SShortSingh.

Programming

0
ProgrammingDEV Community ·

Offline vibration probe diagnoses failing motors and can cut power before breakdown

Project Pulse is a magnetic vibration and sound sensor that attaches to small motors such as fans, pumps, and compressors. In about 30 seconds it reports the ISO severity zone and the likely fault, such as imbalance, looseness, overload, or bearing wear, in plain language. When paired with an optional stop plug, a Guard mode learns each motor's normal pattern, watches for drift, and can cut power if needed. It targets independent repair technicians and small workshops that cannot justify expensive commercial analyzers or subscriptions. All processing runs locally on the device, and an open-weight Gemma 3 1B model writes readable reports, with every figure checked against measured values before display.

0
ProgrammingDEV Community ·

Finoinc: Open-Source, Local-First Personal Finance App for macOS

A developer built Finoinc to track monthly spending without uploading bank data or maintaining a spreadsheet. Users import the OFX statement their bank provides, and the app skips duplicate entries, pairs charges with refunds, and applies custom categorization rules. It also flags recurring charges lacking a rule, avoids double-counting card bills and transfers between own accounts, and tracks bills on a calendar and investments separately. All data stays in a single local SQLite file with no account, telemetry, or default network calls, and the free, MIT-licensed app supports macOS 12+ on Apple Silicon and Intel.

0
ProgrammingDEV Community ·

Test-Driven Development: Writing Tests First as a Design Tool

Test-Driven Development reverses the usual workflow by writing a failing test before the code it exercises exists. The article argues that this turns tests into executable specifications that shape an API's design rather than merely confirming existing behavior. TDD is described as a development process rather than an architecture, though it tends to push code toward loosely coupled, injectable designs similar to Clean or Hexagonal Architecture. The core workflow is a short Red-Green-Refactor cycle: write a failing test, make it pass with minimal code, then clean up the design while the tests guard against regressions.

0
ProgrammingHacker News ·

Developer Builds Free Cloudflare-Hosted WhatsApp Bridge for Claude on Mobile

A developer wanted Claude on mobile to read their personal WhatsApp messages, but found existing options required a WhatsApp Business account or a server to run Baileys. They built a solution that runs for free on Cloudflare using a personal WhatsApp account. The tool also links to Google to pull full names for WhatsApp contacts, since otherwise many show only phone numbers and search works poorly. It is designed as a single-user install because the approach sits in a legal gray area and handles private data, so the repository is meant to be set up by pointing an AI agent at it. The project was shared on Hacker News, where it had 9 points and no comments at the time of reporting.

0
ProgrammingHacker News ·

OmniPTY: New Opinionated Terminal Emulator Launches for macOS and Linux

A developer has introduced OmniPTY, a terminal emulator designed for macOS and Linux, shared on Hacker News under a Show HN post. The project is hosted on GitHub, where the source and documentation are available to users. The post has received 4 points and no comments so far. The source material provides limited detail beyond the project's existence and its opinionated design approach.

0
ProgrammingHacker News ·

Why some developers are running their own LLM proxies to track model traffic

A post on jrdsgl.com questions how developers can know what is happening with their LLM usage if they rely on third-party services. The piece raises the idea of operating a self-hosted LLM proxy as a way to gain visibility into requests and responses. The article was shared on Hacker News, where it received a small amount of attention with three points and one comment. The source text provided is limited to the link and metadata, so the specific arguments and technical details of the post are not available here.

0
ProgrammingDEV Community ·

Why youtube-transcript-api works locally but fails on servers, and three ways to fix it

Developers using the youtube-transcript-api library often find it works on their own machine but returns errors such as RequestBlocked, IpBlocked, or 429 Too Many Requests when run from a server. The cause is that YouTube blocks most IP addresses belonging to cloud providers, and even home or office IPs can be blocked after sending many requests in a short period. The article outlines three remedies: routing requests through rotating residential proxies, running the fetching step from a home connection, or using a third-party transcript service that handles proxies on its behalf. Each option is suited to different situations, from steady production volume to low-volume personal projects or serverless workloads. The author also recommends caching transcripts already retrieved, since captions rarely change and fewer requests lower the chance of being blocked.

0
ProgrammingDEV Community ·

CERT-Bund Flags 35 Drupal Extension CVEs in Single High-Risk Advisory

Germany's CERT-Bund has published advisory WID-SEC-2026-3554 on 23 September 2026, grouping 35 vulnerabilities across Drupal extensions into one high-risk document. The flaws reportedly allow attackers to run arbitrary code, escalate privileges, bypass security controls, manipulate or expose data, and carry out cross-site scripting. Because extensions are maintained independently and run with the same privileges as core, the concentration of issues highlights risk in third-party plugins. The advisory does not publish version ranges, so affected releases must be checked CVE by CVE. A ZoomEye search on 28 September 2026 returned 436,331 Drupal assets, which indicates deployment scale rather than the number of hosts running a vulnerable extension.

0
ProgrammingDEV Community ·

SAP Launches SAP Pay, Bringing Supplier Payments Inside Its Cloud ERP

SAP has introduced SAP Pay, a built-in B2B payments service for SAP Cloud ERP that executes and auto-reconciles supplier payments once an invoice is approved. The service is run by Tereina, which SAP describes as a regulated payment provider founded in April 2026 and a wholly owned subsidiary. It supports traditional rails such as ACH, wire and real-time payments as well as USDC stablecoin transfers via Circle, with intelligent routing across more than 40 currencies. SAP says every transaction is screened for OFAC and KYC/AML compliance and that nothing moves without approval in the existing workflow. The service is generally available in the United States and United Kingdom, with pricing based on transaction volume and reported differences in how ownership and geographic eligibility are described publicly.

0
ProgrammingDEV Community ·

Quote Comparisons Need an Equivalence Policy Beyond Normalization

The article argues that sorting supplier offers by price is straightforward, but establishing what each price actually covers requires a clear policy. Different units, specifications, delivery terms, or validity periods can make offers look comparable when they are not. The author distinguishes normalization, which standardizes data for easier comparison, from equivalence, which judges whether an offer meets the approved need. A fictional mounting-kit example illustrates how a lower headline price can still be ineligible when components or delivery conditions are unclear. The piece recommends recording required attributes, permitted transformations, and review conditions, along with versioned policies and frozen comparison snapshots.

0
ProgrammingDEV Community ·

UUIDv7 Outperforms UUIDv4 in PostgreSQL Insert Speed and Index Size

PostgreSQL 18 introduced native support for UUIDv7, and a benchmark compares it against UUIDv4 for primary key insertion. The tests measured insertion time and on-disk index size across datasets of 1, 10, and 100 million rows. On small datasets the insert times were nearly identical, but at larger scales UUIDv7 was reported to be close to ten times faster. The UUIDv7 index was also consistently about a quarter smaller than the UUIDv4 index. The gap is attributed to how B-tree indexes handle the time-ordered UUIDv7 values compared with random UUIDv4 values.

0
ProgrammingDEV Community ·

New Local Tool Scans Employment Contracts for Clauses That May Affect Side Projects

A developer has built Moonlight, an open-source tool that reviews employment contracts to flag clauses relevant to side projects. The tool quotes the original contract language, highlights specific wording such as intellectual property terms, and suggests questions to investigate, without judging whether the user is legally permitted to run a side business. It is designed to run entirely on a local machine so sensitive contract details are never sent to a server, and it uses Node.js with the gemma3:4b model via Ollama. The project is MIT licensed and published on GitHub, and it was submitted to a Hacktoberfest open-source AI challenge.

0
ProgrammingDEV Community ·

boxr: A Rust-Based Rootless Container Engine Installable as a Single Binary

boxr is an OCI container engine written in Rust that ships as one static binary with no daemon and no need for sudo. The author notes that it is beta software in the 0.1.x series. Installation involves downloading a Linux archive from the project's GitHub releases page, verifying its SHA-256 checksum, and copying the bin/boxr executable onto the PATH. Users can confirm the install with boxr --version, and all state is stored under ~/.boxr in the home directory. The article contrasts this approach with Docker, whose CLI depends on a root-owned daemon, and invites readers to report feedback through the project's GitHub issues.

0
ProgrammingDEV Community ·

WebDecoy: a free WordPress plugin that catches bots without CAPTCHAs

WebDecoy is a free WordPress.org plugin that combines server-side checks, browser signals and hidden decoys to detect bots without showing customers image puzzles. It works locally without requiring a WebDecoy account or API key, and it defaults to a monitor mode that logs detections without blocking requests. The plugin covers comment spam, fake registrations, login brute-force attempts, WooCommerce card-testing patterns, and scanner or scraper activity. Users can optionally connect WebDecoy Cloud for centralized reporting, though the plugin makes no external connections until they choose to do so.

0
ProgrammingDEV Community ·

Coinbase's Autopilot Adds a Quality Loop to Its AI Customer Support Bots

Coinbase's engineering team has described Autopilot, a system that works behind the scenes to maintain its customer support bots rather than interacting with customers directly. The pipeline has agents write, test, probe, score, and analyze the versioned decision-tree procedures that support bots follow. The approach treats support automation as software requiring repeatable validation, since changes to procedures, models, or data interfaces can alter outcomes without warning. Autopilot pairs exploratory agents with deterministic services that grade results, and uses shared test contracts so Coinbase's bots and partner bots can be evaluated on identical inputs. The published account offers detailed architecture but no quantified results after deployment, such as changes in testing time or customer resolution and satisfaction scores.

0
ProgrammingDEV Community ·

Naive RAG pipelines fail in production; seven architectures address specific retrieval failures

A DEV Community article argues that the standard retrieve-and-stuff approach to retrieval-augmented generation works in demos but breaks on real questions, often pulling wrong or incomplete chunks. The author identifies recurring failure modes, including vocabulary mismatch, lost context from torn-out chunks, mismatched question and answer phrasing, and questions whose answers span whole documents. The piece presents seven distinct architectures, each designed to address one or two of these failures rather than being a single universal upgrade. It recommends hybrid search combining vector and keyword retrieval plus reranking as a baseline improvement before adopting more complex designs. Query rewriting techniques such as multi-query and HyDE are also covered as ways to bridge the gap between short questions and long answer passages.

0
ProgrammingDEV Community ·

Developer builds open-source tool to show which apps contact which servers, with Windows port by contributors

A developer built Snitch, an open-source network monitor that shows in real time which application is communicating with which server and in which country. The project was designed to avoid heavy dependencies and telemetry, using local packet capture, a short-lived local database, and offline geolocation so IP addresses stay on the user's machine. The macOS version launched first, and after the project was marked as seeking Windows help, outside contributors filed detailed bug reports and submitted code that produced a working Windows port in version 1.0.5. The project gained around 400 stars in three weeks and added Chinese documentation after a Chinese tech newsletter featured it and drove significant traffic.

0
ProgrammingDEV Community ·

Model Aliases Are Not Versions: Why Silent Drift Breaks Evals

Developers often report regressions with no config changes, assuming randomness, but model names can behave like moving pointers rather than fixed versions. Providers rotate aliases to point at newer snapshots, and the same alias may route to different backends over time, so logging the requested model, echoed model, timestamp, and base URL is recommended. Setting temperature to 0 does not guarantee identical outputs, since batching, floating-point ordering, and routing introduce variance. Fluent-looking output is also an unreliable signal of change, so the article suggests measuring self-consistency, output length, n-gram overlap, and error rates against a baseline. The author proposes a small, unexecuted drift-detection setup using a fixed prompt set and standard tooling.

0
ProgrammingHacker News ·

Hybrid Reasoning Agents Report 35–55% Token Savings and Up to 2.6× Faster Inference

A new arXiv paper introduces hybrid reasoning agents, a technique aimed at making AI reasoning more efficient. According to the reported results, the approach cuts token usage by roughly 35 to 55 percent compared with baseline methods. The same work claims inference speedups of up to 2.6 times. The item surfaced on Hacker News with only a handful of points and no discussion yet, so the claims have not been independently assessed by the community. Full methodology and benchmark details are available in the linked preprint.

0
ProgrammingHacker News ·

Netflix Open-Sources VMAF, a Perceptual Video Quality Metric Fusing Multiple Methods

Netflix has published VMAF, a video quality assessment tool hosted on GitHub. The metric is described as perceptual, meaning it aims to estimate how viewers actually perceive video quality rather than relying only on raw signal measurements. It works by fusing the outputs of multiple quality-assessment methods into a single score. The project drew modest attention on Hacker News, where it received 5 points and 2 comments. The source material provides limited further detail on specific features or benchmarks.

← NewerPage 355 of 3101Older →