SShortSingh.

Programming

0
ProgrammingDEV Community ·

Why scene.remove() Does Not Free GPU Memory in Three.js and How to Fix It

Removing a mesh from a Three.js scene graph does not release the underlying GPU resources such as vertex buffers, textures, and compiled shaders, which must be freed explicitly. Developers need to call dispose() separately on the geometry, each material, and every texture a material references, since material.dispose() alone does not clear texture uploads. Entire object hierarchies, render targets, post-processing composers, and OrbitControls each require their own dispose() calls to prevent cumulative memory leaks. Browsers typically allow only around 16 WebGL contexts per page, so failing to dispose the renderer itself in single-page apps can eventually cause a blank canvas error. Three.js exposes renderer.info.memory to track active geometries and textures, offering a straightforward way to confirm whether resources are returning to baseline after a scene teardown.

0
ProgrammingDEV Community ·

Default NLP Stop-Word Lists Can Break Voice Smart-Home Commands

A common NLP preprocessing step — stripping stop-words — can silently cripple voice-controlled smart-home systems by discarding critical command words like 'on', 'off', and 'all'. Standard stop-word lists treat these as meaningless noise, but in commands such as 'turn on light two', they carry the entire instruction. When filtered out, 'turn on' and 'turn off' reduce to identical tokens, causing the system to toggle the wrong switch unpredictably. The fix is minimal: simply remove the relevant words from the stop-word set before filtering is applied. The episode highlights a broader engineering caution — default library settings encode assumptions that may not suit every domain, and developers must override them when context demands it.

0
ProgrammingHacker News ·

Echo.ai Removes 1,400 Security Vulnerabilities from NanoClaw Container Images

Echo.ai has published a technical blog post detailing how it eliminated 1,400 CVEs (Common Vulnerabilities and Exposures) found in NanoClaw's container images. The effort was part of a security hardening initiative aimed at reducing the attack surface in containerized software environments. CVEs represent known security flaws that can be exploited if left unpatched in software dependencies. The post provides an under-the-hood look at the methods and tooling used to identify and remove these vulnerabilities. The announcement was shared on Hacker News, where it garnered modest early attention.

0
ProgrammingHacker News ·

Researcher Explores Extreme DRAM Manipulation in New Technical Project

A technical project titled 'Skitter Creek Bath Salts' has been shared on GitHub by user xoreaxeaxeax, known for unconventional low-level computing research. The project appears to explore extreme or unusual manipulation of DRAM (Dynamic Random-Access Memory). The submission was posted to Hacker News under the informal title 'Spaghettifying DRAM,' suggesting a focus on stretching or distorting memory behavior in unexpected ways. At the time of reporting, the post had received 3 points and no comments. Limited publicly available details make it difficult to fully assess the project's scope or findings without reviewing the repository directly.

0
ProgrammingDEV Community ·

Agentic RAG Lets AI Models Control When and How They Retrieve Information

Agentic RAG is an evolution of retrieval-augmented generation (RAG) in which the AI agent — rather than a fixed pipeline — decides when to retrieve data, how to frame the query, and whether the returned results are adequate. Unlike classic RAG, which embeds a user's question, fetches top results once, and passes them directly to the model, agentic RAG treats retrieval as a repeatable tool the model can call multiple times until it has sufficient context. The agent can rewrite queries, judge the relevance of results, switch between different data sources such as vector indexes, SQL databases, or APIs, and retry searches before generating a final answer. This added flexibility improves recall and handles complex, multi-step questions better than a linear pipeline, but comes at the cost of higher latency and less predictable compute usage. Classic RAG remains the more efficient choice for straightforward lookups over a single corpus, and the agentic approach is best reserved for scenarios where retrieval quality genuinely needs dynamic oversight.

0
ProgrammingDEV Community ·

BI Analyst Felipe Gabriel Joins DEV Community to Share Data and Power BI Insights

Felipe Gabriel, a Business Intelligence and Data Analyst with over six years of experience, has published his introductory post on the DEV Community platform. He currently works at the Complexo Industrial e Portuário do Pecém, focusing on Power BI, SQL, Python, and ETL pipelines. Gabriel holds a Microsoft PL-300 certification and plans to share practical knowledge across topics including DAX modeling, data integration, Python automation, and applied AI in analytics. He stated that writing is his preferred way to organize learning and give back to the developer community. Future posts are expected to go deeper into code and real-world data workflows.

0
ProgrammingDEV Community ·

Backboard Open-Sources R-CLI Coding Agent That Outscored All Terminal Bench 2.1 Entries

Backboard has released R-CLI, its AI coding harness, under the MIT license, making it freely available for anyone to run using their own API keys. The tool is built on a recursive engine called RLM, which breaks coding tasks into smaller bounded contexts handled by child agents rather than relying on a single large context window. Using R-CLI with Claude Opus, the company recorded an 85.4 percent score on Terminal Bench 2.1, surpassing the same model's published score of 78.9 percent via Claude Code and exceeding all 17 other listed entries, though the run has not yet been submitted for official judge review. A separate test using Moonshot AI's Kimi K2 model via OpenRouter achieved 91 percent accuracy across 89 tasks at a total cost of just $28.72, suggesting frontier-level coding performance may be achievable at a fraction of typical costs. The repository is available on GitHub, and the project requires only Bun, Git, and ripgrep to build from source.

0
ProgrammingHacker News ·

Heart Aerospace Achieves First Flight of World's Largest Electric Aircraft

Swedish startup Heart Aerospace has successfully completed the maiden flight of what it claims is the world's largest electric aircraft. The milestone marks a significant step forward in the company's efforts to bring electric regional aviation to market. Heart Aerospace has been developing electric planes aimed at short-haul regional routes as a cleaner alternative to conventional aircraft. The successful test flight demonstrates the viability of scaling up electric aviation technology beyond smaller experimental models.

0
ProgrammingDEV Community ·

10 Mintlify Alternatives for API Documentation Teams to Consider in 2026

A developer roundup identifies ten alternatives to Mintlify for API documentation, targeting teams that need tools integrating with OpenAPI, CI/CD pipelines, and testing workflows. The list includes Apidog, which combines API design, testing, mocking, and documentation in a single platform, alongside established options like Swagger UI and Redocly. Tools such as Stoplight and Scalar cater to teams prioritizing API governance and open-source flexibility, while GitBook and ReadMe suit those building full developer portals. Docusaurus, Nextra, and a plain OpenAPI-plus-Markdown approach are highlighted for teams preferring Git-based, code-adjacent documentation. The key differentiator across all options, the author notes, is whether documentation stays synchronized with the live API development process rather than the look of the final output.

0
ProgrammingDEV Community ·

AWS API MCP Server Flaw Let Security Policies Silently Fail on Startup

AWS published a security advisory on July 23, 2026, for CVE-2026-16584, a high-severity vulnerability (CVSS v4.0 score of 7.3) in its AWS API MCP Server, which allows AI assistants to interact with AWS cloud infrastructure. The flaw was discovered and reported by an independent security researcher who specializes in AI agent security boundaries. If the server's policy-enforcement data failed to load during startup, the process would continue running without that data, causing security policy checks to be silently skipped for all subsequent requests. This meant configured deny and gate rules protecting AWS operations would not be enforced for the entire lifetime of the affected process. The vulnerability is a textbook fail-open security issue, where an initialization failure leads to permissive rather than restrictive behavior.

0
ProgrammingHacker News ·

How to Build a Low-Cost Stratum 1 PTP Grandmaster Clock

A technical guide published on opscode.io details how to build a Stratum 1 Precision Time Protocol (PTP) Grandmaster clock on a budget. The project uses a Raspberry Pi Compute Module 4 (CM4) paired with an SR1723U10 board. The build aims to provide highly accurate network time synchronization typically found in enterprise hardware, but at a fraction of the cost. The article has garnered early attention on Hacker News, where it was shared as a practical resource for networking and hobbyist communities.

0
ProgrammingDEV Community ·

How to Build Motion Trail Effects in Godot 4 Using Ghost Sprites

A tutorial published on DEV Community outlines a method for creating motion trail and afterimage effects in the Godot 4 game engine. The technique works by spawning fading ghost copies of a character sprite at timed intervals, each placed in world space so they remain fixed as the character moves. Developers create a self-destructing TrailGhost scene and a Trail node that activates and deactivates around movement events like dashes. The system supports customization options including tint colors, z-index layering, and speed-based activation for genres like racing or bullet-hell games. A ready-made plugin called Saltmire Trail is also mentioned for those who prefer a drop-in solution.

0
ProgrammingDEV Community ·

NuxiPro founder shares onboarding lessons from Product Hunt launch

NuxiPro is a task management tool that automatically archives completed tasks to keep workspaces clean and responsive. During its Product Hunt launch, creator observed that users were dropping off the demo without understanding the product's value. Analytics tool PostHog helped identify where users lost interest, revealing the absence of an onboarding guide as the key problem. The founder addressed this by integrating driver.js, a lightweight library that walks new users through a step-by-step product tour toward a core 'aha moment'. The team is now seeking beta testers for the demo ahead of a planned cloud version release.

0
ProgrammingDEV Community ·

Why Low-Power IoT Devices Need Their Own Diagnostic Framework, Not Server Logic

Remote diagnostics for low-power IoT devices fail when engineers apply server-monitoring principles, which assume stable power, constant connectivity, and cheap bandwidth. Battery-powered sensors that wake every 15 minutes and operate over NB-IoT or LTE-M cannot sustain continuous log streaming without draining batteries and overloading narrow links. A more effective approach collects only high-value diagnostic signals — such as battery voltage, radio quality, firmware version, and error summaries — triggered by events or delivered with routine heartbeats. Implementing a structured event contract with fields like boot_id, observed_at, and reason_code ensures diagnostic data remains consistent across device restarts, firmware updates, and delayed deliveries. The guiding principle is that diagnostics should be designed as part of the power and operations model, collecting data by value rather than by default curiosity.

0
ProgrammingDEV Community ·

Why Structured Outputs Beat Raw JSON When Building AI-Powered Features

Developers building AI-integrated applications often rely on asking models to return plain JSON, but this approach does not guarantee a reliable data contract. A technically valid JSON response can still break application logic if required fields are missing, values fall outside expected enums, or data types are inconsistent. Using JSON Schema alongside structured output features — supported by providers like OpenAI and Google — allows teams to define strict contracts that the model must satisfy. The recommended workflow starts with the consuming code rather than the prompt, designing a schema around what the application actually needs at runtime. Validating model responses at the boundary, similar to how external API data is handled, makes AI-powered features significantly more predictable and production-ready.

0
ProgrammingHacker News ·

MCP Memory: Open-Source Agent Memory Tool Built on SQLite FTS5 and Google OKF

A developer has shared MCP Memory, an open-source project designed to provide fast memory capabilities for AI agents. The tool leverages Google's OKF (Okapi BM25 ranking function) alongside SQLite's FTS5 full-text search extension for efficient retrieval. It was posted as a community showcase on Hacker News, linking to its public GitHub repository. The project is lightweight and aims to improve how AI agents store and recall contextual information. At the time of posting, it had received 7 points and one comment on the platform.

0
ProgrammingDEV Community ·

How to detect and fix variable frame rate video uploads before they break your pipeline

Screen recording tools like OBS, QuickTime, Loom, and phone capture apps produce variable frame rate (VFR) video, which causes audio drift when processed by transcoders that assume a constant frame rate (CFR). The drift is subtle at first but becomes clearly noticeable around the four-minute mark, making it a silent but serious problem for platforms that accept user-uploaded tutorials, demos, or gameplay footage. A two-tier detection approach using ffprobe can identify VFR files: a cheap header comparison between r_frame_rate and avg_frame_rate serves as a first-pass filter, followed by a per-frame timestamp analysis to confirm irregular gaps. Once a VFR file is confirmed, FFmpeg 8's -fps_mode cfr flag can normalize it at ingest before it causes downstream issues. The article provides Python 3.11 scripts and ffprobe commands to automate both detection tiers efficiently.

0
ProgrammingHacker News ·

AI-Generated 3D Models Saturate Online Marketplaces With Little Buyer Interest

A surge of AI-generated 3D models has flooded digital asset marketplaces, but demand from buyers remains extremely low. Sellers appear to be leveraging AI tools to mass-produce and list 3D assets at scale, hoping to capitalize on the technology. However, the volume of supply has far outpaced actual purchasing activity on these platforms. The trend highlights a growing disconnect between AI-assisted content creation and real-world commercial viability in the 3D asset market.

0
ProgrammingDEV Community ·

How to Build a Scroll-Aware HLS Video Preload Window Under Safari's Element Limit

A developer tutorial on DEV Community explains how to fix stuttering in vertical video feeds by building a directional preload window using hls.js. The approach attaches HLS sources only to off-screen items ahead of the user's scroll direction, while aggressively releasing elements outside the window to avoid crashing on Safari, which caps the number of active video elements. Unlike the browser's native preload attribute, which mobile browsers largely ignore, hls.js gives developers direct control over when segment fetching begins. The solution uses a reconciler pattern of roughly 120 lines, with separate buffer configurations for active and preloading items to avoid burning users' data plans. A small asymmetric window — four items ahead and one behind — is recommended to match real-world scroll behavior.

0
ProgrammingDEV Community ·

Developers build fused video search using visual and transcript embeddings on a shared timeline

A new developer tutorial demonstrates how to build semantic search inside videos by combining visual shot embeddings and transcript text chunks, both anchored to the same video timeline. The system uses shot detection via PySceneDetect, SigLIP-2 embeddings for visual frames, and Whisper for speech transcription, with results stored in a pgvector-backed database. At query time, both indexes are fused so that a natural-language search query returns a precise video timestamp rather than just a video ID. The approach addresses a core limitation where frame-only search misses spoken content and transcript-only search cannot identify visual elements like diagrams or slides. The tutorial provides working Python code for each pipeline stage, including shot boundary detection, midpoint frame extraction, and normalized vector embedding for both image and text inputs.

← NewerPage 160 of 1335Older →