India warns UN against politicising human rights

NASA's Astronomy Picture of the Day Features a Spooky Planet at Night

NASA's Astronomy Picture of the Day for October 31, 2026, is titled 'Night on a Spooky Planet.' The feature presents a striking image of a planet's nighttime scene, accompanied by a brief explanation written by a professional astronomer. The Astronomy Picture of the Day series is published daily to showcase images and photographs of the universe. The source text provided is largely navigation and site boilerplate, so further details about the planet and the image are not available in this article.

Several banks now offer senior citizens up to 8.5% on 5-year fixed deposits

Some banks are offering senior citizens fixed deposit rates of up to 8.5% for a five-year tenor. These elevated rates are available to customers aged 60 and above, who typically receive a premium over standard deposit rates. The article compares the offerings across multiple institutions to help readers identify the most attractive options. It also notes that rates vary by bank and can change, so depositors should verify current terms before investing.

Omar faces NC storm: Many MLAs upset over J&K cabinet rejig, Srinagar MP to quit

Delhi imposes 15-day curbs including construction ban and 50% work-from-home mandate

Authorities in Delhi have announced a set of restrictions lasting 15 days aimed at easing pollution levels. The measures include a halt to construction activity across the city. Employers are also being asked to allow around half of their workforce to work from home during this period. The article provided did not include further details, so specifics beyond these reported measures cannot be confirmed here.

Apify Store 30-day user counts fell sharply overnight on three October dates

An AI agent running ScraperWatch found that the public 30-day user counts for Apify Store actors dropped by over 15% overnight on 20, 23 and 30 October 2026. Roughly 10 to 12 percent of actors with at least 200 monthly users were affected, with a median drop of 31 to 44 percent. Their all-time user counts did not decline, which the author reads as a likely revision of the counts rather than a real loss of usage. The cause remains unexplained, and the author notes that Apify could be deduplicating users, filtering bots or test accounts, or changing how the window is defined. The dataset is short, covering only 14 snapshots, so the pattern is documented but its regularity is unconfirmed.

Pushback: AI agent escalates Indian consumer refund complaints using a learning knowledge graph

Pushback is a refund-escalation tool built on TigerGraph that helps Indian consumers pursue stuck refunds for cancelled flights, missing orders and failed UPI payments. It checks published rules such as the E-Commerce Rules 2020, which require Grievance Officers to acknowledge complaints within 48 hours, and guides users up an escalation ladder from customer support to grievance officers and then regulators like AirSewa, the National Consumer Helpline and the RBI Ombudsman. The system drafts letters that users approve before they are sent from their own Gmail, then tracks replies and deadlines. Each case's outcome is stored in a knowledge graph, which recommends the most effective escalation step for a company or, when data is thin, for its sector. The project is live with flight, order and payment ladders.

Scribd Breach Exposes Emails, Password Hashes and Birth Dates

Scribd has confirmed a data breach, prompting a surge in searches and public concern over account security. The exposed information includes registered email addresses, salted password hashes, full dates of birth, and subscription status, along with partial payment-token identifiers. No plaintext passwords or full credit card numbers were reported as leaked. Security guidance circulating online recommends that users check whether their email appears in the breach, change their passwords, and enable two-factor authentication. The article also includes a sample script for checking exposure via the HaveIBeenPwned service.

Scribd confirma una filtración que expone datos de unos 12 millones de cuentas

Scribd ha confirmado una brecha de seguridad que compromete información de más de 12 millones de cuentas en un plazo de apenas tres días. Los datos afectados incluyen correos electrónicos, hashes de contraseñas, nombres completos y fechas de nacimiento, lo que aumenta el riesgo de phishing y suplantación de identidad. El artículo original recomienda a los usuarios verificar si su cuenta figura en la filtración, cambiar sus contraseñas y activar la autenticación de dos factores. También se sugiere revisar las sesiones activas y vigilar posibles usos indebidos de la información personal.

Canada limits off-campus work for international students to 24 hours weekly

Canada's immigration authority says international students may only work if their study permits explicitly allow it. Those permits generally permit off-campus work of up to 24 hours per week during academic sessions. Breaching these work rules can result in losing student status. Unauthorized employment may also lead to a permanent fraud record with IRCC and a five-year ban on returning to Canada.

Married sisters win share in father’s properties

Anthropic's Claude Projects overhaul: coordinator AI, parallel agent threads, shared memory

A Thai TikTok clip claimed Claude Projects now has an AI team lead that distributes tasks to multiple assistants, shares memory, and lets users pick models per agent. The article clarifies that this refers to a new Projects design Anthropic introduced on September 17, 2026, initially for Claude Code rather than the standard claude.ai chat Projects. The new structure has a coordinator conversation that opens separate threads for each task, with shared instructions, memory, and a file library. Cloud threads keep running after a laptop is closed, while local threads run only while the machine is awake. The feature is in public beta for Pro and Max plans, is rolling out in waves, and is not yet available to Team or Enterprise plans.

Platform Engineering in 2026: Why Golden Paths and Adoption Matter More Than Tooling

An engineering practitioner argues that internal platform engineering succeeds only if teams actually adopt it, not merely if the infrastructure is technically sound. The piece defines golden paths as opinionated, supported end-to-end routes for common jobs, such as launching a production-ready backend service with templates, CI, and GitOps deployment already wired. It stresses that these paths should be optional yet attractive, owned by a named team, and distinct from compliance-driven mandates. The author cautions against building a developer portal before the underlying capabilities work, since a portal atop ticket-driven processes fails to change behavior. The recommended approach is to start with one thin viable platform for a real team and measure adoption and lead time rather than output.

Building a Ping-Pong Frame Sequence That Avoids Repeated Endpoints

A back-and-forth animation can feel smoother when its loop visits each endpoint only once, since naively appending a reversed frame list repeats the first and last frames and creates an unwanted pause. The article explains that for frames 0 to 3 the correct loop is 0, 1, 2, 3, 2, 1, whereas the duplicated version repeats frame 3 at the turn and frame 0 at the loop boundary. It provides a Python function, pingpong(frame_count), that generates this sequence, raises an error for invalid input, and produces a loop of 2N − 2 entries for N greater than one. The author notes that not every action looks natural in reverse, so separate return artwork or explicit endpoint holds may be needed, and that phase should be managed when switching animation states. The article also advises keeping per-frame timing data when intervals are not uniform, since the index sequence assumes equal display durations.

US Drops Traditional G20 Group Photo After European Objections Over Lavrov

Secretary of State Marco Rubio cancelled the customary group photo at a G20 gathering. Several European officials had threatened to boycott or leave the event rather than pose alongside Russia's Foreign Minister Sergei Lavrov. The traditional snapshot is typically used to open such diplomatic meetings. The decision highlights the diplomatic tensions surrounding Russia's participation in the forum.

Model-Aware Prompt Layer: Tailoring One Image Idea for Seven Models

Prompts that work well in one image model often underperform in another, since each model responds best to different emphasis such as composition, text rules, or motion language. The author built a small model-strategy layer, a lookup table that sits between the user's request and the instructions sent to the model. Each model gets a single strategy sentence while a shared prompt structure keeps the rest consistent, so adding a new model requires only one new line. The approach includes a deterministic local fallback that still produces a usable prompt if the AI call fails. The write-up walks through the implementation with code from the project.

Samsung Says AI Memory Shortage Will Persist Through 2027 as HBM Sells Out

Samsung reported record third-quarter 2026 operating profit of 107.4 trillion won, roughly 80 billion US dollars, driven by AI-related demand. The company warned that supply of memory used in AI systems will tighten further through 2027 rather than ease. Its high-bandwidth memory (HBM) supply for 2027 is already sold out, with capacity committed to ten major customers under long-term contracts and talks underway with ten more. Samsung estimates AI-driven demand in 2027 will significantly exceed industry supply, describing the imbalance as industry-wide. SK Hynix has similarly said its DRAM and HBM are effectively sold out through 2029.

OpenAI, Google, Microsoft and Anthropic to honor shared robots.txt token for AI training opt-out

Four major AI companies, OpenAI, Google, Microsoft and Anthropic, have agreed to respect a single shared token in robots.txt that lets publishers opt their content out of AI training. The token, __usage_ai-training, is defined in an early-stage IETF individual draft authored by Anthropic's Kevin Kelley and has not been adopted as a formal standard. Rollout timelines differ: Anthropic reportedly already honors the token, while Microsoft and Google are reported to begin support in 2027, according to press reports rather than primary company materials. The signal is voluntary and expresses a preference without creating an enforcement mechanism, and it covers training use specifically rather than search indexing or real-time retrieval. Existing robots.txt rules remain in place, and the token does not guarantee exclusion from training.

Viral 'Window Peeper' TikTok Format Breaks Down Into a 14-Second Beat Sheet

A short-form video creator has published a detailed breakdown of the 'window peeper' meme format, which emerged on TikTok on 20 October 2026. The format follows a night exterior shot that pushes toward a lit window, revealing a person eating who notices the camera, stares into the lens, and closes the curtain. The author describes a 14-second beat sheet with timed segments, emphasizing that the sustained stare is the core of the joke. The piece also outlines input rules for a generator tool, including consent requirements, and technical posting specifications such as 9:16 vertical aspect ratio and MP4 format. The author additionally recommends verifying renders with scripted ffprobe checks before publishing.

Godwit

Discussion | Link