Raven, a product of Graylark Technologies formerly known as GeoSpy, is an AI tool that pinpoints the exact property where an interior photo was taken by matching unique room features to millions of images scraped from listing sites and other online sources. Demonstrations show a single washer/dryer photo leading to an Airbnb listing in Charlotte and a bathroom image yielding a Vrbo result in Montecito; other demos return Zillow photos and hotel matches. The system does not perform reverse image search but instead recognizes spatial and stylistic cues learned from vast geotagged image datasets, and Graylark runs city-specific models - one custom model was built for Miami‑Dade. Law enforcement agencies, including the Los Angeles Police Department and Miami‑Dade Sheriff’s Office, have purchased access.
The capability has triggered privacy, legal, and misuse concerns: Airbnb and Zillow insist scraping violates their terms and say they use technical safeguards, and short‑term rental platforms restrict precise location data for listings. Graylark tightened access to government and enterprise users after earlier public availability and reports of misuse. Officials at Miami‑Dade describe use as limited and exploratory with no major case breakthroughs attributable solely to Raven. Graylark’s CEO says interior geolocation emerged during training, and the company plans further investment in mapping interiors of hotels, rentals, and homes - raising risks for stalking, surveillance, and exploitation investigations.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
A bug has been identified in Rust's borrow checker where struct's lifetime requirements can be bypassed through covariant field assignment. This issue could potentially lead to unsafe memory access in Rust programs. (github.com)
CFTC staff were instructed to retract an analysis they conducted on suspicious oil trades involving Iran. The directive raises questions about regulatory transparency and oversight in commodities trading. (bloomberg.com)
Crypto thieves sell modified Ledger Nano X wallets with hidden eSIMs that intercept data, leading to around 100 million dollars in stolen cryptocurrency. Attackers insert hardware to secretly read seed phrases without tampering with the secure chip itself. (heise.de)
Social media engagement for Reform UK figures has declined following Ukrainian strikes on Russian data centres, which significantly reduced bot activity linked to Russian information operations. This decrease in automated activity has led to lower interaction levels on Reform UK posts in the UK and across Europe. (thelondoneconomic.com)
Google releases a second edition of its SRE book, reflecting a decade of lessons and industry changes. The new edition addresses modern challenges like AI workloads, complex dependencies, and automation in site reliability practices. (cloud.google.com)
Michael Smith was sentenced to 18 months for using AI and thousands of bots to flood music streaming platforms with fake plays, earning him over $8 million in royalties. This case is the first DOJ prosecution involving AI-assisted music streaming fraud. (nytimes.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.