Deposition records cited by NPR show TikTok employees repeatedly warned executives about child safety failures, including minors appearing in Live streams where viewers could pay for 'gifts,' and said the company was aware of and profiting from such content. Staff also flagged a signup loophole letting minors bypass age checks via Facebook, Instagram or Google accounts, and a hidden link allowing teens to disable parental 'family pairing' controls. The documents emerged from ongoing lawsuits against TikTok over its handling of child safety; TikTok did not provide NPR an on-the-record response.
engadget.com
· 2026-10-20
A technical explainer argues that speculative decoding does not make large language model inference faster by cutting the number of floating-point operations performed. In fact, total compute (FLOPs) for a given sequence increases under the technique, since a small draft model proposes tokens that a larger model must then verify in a single parallel forward pass.
jbarrow.ai
· 2026-10-19
The visual tells that worked two years ago are disappearing. What still works is checking where a picture came from rather than how it looks.
gokawiil.com
· 2026-10-19
A federal judge issued a preliminary injunction halting enforcement of Utah's SB 73, which required adult websites to detect users' physical locations and block or age-verify anyone using a VPN. The court found the law likely violates the Constitution's restrictions on states burdening people and businesses outside their borders, since it would force sites to verify every visitor worldwide. EFF, which opposed the law, called the ruling a win for digital privacy rights.
yro.slashdot.org
· 2026-10-17
The IEEE Computer Society published commentary arguing that cloud security architectures need to go beyond blocking unauthorized access. It states that systems must also be able to clearly demonstrate how access decisions were made, in order to satisfy regulators and investigators.
wpcms.computer.org
· 2026-10-15
A verified Instagram fan account with over a million followers posted an AI-generated photo of actress Mia Goth wearing a shirt mocking her ex Shia LaBeouf, shortly after her divorce filing became public. The image gained nearly 500,000 likes before being identified as fake, and Google's AI Overview repeated the fabricated narrative as fact when users searched for the story.
theverge.com
· 2026-10-15
Google has quietly changed Android behavior so a device's phone number is automatically verified and linked to every Google account signed in on it, including secondary, work, and throwaway accounts. Users discovered this after receiving unsolicited emails from a Google phone-verification address, and the change appears to have been made to Google's support documentation between August and October 2025. The setting can be disabled via Settings > Google > All Services > Privacy & Security > Phone number verification.
tech.slashdot.org
· 2026-10-14
Android includes a phone number verification setting that is enabled by default, which associates a user's phone number with every Google account signed into that device. Users must manually opt out if they do not want their number linked across accounts.
androidauthority.com
· 2026-10-14
OpenAI published 722 manuscripts and proofs addressing longstanding mathematical problems, generated by an unnamed internal frontier AI model, with only 300 results formalized for computer verification. Researchers who reviewed the work told The Verge they found nonsensical explanations, malformed formulas, and results that duplicated existing research, while OpenAI withheld the model's identity and the prompts used to generate the findings.
tomshardware.com
· 2026-10-13
A bug reported to Rootshell revealed that ICQ's authentication servers could be accessed using any account's UIN paired with a 9-character password, bypassing the real password entirely. The developer of a Linux ICQ client discovered the flaw after a user mistyped a password but was still logged in, tracing it to a buffer overflow in Mirabilis's server-side password check. Rootshell confirmed the bug and published exploit details, including sample code and a 'password' like '123456789' that overruns the 8-character buffer.
opennet.ru
· 2026-10-13
Eurydice is a new tool that converts Rust programs into C code while preserving the original code's structure, rather than optimizing or obscuring it. Started in 2023, it is part of the Aeneas formal-verification project maintained by researchers at Inria and Microsoft, and has already been used to translate post-quantum-cryptography routines from Rust into C.
lwn.net
· 2026-10-09
OpenAI revised its public GitHub repository of formalized mathematical results, adding six new Lean formalizations and modifying 19 existing ones. It also withdrew three results previously listed as formalized. The company said roughly 42% of its top-line results are now formalized and that it will keep updating the repo as it finds errors.
twitter.com
· 2026-10-08
GoKawiil