Ventures Platform raises $83M second fund to expand beyond Nigeria

Pan-African venture firm Ventures Platform has closed its second fund at $83 million, marking a significant increase from its previous raise. The firm, which is headquartered in Nigeria, plans to broaden its investment scope across the African continent. The larger fund signals growing investor confidence in African startup ecosystems. With this capital, Ventures Platform aims to back startups in markets beyond its traditional Nigerian focus.

Maharashtra FDA Chief Tukaram Mundhe Takes Hard Line on Food Adulteration

Tukaram Mundhe, serving as Maharashtra's Food and Drug Administration (FDA) Commissioner, has launched a wide-ranging crackdown on food safety violations across the state. His targets include food adulterators, well-known restaurants and clubs, quick commerce dark stores, and school and government canteens. Mundhe is also going after those who engage in surrogate advertising of banned products. Known for his uncompromising stance, he has faced around 25 transfers throughout his career but has consistently refused to soften his approach. His tenure signals a stricter era of food safety enforcement in Maharashtra.

US Singer Criticizes NYC Event Featuring RSS Chief Mohan Bhagwat

RSS chief Mohan Bhagwat is scheduled to speak at a Manhattan event on Saturday, August 29. The gathering is being promoted as the 'Universal Oneness Celebration,' aimed at bringing together the Hindu American community. An American singer has publicly criticized the event, directing remarks at Zohran Mamdani in connection with it. The event is being held in New York City and has drawn attention amid the controversy surrounding the criticism.

How to Build an AI Observability Platform to Prove Test Automation ROI

Engineering teams using AI-driven test automation are increasingly being asked by stakeholders to demonstrate measurable proof of value beyond simple productivity estimates. An AI Engineering Observability Platform addresses this by tracking key metrics such as agent utilization, token consumption, daily output, time savings, and quality improvements across multiple specialized agents. The proposed architecture uses a telemetry layer to log LLM executions and route data to visualization tools like Power BI, enabling real-time dashboards for leadership. Eight core KPIs — including AI adoption rate, cost per story, hours saved, and defect reduction — are recommended when presenting results to steering committees. A sample executive summary illustrates the approach: one month of AI automation activity generated 3,800 assets, consumed 42 million tokens, and reduced manual engineering effort by 78%.

India’s ₹7L cr market shifts as gold prices shoot

A Unified KPI Framework for Measuring Playwright Test Automation Value

A comprehensive KPI framework has been proposed for teams using Playwright and JavaScript to measure the true impact of test automation beyond basic pass/fail results. The framework organizes metrics across five key areas: execution speed, suite stability, test coverage, maintenance cost, and CI/CD integration. Key targets include a 50% reduction in runtime, over 90% of failures attributed to genuine application defects, and 80% or more regression coverage across functional scenarios. It also tracks automation ROI by comparing manual effort saved against automation costs, with a goal of achieving positive returns within three to six months. Playwright-specific capabilities such as parallel execution, auto-waiting, and built-in trace tools are highlighted as enablers for meeting these benchmarks.

Developer Builds Privacy-First Client-Side N-gram Analyzer for Text Auditing

A developer has created a lightweight, browser-based N-gram Analyzer designed to help writers and content editors identify repetitive phrases and evaluate vocabulary diversity in their text. The tool runs entirely on the client side, meaning user input is never sent to external servers, addressing privacy concerns around sensitive or unpublished documents. N-grams are contiguous word sequences of varying lengths — from single words (unigrams) to four-word phrases (4-grams) — and analyzing their frequency helps assess a document's thematic focus and writing style. The JavaScript utility cleans input text, segments it into words with Unicode support for accented characters, and uses a sliding window algorithm to count phrase occurrences. The project was shared on DEV Community as a technical walkthrough of the implementation logic and the case for local browser processing.

Developer builds open sandbox to test AI shopping agents' checkout flows

A developer has built and released acp-sandbox, a hosted mock merchant tool designed to help engineers test AI shopping agents against the emerging ACP checkout protocol without risking real transactions. The tool addresses a gap in the current testing ecosystem, where Stripe's test mode covers payments but no hosted fake merchant existed to verify the protocol layer of agentic commerce. The sandbox implements the ACP checkout session lifecycle — including create, retrieve, update, complete, and cancel actions — based directly on the April 2026 specification. It logs all requests and responses per API key, allowing developers to diagnose exactly where an agent's behavior deviates from expected protocol. The project is intentionally scoped to ACP only for now, with the developer open to extending support to Google's AP2 and UCP protocols if demand arises.

US Army Left Without Chief of Staff as Hegseth Shows No Rush to Fill Role

The United States Army has been without a chief of staff since Defense Secretary Pete Hegseth dismissed General Randy George from the position in April. No replacement has been named, and officials indicate there is currently no sign that Hegseth is close to making an appointment. The vacancy at the top of the Army's leadership structure has raised concerns about a growing command gap within the service branch. The situation is part of a broader leadership shake-up initiated by Hegseth, who has removed several senior generals since taking office.

How a 1906 Land Purchase Became Delaware's 1,359-Acre National Park

In 1906, William Poole Bancroft started acquiring land in Delaware's Brandywine Valley with the intention of preserving it for public health and well-being. Over time, he accumulated more than 1,300 acres in the region. Portions of this land were subsequently donated to the National Park Service. The site now spans 1,359 acres and serves as the largest unit of First State National Historical Park.

How to Build an Automated QA KPI Dashboard for Playwright and BDD Pipelines

Manual tracking of test automation metrics often produces outdated data and obscures engineering gaps, prompting teams to adopt automated reporting from test suites like Playwright and Cucumber. A structured KPI dashboard can surface key indicators such as pass rates, execution time, flaky test rates, and CI/CD pipeline health in real time. A sample setup with 120 tests — 94 automated and 26 manual — showed a 92% pass rate, 1.5% flakiness, and a 98% CI/CD success rate as of late May 2026. Trend data across six consecutive runs revealed steady improvement in pass rates and execution speed, while dependency resolution errors were identified as the primary cause of pipeline failures. Recommended next steps include pushing automated coverage above 90%, refactoring flaky scenarios, resolving dependency caching issues, and cutting suite runtime below 25 minutes through parallel execution.

‘Achche din’ wreck common man’s budget: Cong

Dolly Parton's final wish revealed

How Codex Manages Its Own Memory: Models, Rust Code, and Autonomous Writes

OpenAI's Codex coding assistant features a cross-session memory subsystem that can autonomously add, modify, merge, and delete persistent memories without requiring user approval for each action. The system operates as a two-model background pipeline: one model extracts reusable information from coding sessions, while a second consolidates that data into a file-based memory workspace. Lifecycle decisions such as scheduling, candidate selection, and pruning are handled by deterministic Rust and SQLite code, while the language models handle semantic judgments about what information is useful or outdated. Forgetting is not a single operation but a set of independent mechanisms, including thread deletion, consolidation-driven removal, and a full memory reset. The analysis is based on Codex commit 8444cf63b50a8a88521e0d2970d49f659b48eac7, reviewed on August 25, 2026; the feature is implemented but remains off by default.

Developer Redesigns Portfolio With Dual Modes, Weather Effects, and DEV Article View Counts

A software engineer has revamped their personal portfolio at a-thedeveloper.vercel.app, replacing a previously formal design with one that reflects more of their personality. The updated site features two toggleable modes — a professional version and an unfiltered one — along with weather-themed visual effects and matching ambient music. The developer also integrated the DEV.to API to display article view counts alongside publicly listed posts, going beyond the platform's default public URL. Originally built in 2023, just two years into their professional career, the portfolio has undergone multiple redesigns over the years. The latest update was shared on DEV Community as a personal project log rather than a technical tutorial.

Developer Spends a Year Building With AI, Finds Gains and Hard Limits

A developer in his fifties spent 2024 using AI coding tools — primarily Claude and GPT — to build an automated trading system and a workplace data-entry automation tool, despite having little prior coding experience. He found that switching between AI models mid-project caused significant disruption, as each model preferred rewriting existing code rather than editing it. The data-entry automation eventually succeeded but left behind a single, unwieldy file of over ten thousand lines of code. Throughout the year, he encountered recurring issues with context loss in longer sessions, usage quota limits, and the absence of documentation practices. He concluded that AI coding tools have real utility for beginners but come with structural limitations that are rarely discussed openly.

Dev builds plugins for Obsidian, VS Code, and Figma — and finds each a walled garden

A developer built text and image integrations for three desktop apps — Obsidian, VS Code, and Figma — and documented the hidden complexities each platform imposes on plugin authors. Figma proved the most restrictive, requiring a multi-step publish flow: create a draft post, obtain a temporary upload URL, PUT the image bytes, confirm media, then finalize the post. Because Figma separates the document-access thread from the UI thread, raw PNG bytes had to be serialized and passed via postMessage before any network call could be made. The plugin UI runs in a sandboxed iframe with a null origin, meaning the backend API had to be specifically configured to accept such requests or the code would silently fail at the last step. The developer's key takeaway is that plugin development means operating entirely within another app's rules — even a basic HTTP request must be carefully earned rather than assumed.

Developer argues early Claude Code was better due to its focused approach

A developer named Alex Kras has published a blog post expressing nostalgia for an earlier version of Anthropic's Claude Code tool. The post, titled 'Focus Is the Main Feature,' argues that the older iteration of the AI coding assistant was more effective due to its simplicity and focused functionality. Kras suggests that subsequent updates have added complexity at the expense of the core feature that made the tool valuable. The article gained traction on Hacker News, reflecting a broader sentiment among some developers about feature bloat in AI tools.

Developer Builds Browser-Only Base64 Tool to Avoid Server-Side Privacy Risks

A developer has created a client-side Base64 encoder and decoder that processes all data locally in the browser, eliminating the privacy risks associated with third-party online tools. The tool was built using a combination of AI assistance and vanilla JavaScript, a workflow the author describes as 'vibe coding.' A core technical challenge addressed was JavaScript's native btoa() and atob() functions failing on Unicode characters such as emojis and non-Latin scripts. The solution uses the modern TextEncoder and TextDecoder APIs to convert strings into UTF-8 byte arrays before encoding or decoding. The finished utility features a split-panel interface with one-click clipboard copying and sends no user data to any server.

Proxmox Cluster: Fixing Corosync Redundant Links and a Hidden Config Bug

A Proxmox VE cluster engineer added a second Corosync network link to establish true redundancy, connecting both the management network and an isolated Corosync-net interface. A misconfiguration using the wrong config key 'priority' instead of 'knet_link_priority' caused the lower-priority link to silently take over traffic, going unnoticed until log analysis revealed the issue. The correct key was identified from the official corosync.conf man page, and once applied, both links were properly indexed and the intended high-priority link resumed control. A failover test confirmed the setup worked correctly — cutting the primary link caused an immediate, seamless switchover to the backup with no loss of quorum.