SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Developer Builds Local RAG Pipeline Using LangChain, ChromaDB, and LM Studio

A developer has published a hands-on technical guide demonstrating how to build a Retrieval-Augmented Generation (RAG) system entirely from scratch without relying on OpenAI or any cloud services. The pipeline uses Python, LangChain, ChromaDB, and a locally hosted Qwen 9B language model running via LM Studio. The system is designed to solve a practical business problem where employees can query large volumes of HR documentation and receive accurate, document-grounded answers in seconds. The pipeline consists of four core components: a document processor that splits text into chunks, a custom embedding service, a ChromaDB vector store for semantic retrieval, and a local LLM for response generation. The guide highlights key engineering trade-offs, such as the tight coupling between chunk size and embedding quality, and the need to use a consistent embedding model across both indexing and querying stages.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Hybrid Search and Reranking Outperform Vector Search Alone in RAG Systems

A technical deep-dive published on DEV Community examines why vector search alone is often insufficient for production-grade Retrieval-Augmented Generation (RAG) systems. The article, third in a series on building reliable RAG pipelines, explains how embeddings translate text into geometric coordinates where semantically similar content clusters together. It argues that even high-quality embedding models fail when fed poorly structured or mixed-content chunks, turning bad input into confidently wrong retrieval results. The piece advocates combining semantic and lexical search methods — known as hybrid search — alongside reranking techniques to narrow large candidate sets down to the most relevant documents. Additional strategies covered include query optimization, metadata filtering, and context compression for more dependable real-world retrieval performance.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Good Code Comments Should Explain Why, Not Just What the Code Does

A technical article published on DEV Community argues that code comments are most valuable when they explain the reasoning behind a decision, not merely restate what the code does. Repeating the logic already visible in the code adds noise and can become misleading if the implementation changes. The piece emphasizes that hidden constraints — such as external system behavior, compatibility requirements, or performance limits — are exactly what comments should document. Without such context, future engineers risk breaking functionality through changes that appear harmless on the surface. The article advises keeping reason-based comments concise, accurate, and updated whenever the underlying constraint changes.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why 'Verified' in a Schema-Change Report Does Not Mean Safe to Release

Data investigation reports often conflate observed changes with release decisions, creating a dangerous gap in schema-change workflows. A report confirming that one field was added between two Delta table versions does not verify the field's name, type, or compliance with release rules. Experts recommend that reports explicitly separate state identity, evidence coverage, and release intent into distinct, labeled checks. Incomplete findings should be recorded as such rather than omitted, so reviewers know exactly what was and was not examined. Whether reviewed by a human or an AI agent, a report must answer which states were compared, what was actually read, and what evidence remains missing before any release decision is made.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI model benchmarks ignore cost — here's the math that actually matters

Popular AI model benchmarks typically rank outputs by quality and token usage, but almost never show what developers are actually charged. The true cost depends on a simple formula multiplying input and output tokens separately against a vendor's listed prices, and those two rates can differ by three to ten times. Using a sample coding task of 40,000 input and 12,000 output tokens, costs across leading models in mid-2026 range from roughly $0.056 to $0.50 per run — a ninefold spread no leaderboard displays. Choosing a cheaper model can backfire if it requires more retries and burns proportionally more tokens, potentially making it costlier than a pricier, more capable alternative. The author recommends logging input and output tokens separately per task, benchmarking top candidates on your own workloads, and recalculating costs quarterly as model prices shift frequently.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

One developer built 2,768 CSS demos solo — here's how he prevented style collisions

A developer behind codefronts.com hand-wrote 2,768 CSS demos across 138 collections, all of which must coexist on shared gallery pages without breaking each other's styles. The core problem is that CSS silently resolves duplicate selectors by source order, meaning two demos sharing a class name like .card or .btn can invisibly overwrite each other with no console warning. Framework solutions such as CSS Modules and Shadow DOM were ruled out because they produce hashed or unfamiliar class names that make copy-pasted code harder for end users to read and adapt. The developer instead settled on a manual, mechanical naming convention applied at authoring time, designed to keep output legible to anyone pasting a snippet into an unfamiliar project. The system prioritizes human-readable code over build-tool automation, reflecting the unusual constraint that a copy-paste demo library must remain transparent to users who never see the underlying build process.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How Azure Data Factory Pipelines Move Data From Source Systems to Production

Azure Data Factory (ADF) is a cloud-based data integration service that orchestrates the movement and transformation of data across enterprise systems. A typical ADF workflow begins with a trigger that determines when a pipeline should run, followed by activities such as data extraction, loading to staging tables, and validation. ADF connects to source systems — including Oracle, MySQL, REST APIs, and cloud storage — through Integration Runtimes, with Self-hosted Integration Runtimes enabling access to data inside private networks. Pipelines can be scheduled, manually triggered, or event-driven, and often work alongside tools like Apache Airflow and dbt to complete end-to-end data transformations. Understanding each component's role — triggers, linked services, integration runtimes, and activities — is key to building and troubleshooting reliable data pipelines in ADF.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AIUC-1 Standard Offers AI Agent Safety Framework Before Products Ship

The Artificial Intelligence Underwriting Company has published AIUC-1, a standard for AI agents covering six operational domains: data and privacy, security, safety, reliability, accountability, and societal risk. Contributors to the standard include Stanford, MIT, MITRE, the Cloud Security Alliance, and Google Cloud. The framework is designed to help SaaS teams assess whether their agent designs are sound before deployment, addressing gaps around logging, authorisation, and incident response that internal responsible-AI guidelines often overlook. Unlike existing compliance frameworks such as SOC 2 or GDPR, AIUC-1 is specifically shaped around agents that can take actions on behalf of users, call external systems, and retain context across sessions. For engineering and product teams building on protocols like MCP or OAuth-connected assistants, AIUC-1 serves as a pre-deployment checklist rather than a formal certification requirement.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Disabling a workflow quietly broke publishing for a month — here is what that reveals

A developer disabled a GitHub Actions workflow that was corrupting static HTML files on their site, believing the fix was straightforward and temporary. A month later, when they tried to publish an article, they discovered the workflow was actually the only mechanism for converting JSON source files into live web pages. The only record of why the automation had been switched off was a single commit message referencing a prior content-poisoning incident — the full reasoning existed only in memory. The incident highlights a broader systems risk: disabling a component does not remove it from the architecture, it simply causes it to return nothing silently, with no downstream error signal. The author concludes that any disabled component should have a named owner and an explicit re-entry condition, not just an off state with no resolution criteria attached.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

ATLOCK Uses Windows NTFS ACLs for File Protection Instead of Encryption

A developer building ATLOCK, a Windows security tool, chose NTFS Access Control Lists over encryption to protect files, arguing that access control and encryption solve fundamentally different problems. File Guard leverages Windows' native permission system to restrict which identities can interact with protected files, keeping them unchanged on disk. The Password Vault, however, uses Fernet encryption with AES-128-CBC and HMAC-SHA256, with keys derived via PBKDF2-HMAC-SHA256, since confidentiality of stored credentials demands a different approach. ATLOCK also includes an Intruder Ops module that captures webcam images on unauthorized access attempts, designed as an asynchronous workflow to avoid freezing the UI during camera initialization. The developer describes the project as a security engineering experiment focused on matching the right protection mechanism to each specific threat.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Google Sheets Can Now Bulk-Geolocate IP Addresses Using a Free Apps Script Tool

A developer guide published on DEV Community explains how to turn Google Sheets into a bulk IP geolocation tool using a free Apps Script integration from IPGeolocation.io. Users paste a ready-made script into the Apps Script editor, store a free API key once, and gain access to custom formulas without writing any code. The formula IPGEO handles individual IP lookups, while IPGEO_BULK can process up to 500 addresses in a single call, returning data such as country, city, timezone, ASN, VPN status, and threat scores. The setup process takes approximately 10 minutes and requires no third-party add-ons. The full source code and formula reference are available on GitHub, making the tool accessible for use cases like fraud review, login monitoring, and ad-click analysis.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How Tally Sticks, Cowrie Shells and Coins Laid the Foundations of Modern Finance

Long before digital payments, societies developed sophisticated financial technologies to record and transmit debt at scale between strangers. Medieval England used tally sticks — notched hazel wood split in two — as a tamper-proof, distributed ledger system that served the English Crown for 700 years until the sticks were famously burned in Parliament's basement in 1834. In West Africa, cowrie shells functioned as official currency for centuries until European traders flooded markets with a cheaper species, collapsing stable monetary systems through supply manipulation. Ancient coinage, emerging around 600 BCE, solved metal-purity verification by backing raw metal with a trusted authority's stamp, though rulers quickly exploited mint control to debase currencies. The article argues that money, in every era, has been a technology for solving the same core problem: how to maintain trusted records of obligation at scale.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Embarcadero Releases Delphi 13 Community Edition for Free Download

Embarcadero has officially launched Delphi 13 Community Edition, making it available to developers at no cost. The release was announced on the company's official blog and quickly gained attention on Hacker News. Community Edition versions of Delphi are typically aimed at students, hobbyists, and small development teams who want access to the tool without a commercial license. The new version continues Embarcadero's practice of offering a free tier alongside its paid professional offerings.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Technical SEO for Developers: Core Web Vitals and Semantic HTML Matter More Than Keywords

Developers often dismiss SEO as a marketing concern, but a significant portion of search ranking depends on technical factors they directly control. Key issues include Cumulative Layout Shift, which occurs when page elements move unexpectedly due to undefined image dimensions, and poor use of non-semantic HTML that makes it harder for crawlers to understand page structure. Single Page Applications built with React, Vue, or Angular can hurt indexability since JavaScript rendering is slower and less reliable for bots, making Server-Side Rendering or Static Site Generation preferable. Proper meta tags, Open Graph data, and Twitter Card tags improve how content appears in search results and on social media, indirectly boosting traffic signals. Managing crawl budget through a robots.txt file and a sitemap.xml ensures search engines focus on the pages that matter most.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Virginia Residents Pay Higher Power Bills as Data Centres Consume 40% of State's Electricity

John Steinbach, a Manassas, Virginia resident, saw his electricity bill jump from roughly $100 to $281 in January 2026, despite no change in his household energy use. A Consumer Reports investigation published in March 2026 linked such spikes to the massive power demands of Data Center Alley, the world's densest concentration of server farms, which now consumes around 40 percent of Virginia's electricity. Data centres currently account for half of all new electricity demand across the United States, and residential prices rose 7.1 percent nationally in 2025 alone. Under existing utility regulations, the cost of grid upgrades needed to supply hyperscalers like Amazon, Microsoft, and Google is spread across all ratepayers, effectively shifting billions of dollars from ordinary households to some of the world's most profitable corporations. Critics argue that residents have had no say in this infrastructure expansion, raising broader questions about who should bear the cost of powering artificial intelligence.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why You Should Stop Using Redundant Instructions With Modern AI Reasoning Models

Modern AI reasoning models, as of mid-2026, are built to verify their own steps and regulate response depth without needing explicit prompts like 'double-check your work' or 'think step by step.' Overloading these models with such instructions wastes tokens, creates friction, and can cause them to over-verify or misinterpret conflicting rules. Experts recommend replacing vague directives with concrete alternatives — such as using official effort-level settings instead of 'think deeply,' and specifying exact output length rather than simply saying 'be concise.' Prompt structures should eliminate duplicate rules, define clear autonomy boundaries, and use verifiable success criteria to guide model behaviour. The advised 2026 prompting framework focuses on role, objective, constraints, output format, and stopping rules — dropping process instructions in favour of clearly defined outcomes.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Cisco ASA/FTD Zero-Day Flaw Enables Unauthenticated DoS on SSL VPN

A actively exploited vulnerability, tracked as CVE-2026-20349, affects Cisco ASA and FTD appliances, allowing unauthenticated attackers to crash the Remote Access SSL VPN service. When the firewall goes down, logging and policy enforcement stop, creating a window for malicious traffic to pass undetected. Enterprises are exposed to risks including lateral movement, loss of network visibility, and potential compliance violations. Cisco has released a patch, and administrators are advised to apply it promptly alongside hardening measures such as blocking untrusted IPs, enforcing multi-factor authentication, and segmenting VPN endpoints from critical assets. Security teams are also encouraged to monitor for anomalous traffic patterns and conduct regular DoS simulation exercises against perimeter devices.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Researchers Hired North Korean IT Operatives via Fake Crypto Startup in Security Sting

Cybersecurity researchers set up a fictitious cryptocurrency startup and advertised developer roles, ultimately hiring three individuals suspected of being North Korean IT operatives. The suspects gained entry not through hacking but by submitting job applications, passing interviews, signing contracts, and receiving legitimate system access. Investigators noticed red flags during onboarding, including geographic inconsistencies in identification documents and the use of a Social Security number linked to mismatched locations. Once given virtual machines, the workers quickly gathered system information, installed remote-access software, and connected personal accounts to corporate devices. The case highlights the need for continuous behavioral monitoring and Zero Trust principles that extend well beyond the initial hiring and authentication stages.

0 comments Read more at DEV Community
← NewerPage 193 of 1339Older →