SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Ubuntu 26.04's sudo-rs leaves ghost processes when SSH drops during sudo su

Ubuntu 26.04 replaced the traditional sudo with sudo-rs, a Rust-based implementation, which handles SIGHUP signals differently when SSH connections drop. A developer discovered that using 'sudo su' over SSH on Ubuntu 26.04 left orphaned processes running indefinitely after the connection closed, reparented to PID 1. Investigation revealed that su blocks SIGHUP, and unlike traditional sudo, sudo-rs lacks cleanup logic to terminate child processes when the terminal disconnects. Switching back to the traditional sudo implementation resolved the issue, confirming sudo-rs is the key difference. Users are advised to use 'sudo -i' or 'sudo -s' instead of 'sudo su', as these alternatives exit cleanly when an SSH session ends.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Telegram Faces Critical Exploits, Mass Data Leak, and Founder's Arrest in 2025-26

Telegram founder Pavel Durov was arrested in France in August 2025 on allegations linked to criminal misuse of his platform, drawing global attention to the app's security shortcomings. Two serious unpatched vulnerabilities have emerged: a zero-click animated sticker exploit enabling remote code execution on Android and Linux, and EvilLoader, a zero-day flaw disguising malicious APKs as video files in chats. Cybercriminals are actively selling EvilLoader exploit kits on underground forums, leaving millions of Telegram for Android users exposed. In January 2026, a 44GB dataset containing over 200 million records surfaced on BreachForums, compiled from three separate databases. Security experts urge users to update the app immediately, avoid opening unsolicited files, and restrict incoming messages to trusted contacts.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer releases 'your-toast', a lightweight React and Next.js notification library

A developer has released your-toast, an open-source toast notification library designed for React and Next.js applications. The library features a glass-style UI and a minimal API, allowing developers to trigger success, error, warning, info, and loading notifications with a single function call. It includes a toast.promise() method that automatically handles loading, success, and error states tied to async operations. The library is architected around a central store and a provider component, removing the need for developers to manage notification state manually across components. A dedicated client-side entry point for Next.js App Router ensures the root layout can remain a Server Component without requiring a full client-side conversion.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

VRM Add-on Breaks on Blender 5.2 LTS and 5.3 Alpha Due to Exclusive Version Cap

Users upgrading to Blender 5.2 LTS or 5.3 alpha have found the VRM Addon for Blender failing to enable, with testing conducted on macOS Apple Silicon on September 22, 2026. The root cause is the blender_version_max field in the extension manifest, which marks the first unsupported Blender version rather than the last supported one — an exclusive upper bound. Add-on version 4.4.0 was the first release to declare a blender_version_max of 5.3.0, making it the minimum version that works on Blender 5.2 LTS; earlier extension versions such as 4.3.0 cap out at 5.1.x. A workaround exists for versions 4.3.0 and newer: installing the legacy add-on zip bypasses the version gate entirely, since it carries no manifest version restrictions. Version 3.9.0 fails on all paths under Blender 5.x because Blender 5.0 removed an API that the add-on imports.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How a misconfigured 1.6 GB VPS froze for hours due to npm ci memory exhaustion

A developer's low-spec VPS running a static sleep-cycle calculator crashed twice in September — first for 10 hours on the 18th, then for 4 hours two days later — due to memory exhaustion triggered by running npm ci during deployments. Kernel logs confirmed the OOM killer terminated the npm process, which had consumed around 600 MB of real RAM on a machine with only 1,612 MB available and zero swap space configured. The combination of no swap and a vm.swappiness setting of 0 left the kernel no fallback mechanism, causing prolonged thrashing rather than a quick process kill. The developer resolved the issue by adding a 4 GB swapfile, adjusting vm.swappiness to 10, and installing earlyoom to trigger faster OOM kills before the system fully freezes. The post-mortem highlights that disabling swap entirely on memory-constrained servers can be more dangerous than the performance overhead it aims to avoid.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Spring Boot Explained: Key Features, Advantages, and Core Java Concepts

Spring Boot is a Java framework built on top of Spring that simplifies and accelerates application development by offering auto-configuration, embedded servers like Tomcat, and starter dependencies. Unlike traditional Spring, Spring Boot eliminates much of the boilerplate code and manual setup, making it well-suited for building web applications, REST APIs, and microservices. One of its key advantages over Spring is that applications can run independently without configuring a separate server. Java 8's Optional class complements such development by providing a safer way to handle null values, reducing the risk of NullPointerExceptions through methods like isPresent(). HTTP status codes also play a central role in these applications, as they are numeric responses from a server that inform clients whether a request succeeded, requires action, or resulted in an error.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

VR Game AmblyoPunch Uses Neuroplasticity Research to Treat Lazy Eye

Seven Sportz is developing AmblyoPunch, a Meta Quest VR game designed to treat amblyopia, commonly known as lazy eye, using dichoptic training techniques. The game's design draws on recent research showing adult brains can actively reconstruct damaged neural networks when exposed to sustained, structured demand. Players complete 15-minute daily sessions across four progressive stages, with streak tracking meant to reinforce consistent engagement over high-intensity bursts. Technical features include per-eye camera rendering and GPU-based Gabor-pattern shaders, keeping frame times under 11 ms on Quest 2 hardware. Session data such as fusion accuracy and stage progression is stored locally with an optional opt-in for research partners, with no personally identifiable information collected.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Git Recovery Guide: How to Undo Almost Any Mistake Without Losing Work

Git almost never permanently deletes work — lost commits are typically orphaned and still stored in Git's object database until garbage collection removes them. Two key safety nets, the reflog and git fsck, allow developers to locate and restore commit hashes even after accidental resets, deleted branches, or botched rebases. Commands like git restore, git reset --soft, and git commit --amend handle most local undo scenarios, while git revert is the safer option for reversing commits already pushed to shared branches. The critical distinction is that reset rewrites history and suits local changes, whereas revert creates a new reversing commit that keeps shared history intact. Knowing which recovery command matches each situation can save hours of work, especially since uncommitted changes discarded via git restore cannot be recovered through the reflog.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Safety Benchmarks Can Fail When the Detector Itself Is Flawed

Developers at AgentSafeLabs discovered a critical flaw in their open-source AI security evaluation framework where the safety detector, not the language model, was producing incorrect results. In standard LLM safety testing, a classifier judges whether a model's response to an adversarial prompt is a refusal, compliance, or ambiguous — and those judgments feed directly into safety reports. During an investigation into apparently inconsistent model behavior, the team found the problem originated in the detector itself, including Unicode normalization failures. The most concerning issue was that a false PASS verdict goes unnoticed, unlike visible signs of uncertainty. The findings raise broader questions about how developers validate the classifiers they rely on to assess LLM safety.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Integrated Web Engine Claims 5,000 Dynamic Sites on One Small VPS Under 1ms Render

A developer experiment demonstrated that a single 2-vCPU, 4-GB VPS could host 5,000 independently managed dynamic websites using a tightly integrated web architecture called Kooboo. The system also achieved sub-millisecond rendering for a dynamic page containing ten blog items on a prepared server-render path. Unlike conventional stacks that chain separate services — HTTP servers, frameworks, databases, caches, and template engines — Kooboo combines the web server, database engine, and render engine into a single unified system. This eliminates process-boundary crossings and repeated data serialization that typically add latency at each layer. The two benchmarks were conducted separately under controlled conditions, and the raw data along with reproduction materials have been made publicly available.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Tool Retrieval Fails 95% of the Time When Users Paraphrase Requests

A structured evaluation of deferred AI tool loading found that retrieval works nearly perfectly when users mirror the exact vocabulary in tool descriptions, but collapses to around 5% recall when they paraphrase naturally. The experiment tested BM25 retrieval across 100 synthetic enterprise tools at three description detail levels — terse, realistic, and verbose — using 200 tasks split equally between vocabulary-matching and paraphrase queries. Verbose descriptions, which cost roughly double the tokens of realistic ones, showed no meaningful improvement in paraphrase recall, exposing that adding more words drawn from the same vocabulary provides no extra retrieval surface. Widening the shortlist from 5 to 10 tools only pushed paraphrase recall from 5% to 10%, doubling context costs for minimal gain. The core finding is that a vocabulary gap between how tools are described and how users actually speak creates a near-invisible failure mode, where the model receives a plausible but wrong shortlist and silently improvises.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How LINQ's Select Operator Cuts Unnecessary Data Load in .NET Apps

A common performance issue in .NET applications involves Entity Framework loading all columns of a database entity even when only a few are needed by the UI. Using LINQ's Select operator, developers can project only the required columns, reducing data transfer, memory usage, and query execution time. For reusable projections, C# Data Transfer Objects (DTOs) — especially using the concise record syntax introduced in C# 9 — are recommended over anonymous types. Select can also handle computed fields, nested related entities, and conditional logic that translates to SQL CASE WHEN statements. The SelectMany operator further extends this by flattening collections across related entities into a single efficient query.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

STON.fi Omniston Widget Offers Standalone TON Connect Mode for Lightweight dApps

STON.fi's Omniston Widget supports a standalone TON Connect configuration that lets developers embed a swap widget with built-in wallet connectivity, without setting up a separate TON Connect instance. In this mode, the developer provides only a TON Connect manifest URL and a DOM mount point, while the widget handles all internal wallet connection logic. The setup suits landing pages, dashboards, and lightweight dApps where the swap widget is the primary feature requiring wallet access. However, standalone mode is not appropriate for applications that already manage their own TON Connect instance; those should use the integrated mode instead. Developers must host a publicly accessible TON Connect manifest file on the same domain as their application, following TON Connect requirements for HTTPS, open CORS policy, and raster image formats for the app icon.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why PostgreSQL with pgvector beats dedicated vector databases for most AI apps

A growing number of engineers argue that specialized vector databases like Pinecone are unnecessary for the majority of AI applications, with PostgreSQL's pgvector extension offering a capable alternative. The debate gained prominence during the 2023–2024 generative AI boom, when startups raised hundreds of millions to build dedicated vector search engines on the assumption that relational databases could not keep pace. Critics of dedicated vector databases highlight four key risks: dual data sources, broken ACID transactional consistency, added network latency, and a fragmented security model. By contrast, storing vectors directly in PostgreSQL via pgvector allows atomic transactions, simpler queries, and zero additional infrastructure cost for datasets in the tens of thousands. Proponents suggest that only applications requiring billions of vectors or highly specialized ANN performance genuinely justify the operational complexity of a standalone vector database.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Rebuilds Home Server Using GitOps to Ensure Full Reproducibility

A developer has begun a blog series documenting a complete rebuild of their home server after realising their existing Ubuntu setup had grown too organically to be reliably restored in the event of a disk failure. The core principle of the new setup is that a Git repository serves as the single source of truth, with nothing permitted to run unless it is defined there. The new infrastructure is built on Proxmox as a hypervisor, with Komodo handling deployments by watching the repository and applying compose stacks automatically. Secrets are the only element kept outside the repository. Future posts in the series will cover hardware choices, the GitOps workflow with Komodo, and decisions the author had to revise along the way.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Merges 256 PRs to Harden Open-Source AI Office Suite GenOffice

A contributor submitted and merged 256 pull requests (PR #187 through #689) to GenOffice, an open-source AI-powered office suite supporting .docx, .xlsx, .pptx, PDF, Markdown, and HTML formats on macOS, Windows, and Linux. The work, backed by roughly 268 local commits, focused almost entirely on bug fixes and regression tests rather than new features. Key areas addressed included malformed XML parsing in DOCX and PPTX files, broken CSV encoding edge cases, and non-finite number inputs that could crash renderers or corrupt exports. Security-related hardening covered path traversal rejection, SSRF redirect limits, zip-handle validation, and sheet-address sanitization. Accessibility and internationalization improvements were also included, spanning bidirectional text, CJK font mapping, keyboard-operable dialogs, and localized ARIA labels.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds interactive cryptid archive on Sanity with live map and creature dossiers

A developer has created Apex Bestiary, a web-based archive of cryptid sightings built as a submission for the Sanity developer challenge. The project displays reported sightings as animated beacons on an interactive world map, where each marker opens a classified dossier with creature illustrations, threat levels, witness accounts, and folklore origins. The platform uses a structured three-document content model in Sanity — separating Creatures, Sightings, and Regions — with real-time listeners that update the map without requiring a page refresh. The frontend is built with Astro and React, uses Leaflet with OpenStreetMap tiles for mapping, and is deployed on Vercel, with all content managed through Sanity Studio and queried via GROQ. The project is publicly accessible without login, and source code has been published on GitHub.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

One Developer Claims AI Agents Now Match a 10-Person Engineering Team

A software developer writing on DEV Community describes running a personal AI agent setup where a coordinator delegates tasks to specialized agents for development, QA, and code review. The workflow allows features to move from roadmap to tested branch with minimal human intervention, with the developer only stepping in for final review and merging. Writing in early 2026, the author claims they can now handle the development workload that would previously have required around ten engineers, given sufficient access to compute tokens. The developer argues that human qualities such as taste, accountability, and defining goals remain the hardest things to automate and therefore the most valuable. They also highlight a practical use case involving AI-assisted data analysis of elderly home care business potential in rural Finland, noting that dynamic dashboards can be generated on the fly from a well-framed research question.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Coding Tools Are Overwhelming CI Pipelines, Demanding Architectural Fixes

The rapid adoption of LLM-assisted coding tools has created a mismatch between how fast code is generated and how fast it can be validated by Continuous Integration systems. AI agents can produce thousands of lines of code in under a minute, while typical CI pipelines take 15 or more minutes to process each batch, leaving developers and agents idle. AI-generated code introduces specific problems including duplicated logic, hallucinated dependencies, and non-deterministic outputs that cause flaky tests and unreliable feedback loops. When autonomous agents use flawed CI results as signals to self-correct, they risk making arbitrary changes that fix one issue while breaking others — a pattern analogous to reward hacking in reinforcement learning. Engineers are being urged to rethink traditional linear pipeline architectures to handle the higher volume, faster velocity, and unique failure modes that come with AI-driven development workflows.

0 comments Read more at DEV Community
← NewerPage 1027 of 3194Older →