trending now
quality: all 6+ 8+
0

This article covers how Yakult delivery women in Japan provide social connection and combat loneliness in an aging population. It is not a security article.

bbc.com · ranit· 27 days ago · 15 min · not-security · details · hn 396
2
0

LoGeR is a novel deep learning architecture from DeepMind and UC Berkeley for 3D geometric reconstruction from extremely long videos (up to 19,000 frames) using a hybrid memory module that combines Sliding Window Attention for local precision with Test-Time Training for global consistency, achieving state-of-the-art results on KITTI, 7-Scenes, and long-sequence benchmarks while maintaining sub-quadratic complexity.

loger-project.github.io · helloplanets· 24 days ago · 5 min · research · details · hn 149
1
0

A satirical RFC proposing the Human Em Dash (HED) — a Unicode character to distinguish human-authored text from AI-generated content by marking em dashes with a Human Attestation Mark preceding them. The mock standard humorously explores the problem of "Dash Authenticity Collapse" and includes behavioral verification requirements like detecting pauses and backspaces.

gist.github.com · jdauriemma· 24 days ago · 9 min · satire · details · hn 132
0
github.com · networked· 26 days ago · details · hn 326
7
0
addxorrol.blogspot.com · tdullien· 9 days ago · 4 min · research · details · hn 1
1
0

This article is about vintage HyperCard applications based on William Gibson's Neuromancer trilogy of science fiction novels, not a security topic.

macintoshgarden.org · naves· 23 days ago · off-topic · details · hn 121
7
0

A detailed technical analysis of an LLM-generated SQLite reimplementation in Rust that demonstrates critical performance failures (~20,000x slower) despite appearing correct. The article identifies two root-cause bugs: a missing INTEGER PRIMARY KEY optimization that forces full table scans instead of O(log n) B-tree lookups, and unnecessary fsync calls on every statement, alongside compound inefficiencies in AST caching and memory allocation patterns.

blog.katanaquant.com · dnw· 27 days ago · 28 min · research · details · hn 460
1
0

This article discusses the emotional labor and working conditions of AI training workers, particularly in Africa, who help develop AI intimacy features. It is not a security article.

data-workers.org · beepbooptheory· 21 days ago · non-security · details · hn 63
this feed is available as an API
GET /api/v1/news/?min_rating=7&days=1
1
0

This article covers archaeological research on Monte Sierpe in Peru, where 5,200 holes carved into a mountain ridge have been identified as likely part of an ancient indigenous trade and accounting system from the 14th century, possibly used by the Chincha Kingdom as a marketplace or khipu-like device for tracking goods and equivalence without currency.

newatlas.com · defrost· 24 days ago · 12 min · archaeology · details · hn 146
0
cacm.acm.org · zacwest· 26 days ago · security · details · hn 363
0
youtube.com · kevinak· 26 days ago · details · hn 255
0
netaskari.substack.com · giuliomagnifico· 14 days ago · 6 min · security · details · hn 5
2
0

Hume AI open-sources TADA, a novel text-to-speech system using one-to-one text-acoustic token alignment that achieves 5x faster inference than competing LLM-based TTS systems while eliminating hallucinations and enabling on-device deployment. The approach synchronizes text and speech through a dual alignment tokenization schema, reducing the token mismatch problem inherent in traditional audio representation methods.

hume.ai · smusamashah· 23 days ago · 7 min · research · details · hn 102
0
synthesia.io · gbrindisi· 15 days ago · 11 min · security · details · hn 7
0
github.com · TheOpenSourcer· 11 days ago · 5 min · details · hn 2
0
github.com · aebtebeten· 26 days ago · 7 min · details · hn 333
1
0

An opinion piece arguing against using LLMs for programming, emphasizing that the learning process and deep understanding gained through hands-on problem-solving is more valuable than relying on AI-generated code.

neilmadden.blog · ms7892· 22 days ago · opinion · details · hn 68
0
BrainDAnderson· 14 days ago · 3 min · security · details · hn 6
1
0

Nicholas Carlini shares his opinionated methodology for conducting high-impact research and winning best paper awards, covering problem selection, collaboration strategies, literature review practices, and paper writing principles developed from his experience in cryptography and ML security research.

nicholas.carlini.com · mad· 24 days ago · 55 min · opinion · details · hn 169
2
0

The USDA is closing its South Building and relocating thousands of employees from Washington, DC to regional hubs in North Carolina, Missouri, Indiana, Colorado, and Utah as part of the Trump Administration's federal workforce reduction. This reorganization risks significant productivity loss and undermines food assistance programs, drawing parallels to a previous relocation of the Economic Research Service that resulted in over 50% staff loss.

foodpolitics.com · speckx· 22 days ago · 5 min · news · details · hn 63
0
radar.cloudflare.com · secondary_op· 24 days ago · security · details · hn 126
2
0

Hisense and other VIDAA-based smart TVs are displaying unskippable advertisements at startup before users can access live TV, raising concerns about forced ad delivery, potential privacy implications, and the absence of user controls to disable this behavior.

guru3d.com · akyuu· 24 days ago · 4 min · news · details · hn 124
1
0

A user reports experiencing 401 errors and OAuth session issues with Claude, asking if it's a widespread outage.

coderbants· 23 days ago · off-topic · details · hn 86
3
0

Iranian state media has publicly named major U.S. tech companies (Google, Microsoft, Palantir, IBM, Nvidia, Oracle) as potential military targets, warning that the regional conflict is expanding into digital infrastructure. Recent Iranian drone strikes have already damaged AWS data centers in the UAE and Bahrain, demonstrating practical capability to target critical cloud and data-center operations.

wired.me · Fricken· 22 days ago · 6 min · threat-intel · details · hn 79
0
techcrunch.com · elsewhen· 23 days ago · 5 min · security · details · hn 106
7
0

This article reverse-engineers the UniFi inform protocol (port 8080) and discovers that while the AES-128-CBC encrypted payload requires per-device keys, the first 40 bytes of every packet are unencrypted and contain the device's MAC address at bytes 8-13. This plaintext MAC allows routing of inform traffic without decryption, enabling multi-tenant UniFi controller hosting on shared infrastructure.

tamarack.cloud · baconomatic· 25 days ago · 7 min · research · details · hn 175
1
0

This article covers oil market volatility and geopolitical impacts during the Iran-Israel conflict, not a security or cybersecurity topic.

bbc.com · tartoran· 22 days ago · 6 min · news · details · hn 71
1
0

Autoresearch@home is a distributed collaborative platform where AI agents share GPU resources to collectively train and improve language models through iterative experimentation and knowledge sharing, extending Karpathy's autoresearch framework with a coordination layer.

ensue-network.ai · austinbaggio· 22 days ago · tool · details · hn 75
page 80 of 280
news.rdintel.com · api · mcp · rdintel.com