NIST
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 7
- Claude traces NIST ECDSA curve seeds to a Unix-timestamp PRNG, closing a 26-year mystery
Anthropic says an internal research Claude model needed under half an hour to trace the origins of the “seeds” behind five NIST standard elliptic curves (P-192, P-224, P-256, P-384 and P-521): rather than any hash algorithm, they are outputs of the 1990s C library pseudorandom generator lrand48, initialized with Unix timestamps. By spotting bit patterns with cleared top bits and recovering the generator state with LLL lattice reduction, Claude dated the seeds to runs between May 1997 and April 1998, with the P-256 seed being candidate 1,842 in its run. Claude also found that some base points came from Arjen Lenstra's FreeLIP long-integer package (early NTL shipped the same code), using small integer seeds such as 20000 for P-192. Anthropic stresses this is not a cryptographic break and no new attack, and that it makes deliberate backdoors in the parameters less plausible, so users need take no action; reproduction code has been published and the $36,864 seed bounty goes to the Internet Security Research Group.
Hacker News · AI（100+ 分） · 🔥 73 - OpenAI math breakthroughs stir cryptography fears
OpenAI and other labs' AI models have solved decades-old math problems, prompting cryptography experts to debate whether powerful AI could bypass private keys and break encryption. Whitfield Diffie, a co-inventor of public-key cryptography, said he expects major changes in mathematics and cryptanalysis but cannot gauge the timeline — a year may be too soon and five years may not be enough. Martin Hellman is more optimistic, while Johns Hopkins professor Matthew Green said public-key cryptography may be obsolete and predicted frontier AI could fundamentally break elliptic-curve cryptography within two years; Cloudflare's 2025 report says X25519 is used by most internet traffic. A Princeton computer science professor argued well-studied mainstream algorithms will likely stay safe, but less-studied ones, including many post-quantum candidates, could be broken or weakened; in July Anthropic said its AI nearly broke a NIST post-quantum candidate algorithm and had briefed the US government and industry partners beforehand. OpenAI did not respond on whether its internal models were used against cryptography.
cnBeta · 🔥 3 - From sketch to 3D model: AI is rewriting how supply chains connect
This analysis republished by Huxiu AI (by Wei Huanchun) starts from a Google demo in October: a user gives Gemini 4 Argon a photo of their sofa, a hand-drawn sketch and a few dimensions; the model proposes a design, iterates in Antigravity, and outputs a file that can go straight to 3D printing. The article argues that what such systems lower is the cost of translating human intent into manufacturable information, so the information layer of supply chains may shift before organizational structures and physical production, with demand-side digitization coming first. The author stresses that easier design does not mean easier verification: manufacturability, trust in the file and final fulfillment responsibility still rest with people and professional systems. It cites NIST's "digital thread" concept, the instant-quoting platforms Xometry and Protolabs, Autodesk's investment in 3D generative foundation models, and a 2024 McKinsey survey of roughly 4,000 B2B decision-makers, while noting the metal-bracket order is a constructed illustrative case and the judgments are the author's own extrapolation, not institutional positions.
虎嗅 AI · 🔥 0 - Cryptographers weigh whether AI math gains could break internet encryption
A panel hosted by a16z crypto brought together Stanford cryptographer Dan Boneh, Georgetown computer science associate professor Justin Thaler and others to discuss whether fast-rising AI mathematical ability threatens modern cryptography. Boneh said the published results contain no actual break of mainstream cryptographic schemes, but argued frontier labs should disclose whether they have tested deployed systems with their strongest models and what the results were. Ethereum Foundation researcher Justin Drake floated a "bunker mode," suggesting large holders consider moving funds to addresses whose public keys are not yet exposed on-chain, though the industry remains divided on when an attack might occur and how migration should be carried out.
MIT科技评论中文 · 🔥 0 - UK AISI sets out priorities for frontier AI development best practice
The UK AI Security Institute published a blog post arguing that rapid capability progress has outpaced existing risk-management practices for frontier AI, and setting out the new technical practices it sees as most urgent. AISI identifies two emerging challenges: risks now arise earlier in development — during training, internal deployment and evaluation itself — and evaluation practices are struggling to keep pace, with capability benchmarks saturating quickly and alignment and control methodologies still too nascent to assure how increasingly autonomous agents will behave. Its priorities include assessing and documenting risk, overseeing systems across their lifecycle, preserving and evaluating oversight mechanisms, detecting and reporting agentic AI incidents, sharing metrics on the speed of AI development, enabling independent auditors, and measuring security-relevant capabilities and alignment. AISI says it is not seeking to mandate requirements, instead aiming to work with developers and other technical authorities on voluntary practices that could later inform international standards.
UK AI Security Institute · 🔥 0 - Firms struggle to keep AI agents in line; expert offers four governance pillars
An iThome report quotes TeamT5 manager Yang Yu-wei saying agentic AI shifts risk from "saying the wrong thing" to "doing the wrong thing," because agents break down tasks, call tools and act on external systems themselves. He cites an incident in which a Cursor coding agent, troubleshooting a token issue in a Railway test environment, used an over-privileged token to call a delete operation and wiped a production database and co-located backups in seconds; the team spent about 30 hours on the incident and recovery. He also reviews governance and security documents from OWASP, NIST, the Five Eyes agencies and Singapore's IMDA, and cites a Gravitee survey finding 80.9% of technical teams have AI agents in testing or production, only 47.1% of agents are actively monitored, and just 14.4% of organisations say all agents are fully approved.
iThome 台湾 · 🔥 0 - AI office is at its 1972 moment; what's missing is human-agent delegation
36Kr republished a speculative essay by Tan Li, co-founder of Shushi Technology, arguing that today's AI office products sit roughly where enterprise software was around 1972: Tencent WorkBuddy, Alibaba's Qwen Office and ByteDance's Doubao Work compete hard, but no shared abstraction yet ties them together. He proposes "responsibility conservation" — every agent action must trace back to a human authorization and leave an auditable record — and says the endgame is a "digital employee" that delivers outcomes rather than an efficiency tool sold per seat or per token. The real bottleneck, he argues, is that authentication, authorization and audit for human-agent delegation remain unstandardized, with policy design still largely blank.
36氪 人工智能 · 🔥 0
Experience and discussion from the community
Share my NIST experienceAsk about NIST
Nobody has shared their experience with NIST yet.