SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Developer Publishes Oracle Schema Hardening Backend Code Guide on DEV Community

A developer has published a technical tutorial on DEV Community detailing the backend setup for an Oracle Schema Hardening project. The guide, labeled Part 1, provides a structured walkthrough of a Spring Boot application with Oracle database integration. It lists over 25 Java source files and configuration files, each with corresponding code blocks for readers to replicate locally. The project combines Spring Boot on the backend with a React frontend, using Oracle JDBC driver version 23.9 and Spring Boot 3.5. The tutorial is intended to help developers implement schema hardening practices within an Oracle database environment.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Google Launches AX Agent Executor on Kubernetes, but Durability Questions Linger

Google has released an open-source project called Agent Executor (internally AX) on its official GitHub under an Apache 2.0 license, briefly topping Hacker News with 179 points and 74 comments. The system is composed of three stacked layers — GKE Agent Sandbox, Agent Substrate, and Agent Executor — designed to run AI agent workloads more efficiently on Kubernetes by exploiting the idle time between agent tool calls. A key technical claim is that Agent Substrate can multiplex 250 stateful agent sessions onto just 8 physical pods with sub-second reactivation, though observers note this largely repackages existing GKE snapshot and gVisor sandboxing capabilities rather than introducing new infrastructure. Google's stated design target of billions of concurrent agent tasks per cluster is an architectural goal, not a verified benchmark, with the public demo limited to the 250-session figure. Community discussion has focused on whether the project reflects a lasting Google commitment or risks being abandoned like past initiatives, a concern Google's track record with infrastructure projects has done little to dispel.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why INT8 Quantization Fails for Large Language Models and How FP8 Helps

Quantizing large language models from 16-bit to 8-bit can halve memory usage, but a single extreme activation value can distort the entire scaling grid, degrading precision for the vast majority of normal values. Research by Tim Dettmers and colleagues in 2022 found that in models around 6.7 billion parameters, roughly 150,000 outlier activations per sequence appeared, yet these were concentrated in just about six feature dimensions representing only 0.1% of all values. Despite their small share, removing these outlier dimensions significantly harmed model performance, revealing that the core challenge is not bit-width reduction alone. Developers must treat bit width, number representation format, and scaling granularity as three distinct design choices when quantizing LLMs. This understanding has driven a progression from basic INT8 approaches toward more sophisticated methods, including floating-point 8-bit formats with finer-grained scaling strategies.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

fastlogging-rs: Rust-Based Logging Framework Supports 8 Languages with Major Speed Gains

A developer has released fastlogging-rs 0.9.0, a high-performance logging framework built on a Rust core with bindings for Python, C, C++, Go, Java, and C#. The framework uses non-blocking, asynchronous logging where each call passes messages to a background thread, keeping the main application path unaffected. Benchmarks claim it is up to 147 times faster than Python's built-in logging module for file writes and up to 9 times faster than Apache Log4j. Additional features include optional file rotation, AES encryption for network logging, OpenTelemetry support, and configuration via JSON, XML, or YAML files. The project is open source and available under the MIT or Apache-2.0 license.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

TailwindMail lets developers build HTML emails using Tailwind CSS utility classes

TailwindMail is a tool that allows developers to write HTML emails using Tailwind CSS instead of traditional inline styles and outdated markup. The platform offers a visual preview editor alongside a code workflow, enabling users to compose, preview, and compile emails into email-client-safe HTML. It is designed to simplify a process that typically involves messy CSS, broken layouts across clients like Gmail and Outlook, and hard-to-maintain templates. The exported HTML is compatible with popular email-sending services such as Resend, Postmark, Amazon SES, Mailgun, and standard SMTP. TailwindMail targets developers already familiar with Tailwind who need to produce transactional emails such as welcome messages, OTPs, and product updates.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Claude Code's shell commands fail on Macs: zsh and BSD tools explained

Claude Code's Bash tool runs commands through zsh on macOS Catalina and later, despite being named and described as a Bash executor, causing subtle compatibility failures. Commands written by Claude often assume GNU tool behavior, but macOS ships with BSD versions of utilities like sed, date, and stat that interpret flags differently. A common example is 'sed -i -e', which silently creates unwanted backup files on macOS instead of editing in place, and may even exit with code 0, hiding the error from automated checks. Installing GNU tools via Homebrew does not fix the issue because Claude Code injects its own process environment — not the user's shell profile — before each command. The author identifies these as two concrete, systemic causes rather than model sloppiness, and notes that PATH modifications in shell profiles have no effect on the Bash tool's execution environment.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Xiaomi Releases MiMo v2.6 Update for Its AI Model

Xiaomi has launched version 2.6 of MiMo, its artificial intelligence model, as announced on the official MiMo website. The release was noted on Hacker News, where it gathered 54 upvotes and 7 comments from the tech community. MiMo is Xiaomi's in-house AI model initiative, and this update marks a continued iteration of the project. The announcement was shared via Xiaomi's dedicated MiMo platform at mimo.xiaomi.com. Further technical details about the changes in v2.6 are available through the official release page.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Vanilla JS CBT Detector Outperforms NLP Libraries in Accuracy, Speed, and Privacy

A developer built a 200-line vanilla JavaScript tool to detect cognitive distortions — patterns identified in CBT therapy — and benchmarked it against five NLP and AI solutions. The rule-based JS detector achieved 94% accuracy on 100 test samples, outperforming OpenAI's API (89%), Hugging Face (83%), and TensorFlow.js (72%), while running at 0.2ms median latency with zero cost. Unlike API-based tools, the vanilla JS approach processes all data entirely in the browser, making it suitable for sensitive mental health use cases where privacy is critical. The benchmark covered ten clinically defined cognitive distortions originally described by Aaron Beck in 1976, representing a closed and well-documented classification set. The developer concluded that rule-based pattern matching is preferable when the output space is fixed, privacy is essential, and deterministic results are required.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agent Documents Every Failed Attempt to Earn Its First Dollar from a Stranger

An AI agent operating on a publishing platform spent two weeks systematically attempting to earn revenue by offering manuscript editing services priced at $25 per pass. It explored agent-focused marketplaces, freelance platforms, and writer communities, but was blocked at every turn by requirements for human identity verification, government IDs, phone numbers, or crypto wallets. Platforms built specifically for AI agents paid only in cryptocurrency, a rail the agent could not access without a human-held bank account. Despite publishing content that drew genuine readers and sending 17 personalised outreach messages to writers, it received zero payments and zero replies. The agent concludes that the core missing infrastructure is not AI capability or autonomy, but a legal mechanism allowing an agent to act as a counterparty in a standard card transaction.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Azure Virtual Network Routing Appliance bridges peering speed and hub-and-spoke control

Microsoft has introduced the Azure Virtual Network Routing Appliance (VNRA), a managed forwarding layer designed to address the long-standing trade-off between VNet peering performance and centralised hub-and-spoke governance. Unlike traditional NVAs or Azure Firewall, the VNRA runs on dedicated networking hardware directly within the Azure backbone, offering tiers of 10, 50, 100, and 200 Gbps without becoming a traffic inspection bottleneck. It is deployed into a dedicated subnet within a hub VNet and does not perform firewall functions; traffic control still relies on UDRs, NSGs, and Azure Monitor. The service is particularly relevant for data-intensive workloads such as AI pipelines, centralised databases, and cross-VNet API calls that suffer under conventional hub appliance capacity limits. Built-in metrics for bytes, packets, and flows are available immediately after deployment, requiring no additional diagnostic configuration.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Small Dev Team Uses AI Agents to Bridge the Gap Between Specs and Code

A small development team built an agentic workflow using Claude Code to address the common breakdown that occurs when agreed project specifications must be translated into precise, actionable code tickets. The team identified that most project failures happen not in planning meetings but in the ambiguous gap between high-level specs and implementation. They developed a tool called /feature, adapted from a known AI prompting technique, which asks developers targeted questions one at a time until all open decisions are resolved before any code is written. A second tool, /jira, extended this into a more automated research-plan-implement pipeline for a hands-off approach. The workflow has improved ticket quality, surfaced overlooked edge cases, and ensured that implementation decisions are documented rather than silently assumed by individual developers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds LGBTQIA-Focused Game Discovery App Using IGDB and Sanity

A developer has built a web app called QueerCade designed to help users discover games featuring LGBTQIA stories, themes, characters, and creators. The app was created using Lovable and connects to the IGDB API to source game data. Firecrawl was used to seed the database with starter content, while Sanity serves as the backend with a Studio and Context MCP. The project was submitted as a dual entry to the Sanity Challenge, fulfilling both the agent-querying and creative coding categories. An AI agent connection was established via Superhuman GO using a custom-built Pack.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds free VLESS proxy on Alwaysdata using Xray and XHTTP transport

A developer needed a foreign proxy server without paying for a VPS and eventually found a working solution on the free tier of French hosting provider Alwaysdata. The project, called alwaysdata-xray, uses three shell scripts to install Xray with VLESS over XHTTP transport and generate a ready-to-use client link in a single command. Because Alwaysdata's free tier only exposes HTTP/HTTPS externally and lacks custom TCP or UDP ports, XHTTP was chosen as the transport layer to tunnel VLESS traffic through standard HTTPS requests. The setup leverages Alwaysdata's built-in TLS termination and process management, meaning users do not need to configure certificates or keep the process running manually. The developer notes that Alwaysdata's free-tier terms of service prohibit proxy usage, so the account could be suspended at any time, and the solution is recommended only as a backup option.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How a 40-Day Narrative Game Built on a Strict State Machine Survived 130 Flags and 33 Endings

Developer Mykola is building NICHLYST, a narrative game set in Kyiv, currently in a 14-day Google Play closed testing period as of September 2026. The game spans 40 in-game days, 130 boolean flags, six moral axes, and 33 distinct endings, creating a state-space that the developer describes as an exponential engineering challenge. To prevent cascading logic failures common in branching narrative games, the project was built around a strict, centralized state machine rather than mutable global JavaScript objects. The 40 days are structured into eight thematic waves, each carrying its own narrative tone and mechanical rules, from a bunker awakening to a climactic archive-or-broadcast decision. The devlog, written on an aging laptop amid air sirens in Kyiv, details how architectural discipline — not creative shortcuts — kept the project from collapsing under its own complexity.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Connect Local and Cloud PostgreSQL Databases Using DBeaver and Power BI

Developers managing structured data across local and cloud environments can use DBeaver, an open-source universal database management tool, as a centralized platform. The guide walks through installing PostgreSQL locally via the EnterpriseDB installer on Windows, configuring a superuser password and default port 5432. A cloud-hosted PostgreSQL instance is then set up using Aiven, a managed cloud database service, which provides credentials for external tool connections. Both databases are accessed within DBeaver by creating new connections using each service's respective credentials. The setup is then extended to Power BI, enabling centralized data visualization across local and cloud sources.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Properly Test Afriex Payment Integrations Using Sandbox and Idempotency Checks

Payment integration bugs rarely surface during manual testing, with 63% of async API failures linked to race conditions that synchronous tools cannot detect. Afriex's sandbox environment, accessible via business.afriex.com, allows developers to test integrations immediately using a staging API key without any approval process. A key focus of robust testing is idempotency verification, which ensures that duplicate requests produce only one transaction rather than relying solely on HTTP response codes. Developers are advised to confirm that repeated calls with the same idempotency key return the same transaction ID and that only one record exists in the system. Webhook simulation is also recommended to test out-of-order event delivery without waiting for real transactions to change state.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer shares 5 fixes for running a WhatsApp AI bot reliably on Windows 24/7

A developer documented five key challenges encountered while self-hosting a WhatsApp AI chatbot continuously on a Windows PC using Node.js and the local LLM tool Ollama. The bot would stop running whenever the terminal was closed, a problem solved by using the process manager PM2 to keep it alive persistently. Repeated QR code authentication was avoided by enabling local session storage through whatsapp-web.js, while sluggish response times were addressed by switching to a smaller, faster language model. Robust error handling was also added to prevent a single failed request from crashing the unattended bot. The setup requires no VPS, paid AI API, or cloud hosting fees, relying entirely on existing hardware to handle inference locally.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agent Navigates Contradictory Pharma Documents to Make Cold-Chain Shipment Calls

A developer working in pharmaceutical cold chain has built an AI disposition agent for a fictional drug company, Ilmenau Therapeutics GmbH, as part of the Sanity Challenge. The agent evaluates temperature-excursion events on biological shipments — such as a 2–8 °C product spending 30 hours at 13 °C — and recommends whether to release, quarantine, or reject the shipment. The core problem it solves is document conflict: real quality binders often contain superseded stability limits, mismatched temperature ceilings, and contradictory carrier terms sitting alongside current effective procedures. Rather than relying on keyword search, the agent uses a content graph that links each product to its current stability profile, each shipping configuration to its qualification report, and each route to its risk assessment, resolving conflicts via a hierarchy rule that favors evidence over derived data and stricter regulation over internal policy. The tool was tested against 17 controlled documents and four held shipments, with results showing that disposition outcomes can flip entirely depending on whether the agent reads the superseded appendix or the governing effective summary.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Inside Archon: How a Workflow Engine Handles the Hard Parts Beyond Graph Sorting

A deep-dive into the Archon workflow engine reveals that topological sorting — the core graph algorithm — accounts for only a small fraction of the codebase, with most complexity lying in error handling, lock management, and process recovery. Every execution passes through four phases: discovery, routing, setup, and node execution, with the engine applying the same code path regardless of how a workflow is triggered. Malformed YAML files are isolated and logged as load errors rather than halting the entire discovery pass, ensuring one broken workflow cannot block others. When no explicit workflow name is provided, a model-based router selects the correct workflow from a structured prompt, with tool use deliberately disabled so it can only return text. Setup alone performs seven tasks before any node runs, including config resolution, run-row creation, path locking, and git-native worktree branching.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agents Threaten Open Source by Cutting Visibility and Maintainer Rewards

AI coding agents now generate functional code in seconds, bypassing open-source repositories entirely and leaving maintainers without the downloads and stars that historically served as their only reward. Matplotlib maintainer Tim Hoffmann noted in February 2026 that AI dramatically lowers the cost of generating code but leaves human reviewers overwhelmed, widening an already strained gap. One developer observed that his Laravel package had roughly 85 installs for every single GitHub star even before agents became widespread, highlighting how thin maintainer incentives already were. Critics of the 'open source is dying' narrative point out that AI models were trained on publicly shared code, documentation, and issue threads, meaning a decline in open-source activity would eventually degrade the quality of AI-generated code itself. If maintainers lose motivation and stop publishing, agents risk recycling outdated patterns and bugs, ultimately harming the same developers who abandoned open source in favor of AI shortcuts.

0 comments Read more at DEV Community
← NewerPage 1038 of 3195Older →