Dhanik.online Launches Free Mutual Fund Screener for Indian Investors

A new web platform called Dhanik.online has been launched to simplify mutual fund research for investors in India. The tool allows users to screen and compare over 14,000 mutual funds side by side, covering returns, risk metrics, and expense ratios. It also includes an XIRR calculator to measure actual returns from SIPs and a rolling returns analyzer to assess fund consistency across market cycles. Users can additionally track their entire mutual fund portfolio from a single dashboard. The platform is built using React, TypeScript, Tailwind CSS, Supabase, and is hosted on Netlify.

DeepInfra Offers LLM API Costs Up to 27x Lower Than OpenAI, With Caveats

DeepInfra is an open-source LLM inference platform offering pay-per-token pricing significantly below major providers like OpenAI and Anthropic, with models such as DeepSeek R1 priced at $0.55 per million input tokens versus OpenAI o1's $15. A week-long benchmark found the savings are real for high-volume tasks like classification and batch processing, where switching from GPT-4o-mini to Llama 3.1 8B can cut costs by 67%. However, the platform lacks support for certain features including structured output mode, function calling, and does not host Anthropic's Claude models. Users also face practical limitations such as a 30 requests-per-minute cap on the free tier, frequent model version changes requiring prompt re-tuning, and no per-feature cost tracking in the dashboard. DeepInfra offers a $5 free credit for new users, and self-hosting becomes cost-competitive only beyond approximately one billion tokens per month at high GPU utilization.

Courts can examine the correctness of answer keys: Delhi high court

CGHS officials get financial powers to cut claim delays

A Korean Developer's 1980s Journey: Tape Drives, Game Addiction, and Typed Code

A software developer recalls how a discontinued math competition in Busan, South Korea, was quietly converted into a computer class in 1985, giving him his first exposure to personal computers. He went on to represent Busan at a national computing contest in Seoul in 1986, where participants brought their own machines and sat four-hour exams. A Samsung computer he earned by qualifying for nationals became his gateway to cassette-loaded games and early programming. He developed patience and coding skills by manually typing source code from a computer magazine line by line. By middle school, floppy disks and role-playing games like Ultima had taken over, before exam pressures eventually steered him toward a formal computer science education in the 1990s.

A Farmer Arrested for Going 5 Seconds over His Time Limit at Data Center Meeting

Article URL: https://www.gadgetreview.com/arrest-him-the-moment-police-handcuffed-a-farmer-for-going-5-seconds-over-his-time-limit-at-data-center-meeting Comments URL: https://news.ycombinator.com/item?id=48701342 Points: 28 # Comments: 8

Why Redis Uses Approximate LRU Instead of a Perfect Eviction Algorithm

Redis does not maintain a perfectly sorted LRU structure because updating it on every read — even with a doubly-linked list — would add significant overhead at millions of requests per second. Instead, when memory is full, Redis randomly samples a small set of keys (default: five) and evicts the least recently used among them. To improve accuracy, Redis retains the best eviction candidates in an internal pool of 16 entries, merging new samples with existing candidates across iterations. This approach means eviction cost is only paid when memory is actually exhausted, keeping the far more frequent read operations fast. The result is a near-perfect approximation of true LRU that scales efficiently without sacrificing meaningful cache hit rate accuracy.

Dev Adds Self-Healing AI Loop to Open-Source Trust Language LOOM on Day 7

A solo developer building LOOM, an open-source language designed as a trust layer for AI-written code, marked a significant milestone on Day 7 of the project. The automated system that grows the language now self-improves, expanding LOOM's verifiable checks from 308 to 333, all passing. A key update introduced 'metered attestation,' bundling what foreign code can do, how often, and a non-AI authority signature into a single capability unit. After the automated system proposed a change that inadvertently weakened a security guarantee, the developer rejected it and patched the pipeline so proposals must pass both a self-check and adversarial self-test before reaching human review. The project, built solo from Ukraine and published under the MIT license, has now frozen its trust layer and is shifting focus toward a WebAssembly backend and real-world users.

Teen Developer Proposes PAL Standard to Let AI Agents Control Physical Hardware

A 19-year-old developer has proposed an open standard called PAL (Physical Agent Link) that enables AI agents to interact directly with real-world hardware such as GPIO pins, sensors, and I2C buses. The proposal argues that existing solutions, including Espressif's Chat Coding framework and Raspberry Pi-based setups, are either too resource-heavy, too rigid, or too slow for iterative agent use. PAL's core insight is that Python's REPL loop and an AI agent's reasoning loop are structurally identical, making a MicroPython REPL over WebSocket a natural and lightweight interface for hardware control. The architecture splits an ESP32-S3 into two cores — one running stable C/FreeRTOS drivers for hard real-time safety, and the other running a MicroPython environment where the agent freely executes generated code. The standard uses simple JSON messages to send and receive Python execution results, requiring no tool schemas or pre-registration of hardware components.

Northern and Central Europe Shatter Heat Records Amid Intense Heatwave

A powerful heatwave sweeping across northern and central Europe has broken all-time temperature records in multiple countries. Germany, Denmark, and the Czech Republic are among the nations registering their highest temperatures ever recorded. The extreme heat event marks a significant climate milestone for a region not traditionally associated with such intense heat. Authorities across affected countries are monitoring the situation as temperatures continue to climb to unprecedented levels.

7 India-bound cargo ships cross Hormuz in 3 days, 15 in queue

React vs. Angular in 2026: Which Framework Suits Enterprise Architecture Best

As both React and Angular reach full maturity in 2026, the choice between them is less about quality and more about organizational fit. Angular's opinionated, standardized structure makes it well-suited for large enterprises with multiple distributed teams, offering built-in tools that reduce reliance on third-party libraries and lower security risks. React, by contrast, offers a lightweight and flexible approach that empowers senior developers to build highly customized, performance-tuned applications, though it requires strong technical leadership to avoid inconsistent codebases. Experts recommend that teams prioritize cross-team consistency and long-term maintainability should lean toward Angular, while those needing rapid UI iteration with disciplined developers may benefit more from React. Ultimately, the decision should be driven by team size, security requirements, and how much architectural freedom an organization can responsibly manage.

US issues limited commemorative passports featuring Trump's image for 250th anniversary

The United States is releasing a special edition commemorative passport to mark the country's 250th anniversary. The limited passport features an image of President Donald Trump, making it an unusual addition to official travel documents. These passports will become available starting July 6, 2025. Unlike standard passports, applicants must apply for this special edition in person rather than through regular channels.

ContextVault 1.3 Offers Local-First Memory Engine to Unify AI Conversations and Sessions

ContextVault, a tool built by an independent developer, has reached version 1.3 and expanded from a simple browser conversation recorder into a local-first context engine for AI-assisted development. The tool addresses "context fragmentation," the problem of project knowledge becoming scattered across multiple AI platforms, coding agents, and terminal sessions. It operates across three layers: a Chrome extension that captures browser-based LLM conversations, a terminal integration, and a unified vault that makes all captured context searchable. All data is stored locally in IndexedDB with no backend server, no user account requirement, and no external data transmission. Conversations can be exported as Markdown files or ZIP archives, each tagged with metadata such as platform, model, and date.

Developer Builds Self-Hosted Scratchpad to Stop Losing Temporary Notes

Developer Sangeet Verma published a post on DEV Community on June 27 describing a self-hosted scratchpad tool he built. The project was created out of frustration with notes and jottings repeatedly disappearing. Verma developed the workspace using Go as the primary programming language. The tool is open source and falls under the web development category. The write-up is a roughly five-minute read tagged under selfhosted, Go, opensource, and webdev.

Venezuela hit by third major quake as death toll from earlier tremors tops 1,430

A 5.6-magnitude earthquake struck Venezuela's Aragua region, adding to the destruction caused by two earlier tremors that have killed more than 1,430 people. Thousands more have been injured and left homeless as a result of the successive quakes. Rescue operations remain active across heavily affected areas while authorities work to restore power and address critical supply shortages. International aid has begun arriving to support relief efforts on the ground. The United Nations has estimated significant economic losses stemming from the disaster.

DataBench Offers 25+ Developer Utility Tools in One Free Browser-Based App

A developer has launched DataBench, a free, browser-based workbench that consolidates over 25 commonly used developer tools into a single tab. The platform includes utilities such as a JSON formatter, regex tester, CSV-to-SQL generator, Base64 encoder, and PDF compressor, among others. All tools run entirely client-side, meaning no data is sent to external servers. DataBench requires no sign-up, no installation, and has no paywall, making it immediately accessible at databench.app. The project is actively under development, with the creator inviting user feedback to expand the tool set further.

Developer Open-Sources Solana Prediction Market Smart Contract Built with Anchor

A developer has released an open-source prediction market smart contract on Solana, built using the Anchor framework and available on GitHub. The contract supports binary YES/NO outcome markets with bonding curve trading in SOL tokens, along with platform and creator fees, Metaplex metadata, and whitelist functionality. An optional backend co-signing feature allows for controlled market launches, while whitelisted users benefit from reduced fee tiers. The project includes a full stack implementation covering on-chain Rust logic, a TypeScript backend with WebSocket support, MongoDB state management, and a responsive frontend. Developers can clone the repository, run it locally against a validator or devnet, and explore a live demo at takes.fun.

How AI Agents Store and Recall Memory Across Conversations

AI agents do not retain memory natively; instead, they rely on external systems that save key information and reinsert it into the prompt when a user returns. Short-term memory keeps a live session coherent, while long-term memory preserves preferences, names, project details, and past decisions across separate conversations. Effective memory systems are selective, storing only durable and relevant facts rather than entire chat histories to avoid noise. Retrieval quality is critical — the best systems rank stored memories by semantic relevance, recency, and importance to surface the most useful context. Common failure points include saving outdated or duplicate information, retrieving unrelated memories, and lacking mechanisms to correct or delete stored data.

Ahead of VB G RAM G coming into force on July 1, 19 states and UTs notify scheme