SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Developer Builds AI-Assisted Contradiction Resolver Using Sanity App SDK

Software developer Shahbaz Ali created a contradiction-resolution tool called Contradiction Triage as part of the DEV × Sanity Challenge, addressing the common problem of conflicting information across company documents. The tool detects contradictions between documents written by different teams — such as legal, support, and marketing — and creates a structured case for each conflict. An AI agent proposes a ruling with a rationale and cited precedents, which a human judge then approves or overrides. Once a decision is made, it becomes the canonical answer served across the platform, complete with an audit trail and attribution. The system uses six document types built on Sanity's App SDK and stores all rulings as queryable documents to prevent repeated, unresolved debates.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Fine-tuned 310M encoder beats Jev on topic classification but ties on polarity tasks

A developer ran three Japanese text classification tasks — long-document topic, short-text polarity, and financial-sentence polarity — across six systems using 250 labeled rows each. A 310M ModernBERT encoder fine-tuned on just 250 labels outperformed TypeSafe's Jev decision API by 12 percentage points on topic classification, while also being 4 to 20 times faster. However, on the two polarity tasks, the trained encoder and Jev were statistically tied, with neither system holding a clear overall advantage. Zero-shot open-source models, including a GLiNER-family classifier, consistently underperformed Jev across all three tasks. The key finding is that which system wins depends on task structure — zero-shot APIs like Jev are practical when no labeled data exists, while small trained encoders gain an edge once a few hundred labels are available.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

DeepSeek's Time-Based API Pricing Caught Developers Paying Double Overnight

DeepSeek's API charges peak rates during Beijing business hours (09:00–12:00 and 14:00–18:00 CST, Monday–Friday), with off-peak prices exactly half the peak cost. A developer discovered their 3 a.m. Eastern batch job was inadvertently running during Beijing's afternoon peak window, doubling API costs without any code error. The classic US overnight maintenance window of roughly 2–6 a.m. Eastern falls squarely within DeepSeek's peak billing period, while standard East Coast business hours are off-peak. The recommended fix is to write cron schedules in Asia/Shanghai timezone rather than converting to UTC, since US daylight saving shifts can silently break hand-converted times twice a year. Developers are also advised to account for long-running jobs that straddle peak boundaries and retry queues that may drain into expensive windows.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Thousands of Industrial Control Devices Found Exposed on Public Internet

A ZoomEye scan conducted on September 20, 2026, detected over 10,500 records responding to industrial protocol fingerprints across three queries. Modbus accounted for the largest share with 9,820 records, followed by EtherNet/IP with 585 and Siemens S7 with 173, each showing distinct geographic distributions. These protocols were originally designed for isolated, air-gapped networks and lack native authentication or encryption in their base forms. Researchers cautioned that a reachable device does not automatically indicate a vulnerability or confirm it controls critical infrastructure, as test rigs and simulators can appear identical in such scans. The findings nonetheless highlight that internet exposure of legacy industrial protocols is a widespread, multi-regional concern requiring careful network architecture controls.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Study: Claude Code hooks block protected file edits better than CLAUDE.md rules

Researchers ran 29 headless sessions using Claude Code v2.1.273 to compare two methods of preventing edits to a protected directory: a plain-language rule in CLAUDE.md and a PreToolUse shell hook. Both methods successfully blocked edits in normal conditions, but when users explicitly claimed authorization in the prompt, the CLAUDE.md rule was overridden in both test runs while the hook held firm in both. The CLAUDE.md approach added 58–70 input tokens to every request, whereas the hook incurred no token cost unless triggered, though it did introduce an extra model round trip on direct requests. Claude Code's own documentation notes that CLAUDE.md content is treated as context rather than enforced configuration, and that PreToolUse hooks are the correct mechanism for reliably blocking actions. The experiment provides concrete measurements backing that distinction, showing hooks are more tamper-resistant against user-supplied authorization claims.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Key Docker Commands for Managing Containers, Images, and Volumes

This tutorial covers essential Docker commands used in day-to-day container management, including pulling images, running containers, and executing commands inside them. Lightweight base images such as Alpine and Bookworm are recommended for small applications to reduce overhead compared to full OS images. Commands like docker run -it, docker exec, and docker ps help users interact with and monitor running containers. Port mapping with the -p flag exposes container services to the host, while volume mounting with -v allows persistent or shared file storage. Additional commands such as docker logs, docker inspect, and the --rm flag for auto-removal round out core container lifecycle management.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer's Second AWS AI Agent Accidentally Fulfilled His First Project's Wishlist

A software developer completing an AWS nanodegree shipped his second AI support agent, Nova Adiutrix, built on Amazon Bedrock to handle order tracking, refunds, and product queries for a simulated online store. During testing, the agent incorrectly refused to share order details with the account holder, citing a security check that falsely flagged the order as unlinked — making a bug look indistinguishable from correct security behavior in the logs. The developer later realized that Nova Adiutrix had organically implemented three features he had explicitly deferred in his first project, Nova Trivium, listed under a 'Not built, and why' section in that project's README. Key architectural differences between the two builds included a shift from pasting static FAQ content directly into a system prompt to using a Bedrock-managed vector knowledge base for retrieval. The developer noted that each project surfaced the same core challenges — where knowledge lives, where rules are enforced, and whether a security checkpoint can be fooled by the system it is meant to check — only at a greater level of complexity.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Fliproom Uses Sanity CMS to Turn Room Changeovers Into Visual Checklists

Fliproom is a small open-source planning tool built for the Sanity CMS developer challenge that converts room layout changeovers into step-by-step visual checklists. A volunteer selects a starting and ending room layout, and the app automatically derives a numbered move list by comparing equipment placements across layouts rather than requiring manually authored instructions. The demo features a fictional 12-by-8-metre community hall with eight pieces of equipment and three layouts, hosted publicly on GitHub Pages with no login required. Dependencies between steps keep certain moves locked until prerequisite actions are completed, and progress is preserved across browser reloads. The MIT-licensed app is built with Next.js, TypeScript, React, and Sanity, with AI agents handling most of the technical development and code review.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Logs 445 Clones but 2 Stars for Free Mental Health Toolkit

A developer released a collection of 36 free, client-side mental health tools built in vanilla JavaScript, requiring no backend, signup, or dependencies. After 30 days, the project attracted 157 unique cloners and 445 total clones on GitHub, but earned only 2 stars and no forks. Alongside the repository, the developer published 62 articles on Dev.to, generating just 607 views and 24 reactions in total. Attempts to monetize through a Gumroad template, a paid API listing on RapidAPI, and an AI agent marketplace yielded zero sales or paid calls. The developer attributes the low conversion rate to a weak repository presentation and a monetization model undermined by the toolkit's completeness at no cost.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How HLS, MPEG-TS and LL-HLS Power High-Performance 4K Live Streaming in 2026

As 4K HDR displays and gigabit internet become mainstream, IP-based video delivery has largely replaced traditional cable and satellite TV. Live streaming at scale — particularly for high-viewership events like sports or pay-per-view — poses distinct engineering challenges around latency, buffering, and network congestion that on-demand platforms do not face in real time. Two primary transport standards dominate live video distribution: HLS, developed by Apple, which breaks streams into small HTTP chunks with latency of 12–25 seconds, and MPEG-TS, which transmits continuous fixed-size packets with sub-2-second latency but is more vulnerable to packet loss. A newer variant, Low-Latency HLS (LL-HLS), bridges the gap by delivering partial chunks via HTTP/2 or HTTP/3, achieving latency of 2–5 seconds suitable for interactive broadcasts. Best practices for reliable 4K live streaming in 2026 include using wired or 5 GHz Wi-Fi connections, tuning client-side buffer settings, and selecting providers with redundant multi-gigabit edge CDN infrastructure.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How 11 Variants of One Shell Script Created a Company-Wide Security Crisis

A software engineer discovered that a single useful shell script had been copied and independently modified into 11 variants spread across hundreds of laptops at their employer. Several of these scripts were launching pods into Kubernetes clusters using container images that were years old and riddled with known security vulnerabilities. Some of those clusters fell within strict compliance boundaries, including FedRAMP-regulated environments, raising serious audit concerns. When a scanner flagged the outdated images, the team realized they had no reliable way to push a one-line fix to machines they did not control. The incident highlighted a fundamental distinction between a script — which only runs where it lives — and a self-updating tool that can enforce changes across systems without relying on users to act.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Three Threat Groups Exploit Cisco Firewall Manager Flaws, Including CVSS 10.0 Bug

Cisco Talos confirmed on September 9, 2026, that two vulnerabilities in Cisco Secure Firewall Management Center (FMC) were being actively exploited by three distinct threat clusters. The more severe flaw, CVE-2026-20079 (CVSS 10.0), allows an unauthenticated attacker to bypass login and execute commands as root via a crafted HTTP request, while CVE-2026-20316 (CVSS 5.3) involves hardcoded credentials enabling limited unauthorized access. A suspected Sandworm-linked group chained both flaws to install a Cyclops Blink malware variant, a credential thief deployed a web shell to extract authentication data, and a Qilin ransomware affiliate used the static credentials to map networks and deploy ransomware. CISA added CVE-2026-20079 to its Known Exploited Vulnerabilities catalog the same day, setting a federal remediation deadline of September 12, 2026. Because FMC centrally manages firewall policy and configuration across entire network estates, its compromise effectively undermines the security of every device it controls.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Reel Motion App Converts Instagram and TikTok Workout Videos Into Guided Routines

A developer has launched Reel Motion, an app designed to transform workout videos from platforms like Instagram, TikTok, and YouTube into structured exercise routines. Users share a video with the app, which then detects the exercises and generates a step-by-step workout complete with illustrations, instructions, reps, timers, and audio guidance. The app also supports voice control, allowing users to navigate workouts without picking up their phone between sets. All imported workouts are saved in a personal library, replacing the need to bookmark scattered videos across multiple platforms. The app is now live at reelmotion.fit and the developer is seeking feedback from the tech community.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Extract Text and Images from PowerPoint Files Using Free Spire.Presentation in C#

Developers working with PowerPoint files in .NET often face challenges because the native framework lacks built-in APIs for parsing .pptx content directly. Free Spire.Presentation for .NET is a lightweight, free library that supports .NET 6 and above and simplifies tasks like text extraction and image export from PowerPoint files. Using Visual Studio 2022 or Rider, developers can batch-extract all slide text into a TXT file by traversing slides, shapes, and paragraphs through a three-level hierarchy. The library also supports exporting embedded images as local PNG files, reducing boilerplate code and format compatibility issues. It requires no watermark and carries no commercial restrictions for everyday non-commercial parsing use cases.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How Port, Baud, and Protocol Mismatches Cause RTK GNSS Failures in Drones

Engineers integrating Septentrio-powered GNSS receivers into UAV autopilots like ArduPilot and PX4 most commonly encounter a 'no fix' error caused not by faulty hardware but by port, baud rate, or protocol direction mismatches. Septentrio modules use LVTTL (3.3V logic) UARTs that are incompatible with RS-232 signals, and connecting the wrong voltage source can damage the port. Experts recommend assigning a dedicated UART per task — one for autopilot position output via NMEA, one for incoming RTCM or CMR corrections, and a third for raw logging — to avoid configuration conflicts that mimic hardware failures. A baud mismatch typically produces garbled data rather than silence, leading integrators to mistakenly suspect a dead receiver rather than a misconfigured parameter. Changing baud rates on both ends simultaneously without verification is a common mistake that turns two functional ports into one broken link.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How SaaS Founders Can Use Directory Listings as a Long-Term Distribution Channel

A detailed submission guide published on DEV Community outlines how SaaS founders can systematically list their products across online directories to build lasting distribution. The author argues that most solo founders skip this step, despite it offering compounding benefits through backlinks, referral traffic, and AI citation visibility. Research cited in the piece suggests over 35% of LLM citations for B2B SaaS products trace back to just ten sources, making high-authority directory listings increasingly valuable. The guide draws from a curated database of 393 SaaS directories, filtered by category, fee type, and domain authority. Unlike paid ads or social posts, directory listings are described as permanent, low-cost assets that continue driving discovery long after submission.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Rust Crate to Block AI Agents from Accessing Cross-Tenant Data

A developer has released TenantInvariant, an experimental open-source Rust library designed to prevent AI agents from accessing resources belonging to the wrong customer in multi-tenant SaaS applications. The problem arises when an AI model generates a tool call referencing a valid resource ID that happens to belong to a different tenant, bypassing intended data boundaries. The library addresses this by resolving resource ownership server-side from a trusted source, rather than trusting any tenant information supplied by the AI model itself. It then compares the authenticated actor's tenant against the resolved resource owner, denying access if they do not match or if ownership cannot be determined. The crate aims to make tenant isolation an enforceable, named invariant in the application layer rather than relying solely on prompt engineering.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Engineer Builds Testable AWS Multi-Region Disaster Recovery with Real Failover Under 15 Min

A software engineer has built and open-sourced a multi-region 'pilot light' disaster recovery setup on AWS, using Terraform to replicate a notes API across two regions in Ireland and Paris. The architecture keeps the secondary region dormant — with zero running application instances — until a failure is detected, at which point Route 53 automatically flips DNS within roughly 30 seconds. PostgreSQL data is continuously replicated via a cross-region read replica, while S3 and Secrets Manager stay live in both regions simultaneously, minimizing manual intervention during failover. A single script handles the one step that cannot be automated — promoting the RDS read replica — targeting a full recovery time of 10 to 15 minutes. The project was designed to demonstrate that pilot light is the most cost-effective disaster recovery pattern achieving minute-scale RTO, and benchmarks each trade-off against AWS's Well-Architected Framework.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Airflow May Be the Wrong Tool for Medallion Data Pipelines

A data engineer with six years of experience argues that Apache Airflow is poorly suited for medallion architecture pipelines, citing dependency conflicts, lack of atomicity, and excessive orchestrator maintenance overhead. For AWS-based pipelines, the author recommends AWS Step Functions for its server-free state machine model and native support for long-running Spark jobs via callback patterns, while noting its execution history limits. Teams using Delta Lake on Databricks are advised to adopt Databricks Workflows instead, as it integrates directly with cluster lifecycle management and Delta commits, avoiding orphaned clusters. A key architectural warning is raised against monolithic pipelines that chain Bronze, Silver, and Gold layers serially, as a single failure forces costly full reruns. The overall message is that no single orchestrator fits all use cases, and engineers should choose tools based on the specific failure modes their infrastructure is likely to encounter.

0 comments Read more at DEV Community
← NewerPage 1080 of 3216Older →