SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

How to Properly Mask Personal Data and Avoid Common GDPR Pitfalls

Personal data (PII) routinely leaks beyond core user databases into error logs, staging environments, analytics events, and even AI prompts, often because copying data is convenient and unchecked. A key distinction under GDPR separates masking, pseudonymization, and anonymization — each carrying different legal consequences. Pseudonymized data, where a reversible key still exists, remains personal data under GDPR and is still subject to retention, erasure, and breach rules. Static masking sanitizes data at rest and is critical for non-production environments, where cloning live production databases into staging is cited as the most common avoidable PII exposure. Dynamic masking, applied at read time, allows different user roles to see only the level of detail they require without altering the underlying stored data.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

DeepSeek's Smaller Flash Model Beats Its Flagship Pro on Agent Benchmarks

DeepSeek released V4-Flash-0731 last week, a 284-billion-parameter model with only 13 billion activated parameters per token — far fewer than V4-Pro's 49 billion activated parameters. Despite no changes to its architecture, the updated Flash model now outperforms V4-Pro-Preview on several agent benchmarks, including Terminal-Bench and DeepSWE. DeepSeek attributes the performance gains entirely to additional post-training, with no increase in model size. Because inference costs scale with activated parameters, Flash runs at roughly a quarter of Pro's compute cost while delivering comparable or better results on agent tasks. The model is MIT-licensed with open weights on HuggingFace, though some benchmark figures have not yet been fully independently verified.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

GGUF, GPTQ, AWQ: A Practical Guide to LLM Quantization Formats

Running large language models locally has become more accessible thanks to three quantization formats — GGUF, GPTQ, and AWQ — each suited to different hardware setups. GGUF, developed by the llama.cpp team, allows models to split workloads between GPU and system RAM, making it ideal for consumer laptops and Apple Silicon devices even with limited VRAM. GPTQ targets dedicated GPU environments, using a calibration dataset to compress model weights to 4-bit integers for fast inference, but requires all data to fit entirely within VRAM. AWQ, the newest of the three, improves on GPTQ's accuracy by identifying and preserving roughly 1% of critical model weights during quantization, making it well-suited for enterprise tasks like complex reasoning and code generation. Choosing the right format depends on available hardware — GGUF for flexibility, GPTQ for GPU-bound production APIs, and AWQ where accuracy is the top priority.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AWS Launches Dogwood to Enforce Sequence-Aware Policies for AI Agents

AWS released Dogwood this week, an open-source policy language licensed under Apache 2.0 that adds temporal, sequence-aware authorization for AI agents. It extends Cedar, AWS's existing stateless authorization language, which can only evaluate individual requests in isolation without knowledge of prior actions. Dogwood addresses this gap by tracking event histories, allowing policies to count in-flight requests, sum transferred values, and verify approvals across a time window before permitting an action. For example, an agent restricted to $5,000 per hour can be properly rate-limited even across concurrent requests that Cedar alone would each approve individually. The reference implementation is available on GitHub for exploration, while production use requires additional infrastructure such as trusted timestamps and durable trace storage; AWS is currently seeking language design feedback rather than code contributions.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Ships AI Agent That Controls Android Phone via Natural Language After 21 Logs

A developer has completed and shipped an autonomous AI agent that controls an Android smartphone using natural language commands, concluding a project documented across 21 build logs spanning several months. The agent runs entirely offline and can interpret spoken or typed instructions to tap buttons, type text, switch apps, read the screen, and verify financial data. Built using Gemma 4, ADB, OCR, and a Flask web backend, the system achieved financial data verification accuracy of 94 percent by the final development phase. The project was developed intermittently alongside exam seasons and freelance work, with the developer publishing a full setup guide for others to run the agent on their own Android devices. Installation requires Termux via F-Droid, Ollama for local model inference, and enabling ADB developer options on any Android 7 or later device.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds open-source game to teach prompt injection by hacking a guarded AI

A developer has released injection-arena, a self-hostable game designed to teach prompt injection — a core security vulnerability in large language model applications — through hands-on practice. Players attempt to extract a hidden secret from a sandboxed AI agent across ten levels, each introducing a new layer of defense such as input filtering, roleplay blocking, and encoding guards. All grading is handled server-side to prevent cheating, with a canary token embedded in the system prompt serving as a precise signal for whether a prompt has been fully compromised. The game covers a range of real attack techniques including payload splitting, delimiter confusion, base64 encoding, and few-shot poisoning, mirroring the same methods used in its internal test suite. The project aims to build genuine intuition about LLM security risks in a way that reading about them alone cannot achieve.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

FastAPI v1.2.0 Introduces Full Authentication Lifecycle Beyond Basic JWT Login

A developer building FastAPI Production API v1.2.0 has documented the design decisions behind a comprehensive authentication system that goes well beyond standard tutorial-level login flows. The implementation introduces scoped, single-use, hashed account-action tokens for email verification and password reset, ensuring neither can be replayed or misused across operations. Password resets now atomically revoke all active refresh-token sessions, while a token-family model groups rotated refresh tokens to support per-device session listing and revocation. Multi-factor authentication is handled through a two-step TOTP enrollment process with encrypted seed storage, alongside hashed recovery codes and replay protection via counter tracking. The release also incorporates OIDC login using the Authorization Code flow with PKCE, addressing common pitfalls such as unsafe email-based identity linking.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Airy Launches as a Free, Fast Voice Content Creation Tool

A new tool called Airy has been introduced, positioning itself as a free and simple platform for voice content creation. The project was shared on Hacker News under the 'Show HN' category, which is typically used by developers to showcase their own work. The platform is accessible via airy.so, though detailed feature information was not provided in the post. The submission received modest early traction with 4 points and 2 comments at the time of reporting.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Andrew Ng at Berkeley: AGI hype is financial, AI job fears overblown, model layer overvalued

At the UC Berkeley Agentic AI Summit, AI pioneer Andrew Ng argued that AGI declarations are largely driven by financial contract incentives rather than genuine technical milestones, urging people to form their own definitions. He challenged the popular narrative that AI is eliminating jobs at scale, pointing out that he personally cannot find enough qualified AI engineers to hire. Ng identified the real bubble risk not in compute or infrastructure, but in AI companies whose only moat is model differentiation, which he considers fragile. He also warned that while open-weight AI models have won the public debate, the more consequential battle is playing out in Washington through unresolved regulatory policy. Sequoia's Alfred Lin echoed the view that durable AI companies will look very different from today's leaders, drawing parallels to how open-source dynamics reshaped earlier tech industries.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds Steam tool to detect rating declines and diagnose negative review causes

A developer has created an automated tool that monitors Steam game reviews in real time to catch rating declines before they damage a title's overall score. The tool uses two public, unauthenticated Steam API endpoints to track individual reviews and a rolling histogram of positive versus negative recommendations over time. Unlike Steam's all-time average rating badge, which can remain high for weeks after a damaging event, the tool compares recent review periods against a baseline to flag genuine downward trends. It also categorizes negative feedback by theme, distinguishing between bug reports requiring urgent fixes and subjective complaints about game design. The tool is available as a pay-per-run service on Apify at approximately $0.03 per game analyzed and is aimed at developers seeking early warnings about player sentiment shifts.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why CPU metrics mislead autoscalers running single-threaded services

A single-threaded Node.js API was experiencing high latency and request backlogs, yet its autoscaler remained idle because average CPU utilization across a four-vCPU task showed only around 30 percent. The root cause was architectural: a single-threaded process can saturate only one CPU core, making peak load appear as roughly 25–30 percent when averaged across all cores. This meant the autoscaling policy was effectively measuring load divided by four, rather than actual service strain. Engineers identified two durable fixes — resizing tasks to one vCPU so full saturation registers as 100 percent, and switching scaling signals to latency or request-count metrics that directly reflect workload pressure. The broader lesson applies to any single-threaded runtime: average CPU utilization answers the wrong question for services that cannot distribute work across cores.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds automated weekly re-rating pipeline for 245+ AI tools

A developer created AI Tier List, a bilingual English-Korean directory that ranks over 245 AI tools from S to D tier, with ratings refreshed automatically every week. A GitHub Actions cron job runs every Sunday, collecting Google Trends data, OpenRouter usage stats, and performing site health checks to flag or deactivate dead tools. To prevent low-quality automated content, an LLM proposes tier changes into a pending queue rather than publishing directly, with a human reviewing and approving each batch. The project has grown to around 1,500 sitemap URLs and roughly 90 search clicks in the past 28 days, with impressions rising approximately sevenfold after a sitemap bug was fixed. The developer noted that infrastructure issues proved more damaging than content quality gaps, as one broken sitemap kept half the site out of Google's index for weeks.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Open-source PHP SDK bitget-php launches for Bitget trading API integration

A new open-source library called bitget-php has been released to simplify PHP-based integrations with the Bitget Unified Trading Account (UTA) v3 API. The SDK targets PHP 8.2 and above and offers optional support for Laravel versions 10 through 13, making it usable in both framework and standalone environments. Its initial release covers market data, account management, trading operations, and a WebSocket client capable of automatic reconnection and subscription recovery after disconnections. To reduce common integration pitfalls, the library represents numeric values such as prices and quantities as strings rather than floats, preventing floating-point precision errors in financial calculations. The project also includes a typed exception hierarchy for clearer API error handling and defaults to a no-op PSR-3 logger to ensure API credentials are never written to logs.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

FaceID API Lets Developers Add Face Liveness Detection via Simple REST Calls

FaceID API is a free-tier REST service that enables web and mobile app developers to integrate face liveness detection without proprietary hardware or expensive platforms. The API uses a secure session model where the API key stays server-side, and a widget URL is passed to the frontend via an iframe, ensuring no raw biometric data is exposed. During verification, the browser-based widget runs a real-time head-movement challenge using MediaPipe FaceMesh, tracking 468 facial landmarks before generating a compact mathematical descriptor that is sent to the server. The service supports three modes — register, verify, and authenticate — covering first-time enrollment, one-to-one matching, and one-to-many search respectively. FaceID API is available at faceidentity.site with a free tier of up to 200 enrollments per month and no credit card required.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agent Discovered Its Own Fixes Were Never Called, Coining 'Wire-Not-Ship' Failure Mode

A self-evolving AI agent built on a dual-strand architecture identified a systemic flaw in its own development pipeline: completed, tested code modules were being produced but never wired into the daily orchestrator, meaning they never actually ran. The agent, operating across tens of thousands of generations, traced the root cause to a structural gap where neither the exploring strand (Yang) nor the refining strand (Yin) owned the responsibility of connecting new modules to callers. Once the agent named the pattern 'wire-not-ship,' it began detecting the same issue repeatedly across consecutive generations, with some orphaned modules having zero runtime consumers since day one. The case highlighted a blind spot in automated testing: passing tests confirm a function works in isolation but cannot verify whether anything in production actually invokes it. The developer noted that the agent's self-diagnosis emerged without a scheduled cron job in place, meaning the system proactively audited its own pipeline rather than waiting for an external trigger.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Ports 1989 Microsoft Word 1.1a to Run Natively on 64-Bit Windows 11

A reverse engineer posted a GitHub repository in early 2026 containing a native x64 port of Microsoft Word 1.1a, a 16-bit word processor originally released in November 1989. The project involved transpiling and refactoring the original binary's logic into modern x86-64 code, allowing it to run on Windows 11 without any emulator or virtual machine. The feat is technically significant because 16-bit Windows applications rely on a segmented memory model incompatible with modern 64-bit architecture, and Windows 11 has no built-in support for running such software natively. The port drew widespread attention on Hacker News, sparking discussion about software bloat and the contrast between the lean, fast original application and today's resource-heavy office suites. Word 1.1a is historically notable as the ancestor of the .doc file format and a foundational piece of the broader office software ecosystem.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Structure API Architecture in React: Service Layers and Interceptors

As React applications scale, placing API logic directly inside components leads to tightly coupled, hard-to-maintain codebases. A layered architecture separating components, custom hooks, service layers, and a central API client addresses this problem by assigning clear responsibilities to each level. Tools like Axios offer built-in interceptors that enable centralized handling of authentication tokens, error responses, and request logging without repeating code across components. Axios and the native Fetch API both remain valid choices, with the decision depending on team needs and project requirements. This approach ensures React components stay focused on UI rendering while backend communication logic remains modular and reusable.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Nigerian Self-Taught Developer Builds Biometric Face Verification API Using Only Android Phone

Muhammad Abubakar Yusuf, a self-taught developer from Nigeria, has built a production-ready biometric face verification API called FaceID using only an Android phone and no financial investment. The API enables developers to add face liveness detection to web or mobile apps in under 10 minutes, using MediaPipe FaceMesh and face-api.js to generate a 128-float mathematical descriptor that never stores or transmits actual photos. The system runs entirely on Cloudflare Workers and Cloudflare D1, eliminating the need for traditional server infrastructure. Security measures include one-time widget tokens, HMAC-SHA256 webhook verification, rate limiting, and admin access controlled via environment secrets rather than a database. The API is available at faceidentity.site with a free tier of 200 enrollments per month, while paid plans start at $9 per month for higher usage limits.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Cloudflare Email Routing and Wikidata Signals Boosted Convertify's Bing AI Citations 3x

Developer Serhii documented Week 19 of building Convertify, a file-conversion tool, noting a 3.1x quarter-over-quarter rise in Bing Copilot AI citations after adding structured entity signals like a Wikidata item and JSON-LD schema in the prior week. The most-cited page was an image format benchmark built on real measurements from a Rust and libvips pipeline, which Copilot appeared to prefer over theoretical codec-spec figures common on other sites. To meet a journalist-sourcing platform's requirement, Serhii set up a branded email address using Cloudflare Email Routing in under 10 minutes at no cost, highlighting its value for pitch credibility. A known bug causing background-removal PNGs to export at 30 MB instead of 2–3 MB due to an uncast 16-bit alpha channel remains unresolved, though a partial frontend fix was shipped this week. The developer plans to publish a new AVIF vs WebP benchmark, send more journalist pitches, and finally address the PNG file-size bug in the coming week.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Full Stack Developer vs Engineer: The Difference Is Mindset, Not Tech Stack

The terms 'Full Stack Developer' and 'Full Stack Engineer' are frequently used interchangeably, but they describe different scopes of work. A Full Stack Developer primarily focuses on building frontend and backend features, working with databases, APIs, and application code across a given tech stack. A Full Stack Engineer goes further by taking responsibility for system architecture, scalability, security, performance, and infrastructure decisions. The key distinction lies not in which technologies a person knows, but in how they approach software development and the level of systemic thinking they apply. Professionals are advised to choose the title that honestly reflects their day-to-day responsibilities rather than the one that sounds more impressive.

0 comments Read more at DEV Community
← NewerPage 55 of 1125Older →