UPI Now Enables PF Withdrawals, But Experts Urge Caution Over Pension Risk

The Unified Payments Interface (UPI) has been integrated with the Employees' Provident Fund system, making it easier for members to withdraw their provident fund savings. While the move simplifies access to retirement funds, financial experts are cautioning against hasty withdrawals. A key concern is that premature withdrawal can disqualify members from lifelong pension benefits under the Employees' Pension Scheme. These pension benefits are designed to provide a steady monthly income after retirement, making them a critical component of long-term financial security. Experts advise members to carefully weigh short-term liquidity needs against the long-term value of retaining their pension eligibility.

Engineering Team Cuts LLM API Costs by 60% Using Caching and Token Monitoring

A software engineering team shared how they reduced their large language model API costs by 60% on production AI projects by systematically identifying and addressing cost drivers. They found that the bulk of expenses came from repetitive input tokens — such as repeated system prompts and retrieved documents — rather than output tokens. The team built middleware to log token counts and estimated costs for every LLM call, enabling data-driven decisions instead of guesswork. Their single biggest saving came from implementing semantic caching, which returns stored responses for queries that are similar in meaning rather than only identical in wording. The approach, documented with code examples for Django projects, prioritizes measuring usage first before attempting any optimization.

India Dismisses Pakistan's Allegations Amid Karachi Attack Fallout

India has rejected Pakistan's claims linking it to a recent attack in Karachi, calling them baseless and urging Pakistan to look inwards. Pakistan's information minister confirmed that the country carried out overnight airstrikes on eastern Afghanistan in retaliation for the deadly attack. The strikes were aimed at targeting terrorists believed to be responsible for the incident. The episode has heightened tensions in the region, with India firmly denying any involvement in the attack.

How the useDebounce Hook Fixes Common React Debouncing Mistakes

When users type in a search box, React components can fire an API request on every keystroke, generating redundant and stale calls. A common workaround is writing debounce logic manually with setTimeout inside components, but this approach introduces bugs like memory leaks on unmount, stale closures, and scattered duplicate code. The useDebounce hook from @reactuses/core addresses all three issues by wrapping lodash.debounce internally, handling edge cases like leading and trailing execution. It works by maintaining two separate values: a fast-updating one bound to the UI input, and a debounced one used to trigger side effects only after typing pauses. This pattern keeps the input responsive while reducing API calls to one per typing pause rather than one per keystroke.

Pakistan Ground Operation Near Afghan Border Kills 29

Pakistan has launched a ground military operation along its border with Afghanistan, resulting in 29 deaths. The operation targets militant groups active in the border region. It comes just one day after armed militants attacked the regional headquarters of the paramilitary Rangers in Karachi, Pakistan's southern port city. That attack, carried out with guns and explosives, killed three soldiers. The back-to-back incidents highlight the ongoing security challenges Pakistan faces from militant activity.

How to Build Autonomous AI Agents in Python Using Orchestration Patterns

Autonomous AI agents are software systems that can plan multi-step workflows, use external tools, self-correct, and retain memory without constant human intervention. At their core, these agents require a reasoning engine such as an LLM, tool access, memory systems, and a planning loop following an observe-think-act cycle. A developer guide published on DEV Community demonstrates how to implement a minimal working agent in Python using the ReAct pattern with OpenAI's GPT-4 API. The guide recommends starting with three to five well-defined tools and a narrow task before incrementally adding complexity. It also outlines practical use cases including code review bots, research assistants, and customer support agents, each with varying tool requirements and complexity levels.

Developer releases read-only TypeScript tool to monitor Stripe Connect balance risks

A developer has released stripe-connect-reckon, an open-source TypeScript library designed to monitor Stripe Connect accounts for financial risk signals. The tool was inspired by a real incident in which a connected marketplace account went negative, triggering silent payout suspensions and queued refunds that were only discovered through a customer complaint. The library detects issues such as negative balance risk, failed payouts, unreconciled refunds, and gaps in Stripe's event window, reporting each with a severity level. It is strictly read-only, calling only list and retrieve endpoints without writing or remediating anything, and is backed by 47 unit tests using fixtures with no network dependency. Released as v0.1.0 under the MIT license, the package is available on npm and GitHub, with the developer actively seeking feedback on additional signals worth monitoring.

smolagents Enables Python-Based AI Agents But Demands Clear Safety Boundaries

smolagents is an open-source Python library by Hugging Face that lets developers build AI agents in minimal code, with a key feature being 'CodeAgent', which expresses actions as executable Python rather than JSON or plain-text tool calls. This design allows agents to perform complex tasks involving loops, conditionals, and tool composition, but also raises the stakes if execution boundaries are not properly defined. The library integrates with a wide range of model providers, tool sources like MCP servers and LangChain, and optional sandboxed environments such as Docker, E2B, and Modal. Security experts and the Doramagic project both advise a staged onboarding approach: starting with no-tool agents, then adding read-only tools, and explicitly deciding the execution environment before granting real system access. The core safety question is not whether the package installs correctly, but whether the host environment, tool permissions, and sandbox policies are properly configured before deployment.

Has anyone used DeepSeek? Is it really good?

Seoul Developer Builds Self-Reinforcing K-pop Music Pipeline on OCI Free Tier

A Seoul-based backend developer has built k-cosmos, a web-based 3D music space that maps K-pop tracks using 768-dimensional vector embeddings, after finding no structured K-pop metadata or emotional tag datasets publicly available. The self-reinforcing data pipeline runs on Oracle Cloud's free tier and uses Spring Boot with pgvector to continuously enrich its own music database. To prevent database connection exhaustion, the developer split external API calls and embedding generation into three decoupled transaction phases, ensuring heavy network I/O occurs outside active database connections. A two-stage SQL window function enforces artist diversity in recommendations, preventing any single artist's large discography from dominating the suggestion space. Budget controls randomize and flatten the processing queue nightly to evenly distribute API quota usage and avoid hitting free-tier LLM limits.

India assistant coach admits 'disbelief' after shock loss to Ireland

India suffered a surprising defeat to Ireland, leaving the coaching staff in a state of disbelief. Assistant coach Ryan ten Doeschate acknowledged that the team was slow to adapt to the conditions they faced. The wind and other environmental factors played a significant role in unsettling India's game plan. Ten Doeschate credited Ireland for putting in a 'fantastic' performance to secure the win. The result has raised questions about India's ability to adjust quickly to unfamiliar playing conditions.

YRF makes first bet on vertical entertainment

Watch: Massive black smoke billows from Russian oil refinery after Ukrainian drone strike, 1 dead

Italy opens antitrust probe into Microsoft over AI-driven subscription upgrades

Italy's competition authority has launched a formal investigation into Microsoft over its handling of AI feature integrations in Microsoft 365 plans. The regulator alleges that subscribers were automatically moved to more expensive plans that bundled new AI tools, including Copilot and Designer. Authorities are concerned that customers were not adequately informed about these changes or the associated price increases. The watchdog believes this conduct may constitute an unfair commercial practice and could have limited consumer choice.

Iceland Cricket trolls Gambhir after India's loss

Putin Acknowledges Fuel Shortage After Ukraine Strikes Russian Energy Sites

Russian President Vladimir Putin has publicly admitted to a fuel shortage in Russia following Ukrainian strikes on energy infrastructure. Ukraine has carried out attacks targeting Russian fuel and energy facilities, which Kyiv describes as justified retaliation. Ukrainian officials argue the strikes are a direct response to Russia's near-daily bombardment of Ukrainian civilians and energy infrastructure. The conflict, which began with Russia's full-scale offensive in February 2022, has seen both sides targeting each other's critical infrastructure.

Venezuela Earthquake Death Toll Approaches 1,500 as Humanitarian Crisis Deepens

Venezuela has been struck by twin earthquakes in what is being described as one of Latin America's most devastating seismic disasters. The death toll is approaching 1,500, with rescue hopes diminishing as time passes. Millions of survivors are feared to be without access to sanitation and other basic necessities. The scale of the disaster has triggered a widespread humanitarian emergency across the affected region.

Temporary employees can get pension if eligible: SC

Python's itertools Module: Five Functions to Build Efficient Data Pipelines

Python's standard library includes the itertools module, a collection of iterator-building tools designed to help developers write faster, more memory-efficient code. All itertools functions are implemented in C and use lazy evaluation, meaning values are generated on demand rather than loaded into memory all at once. A DEV Community tutorial highlights five high-impact functions — count(), cycle(), chain(), and others — each demonstrated with practical, real-world examples. Functions like cycle() simplify round-robin server load balancing, while count() automates row indexing without requiring a separate counter variable. The module is particularly valuable when processing large datasets or building data pipelines where memory efficiency is critical.

How CLAUDE.md and AGENTS.md Rule Files Keep AI Agents Consistent in FastAPI Projects

AI coding agents like Claude Code and Cursor automatically read project-level rule files such as CLAUDE.md, .cursorrules, and AGENTS.md as system-level instructions, allowing teams to enforce consistent code style across a repository. In FastAPI projects, these files address common inconsistencies like varying endpoint structures and mismatched dependency injection patterns. A practical rule file covers the tech stack, router design conventions, dependency injection standards, prohibited patterns, testing setup, and error-handling approaches. Experts recommend keeping each rule file under 50 lines and framing rules as positive patterns rather than prohibitions, since agents respond better to instructed structures than to restrictions. Rule file changes should be version-controlled and included in pull request reviews, treating them with the same rigor as infrastructure-as-code configurations.