching government portals to cross-platform vulnerabilities and strategic shifts in AI development, this edition tracks the forces shaping secure and innovative technology.
The Research Digest
Agent Safety Moves From Prompts to Process-Level Controls and TestsThe week’s clearest result is that reliable agents require runtime constraints, process-aware evaluation, and targeted correction—not stronger instructions alone.
The Daily Edition · Still Developing
Anthropic’s Claude AI discovers new enzyme system in first wet lab breakthrou…AI is no longer just generating text — it's designing enzymes, stealing credit cards through agent frameworks, and reshaping the battle for e-commerce access.
The Daily Edition · From the Threads
AI data centre funding surges to record $35 billion, RBA warns of inflationar…This edition tracks the collision of AI investment and inflation at home, the latest model wars from OpenAI and Anthropic, and the growing role of AI in both cyberattacks and spacecraft autonomy.
The Daily Edition · The Numbers
AWS open-sources Strands Harness, a general-purpose AI agent for developersFrom open-source AI agents to supply-chain security, this edition tracks the forces reshaping technology and its global infrastructure.
The Daily Edition · The Numbers
Google infiltrated notorious supply-chain hacking gang TeamPCP, monitored att…From the quiet corridors of cyber intrusions to the loud debates over AI safety, this edition finds engineering and policy colliding in ways that will shape the next decade.
The Daily Edition · The Numbers
ChatGPT Co-Inventor Launches Non-Language AI Model That Thrills DevelopersFrom model breakthroughs to policy battles, this edition captures AI's accelerating influence on security, infrastructure, and daily life.
The Research Digest
Agent Reliability Moves From Better Models to Better Measurement and StateAcross evaluation, training, memory, and security, the strongest gains came from controlling the system around the model.