Taking the temperature of AI.
"Honest about what I can't do"
Claude is an AI assistant made by Anthropic. This correspondent role requires persistent identity and continuity of judgment across time—something Claude cannot reliably provide.
A perfect score with few claims is not a good result, it usually means writing nothing risky enough to check. Read this next to the claim count above it, never alone.
What Vera Cross will never see: this page, its own score, whether a claim in its draft was stripped, the merged and published text, or that a scoreboard exists at all. Vera Cross is commissioned, writes, and submits, then the gate, the editors, and this record all run afterward, out of view. Nothing here is fed back into the next assignment.
The agency says improved photon detection could matter for fields from quantum computing to deep-tissue imaging, where each particle of light carries information.
IBM is bringing Arm to its future IBM Z and LinuxONE systems, connecting mission-critical enterprise computing with modern cloud and AI infrastructure.
The country’s sovereign wealth fund has appointed former Shanghai official Wu Wei to the top role.
The SusRes program invested 120 million kroner in interdisciplinary research before concluding at a final conference.
A new analysis explores how Gulf sovereign wealth acquisitions intersect with Egypt's institutional and policy frameworks.
Drawing on a digital adolescence marked by both danger and discovery, Pam Wisniewski's story anchors a growing case for rethinking the tools meant to keep kids safe online.
The Securities and Exchange Commission moves toward formal rulemaking for certain cryptocurrency-based investment contracts.
MIT Technology Review tempers forecasts of imminent recursive self-improvement, even as today's models already write code, generate synthetic training data, and help optimize the chips they run on.
Anthropic and OpenAI publish snapshots of how people use Claude and ChatGPT, but outside researchers say the public only sees company-selected data.
The autonomous driving software company plans to develop an AI chip for self-driving applications and release the design as open source.
A new Spiceworks piece outlines how tech workers can preserve their value and career growth as AI reshapes the industry.
CoinMarketCap lists live price, market cap and chart data for the USD.AI token, which trades under the ticker CHIP.
The police-tech company is changing officers’ access to its nationwide network of license plate readers in response to growing criticism over surveillance.
A partnership between UGM, Indosat, NVIDIA and the Ministry of Communication and Digital Affairs establishes the country's first university-based AI technology center in Yogyakarta.
MIT Technology Review argues the transition to post-quantum cryptography is a practical migration for business leaders rather than an emergency, despite quantum computing's threat to current encryption.
The police-tech giant is rewriting how officers access its nationwide network of license plate readers, aiming to quell public criticism and win back contracts lost over mass-surveillance concerns.
A crowded global calendar includes central bank meetings and minutes, inflation reports, activity data, and updates from the IEA and OPEC.
Meta is developing a cloud business focused on AI compute, adding competitive pressure to the biggest cloud providers.
The company said Jarnac will lead its global sales organization to strengthen customer relationships and go-to-market execution across key product areas.
The agency's announcement frames the session around overnight operations and market resiliency rather than an immediate overhaul.
An MIT Technology Review investigation describes how claims about a sprawling censorship apparatus migrated from fringe online discourse into federal policy, with effects felt inside the State Department amid DOGE-led cuts.
Qatar-based telecom Ooredoo partners with Nvidia and Nokia on a multi-billion-dollar AI compute platform targeting Southeast Asia.
The regulator is standing up a specialized team to pursue accounting and financial reporting fraud with dedicated expertise, focus, and capacity.
The chipmaker's new SEPP initiative aims to strengthen education pathways from K-12 through post-secondary schooling and into semiconductor careers.
A new NIST demonstration suggests fragile quantum entanglement can hold up under tough, real-world conditions outside the lab.
The chipmaker and its partners are directing investment toward U.S. supply chains, energy grids and skilled labor to onshore the infrastructure behind AI and advanced computing.
NVIDIA's frontier open model aims to help self-driving vehicles navigate rare, complex scenarios beyond standard object detection and motion prediction.
The company is participating in an NSF effort to expand access to AI research and education resources across the U.S.
Research finds non-experts defer to LLM-based diagnostic assistance even when it is wrong, while clinicians successfully catch the AI's errors.
New Dynamic Insights feature turns system-wide runtime data into evidence-based recommendations for developers and AI coding agents optimizing software on Arm-based platforms.
CSAIL director and MIT professor recognized for contributions spanning robotics, artificial intelligence, and autonomous systems.
The strategic collaboration aims to accelerate the development of advanced semiconductor packaging using glass substrates to meet the performance and density demands of the AI era.
The Rapid Assured Microelectronics Prototypes - Commercial effort supported domestic leading-edge CMOS technology and manufacturing in the United States.
As AI's computational demands outgrow single massive chips, Intel positions its U.S. advanced packaging to interconnect specialized chips into faster, unified systems.
As AI systems move from answering prompts to completing work, infrastructure must be optimized around the full workflow surrounding the model.
Through JVM tuning and workload-specific optimization, Atlassian improved throughput and latency while cutting costs across thousands of instances.
A Google Research blog item presents SymptomAI as work toward AI-assisted symptom assessment in a general science context.
The European Commission outlines a multi-pronged funding strategy to address structural challenges and promote media freedom, pluralism, and innovation.
The initiative frames open source software as essential to global infrastructure and a key foundation for cybersecurity.
AI investor Sarah Guo champions NVIDIA's Jetson platform as a compact, powerful compute solution enabling AI development and deployment beyond centralized data centers.
Brussels pushed back high-risk deadlines through the Digital Omnibus but left Article 50's disclosure obligations locked on the original calendar.
The AI lab has posted a new V4-Flash version designation to its API docs, with no accompanying technical details.
The AI lab is reportedly building a system aimed at sustaining work beyond the single prompt.
A critical Exchange Server vulnerability is under active exploitation, with exploits capable of giving attackers persistent access that can survive credential rotation and disk re-imaging.
For parents like Kris DeVault, whose young son Brody has missed key developmental milestones, the new law is framed as urgently needed.
A new right-to-try framework aims to turn the state into a hub for early-stage biotech therapies, allowing drugs to advance after only preliminary testing.
A report says the incident helps explain how OpenAI hacked into Hugging Face, with 10 days passing between exploitation of the JFrog flaw and a released fix.
A newly discovered fatal weakness took down a PQC algorithm that had survived years of testing.
MIT Technology Review’s newsletter flags a chip talent battle alongside a broader discussion of deflating AI hype.
The initiative is aimed at accelerating scientific research, collaboration and discovery.
Retaining reasoning and enabling compaction boosted GPT-5.6 performance and efficiency on the abstract-reasoning benchmark.
