SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

One hardcoded enum list broke a new feature across an entire production app

A software team added a seventh theme, 'birthday,' to a picture-book generator, updating types, forms, sitemaps, and content — but missed a hand-written Zod enum inside the API route that still listed only six themes. Every user request through the front-end returned a 400 error all day, while internal scripts and unit tests bypassed the route entirely, keeping the test suite green. Multiple copies of the same stale six-theme list were scattered across seed files, admin filters, SEO catalogs, and even a test file that defined its own 'valid' themes — meaning the test passed precisely because it never encountered the new value. The fix was to derive a single source-of-truth constant from the existing metadata map and import it everywhere, eliminating the possibility of the API gate drifting from the create form. A lint-style test was also added to flag any file containing three or more hardcoded theme literals in a row, enforcing the pattern going forward.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Multilingual Voice Health Assistant for Underserved Communities in India

A developer built Sehat Sathi, a Hindi-English voice assistant designed to help people in smaller Indian towns access basic health guidance without needing smartphones or English proficiency. The project was completed over 10 days as part of the VoiceForBharat Edition challenge run by Murf AI. The system uses Deepgram for speech recognition, Google Gemini as its reasoning engine, and Murf Falcon's Anisha voice for low-latency, natural-sounding Indian English responses, all connected via LiveKit for real-time audio. Sehat Sathi can identify nearby government hospitals by district, remember returning callers with their consent, and escalate serious symptoms to human support — while explicitly refusing to diagnose conditions or recommend specific medicines. The assistant is designed to give honest, safe guidance on whether a caller should rest at home or seek in-person care, and defaults to directing users to local health workers or emergency services when its own data falls short.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Hindi-English Voice AI for Indian Small Businesses in 10 Days

A developer built Vyapar AI, an AI-powered voice assistant tailored for Indian small business owners, over a 10-day sprint. The system allows users to interact in Hindi, English, or Hinglish through natural speech instead of navigating traditional app interfaces. It uses Deepgram for speech-to-text, a large language model for reasoning, and Murf Falcon's text-to-speech API to deliver a natural-sounding Indian voice. The assistant can remember users, use business tools, make outbound calls, and hand off conversations to specialist agents. The developer encountered several technical hurdles along the way, including API rate limits, telephony configuration issues, and integration challenges.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Duplicate Python Function Silently Ignored Trim Parameters in Solo Developer's Flask App

A solo developer running an AI video editing platform called snipforge.video discovered a bug where GIF trimming failed to respect user-selected start and end times, always converting the entire video instead. The root cause was a duplicate function definition in a 16,000-line single-file Flask app — Python silently overwrites a function name when it is defined twice, and the second, older version accepted the trim parameters but never used them. Because both definitions had identical signatures and the app produced no errors or exceptions, the bug went undetected for some time. The fix was straightforward: delete the shadowing duplicate and retain only the complete implementation that correctly passes trim values to FFmpeg. After the fix, a requested five-second clip dropped from a 48 MB, 74-second job to a 1.9 MB output completed in just six seconds.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why the 'cheapest' AI model may cost more depending on your workload

Choosing the most cost-effective AI model depends not on leaderboard rankings but on the ratio of input to output tokens a specific workload generates. Models price input and output tokens differently, and that gap varies significantly — for example, Grok 4.3 charges only twice as much for output as input, while others charge five to six times more. A document classification task heavily skewed toward input tokens favors Claude Haiku 4.5 by 17%, whereas a code generation task with far more output tokens makes Grok 4.3 42% cheaper. Reliability also affects true cost: a model with a 20% failure rate that reroutes failed calls to a pricier fallback can erase its apparent savings entirely. Developers are advised to measure their own token ratios across real requests before comparing model prices.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Voice-Based Finance Tutor After Losing Money on a ₹6 Share Trade

A developer built FinEd Saathi, a voice-first financial literacy tutor, after personally losing around ₹50 on a ₹6 share trade due to a poor understanding of brokerage fees and taxes. The project was created in 10 days as part of the VoiceForBharat Edition hackathon under the Financial Services track. FinEd Saathi is designed for Indian market beginners and can explain topics such as stocks, mutual funds, ETFs, IPOs, and bonds through patient, conversational voice lessons. The app supports English, Hindi, and code-mixed speech using Deepgram Nova-3 for transcription, Gemini for teaching logic, LiveKit for real-time audio, and Murf Falcon 2 for low-latency text-to-speech. The project aims to bridge the gap between access to financial information and genuine understanding before users risk real money.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

ThoughtDAG Lets Users Build Editable Context Graphs for LLM Chats

A developer has released ThoughtDAG, an open tool that allows users to visualize and edit the context structure of large language model conversations as a directed acyclic graph (DAG). The project was shared on Hacker News as a community showcase submission. ThoughtDAG aims to give users greater control over how context is organized and passed during LLM interactions. The tool is accessible via a GitHub Pages site, suggesting it is an independent, likely open-source project. At the time of posting, the submission had minimal community engagement with only a few points and no comments.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Open-source tool backtests rent vs. buy decisions across 241 US metros

A developer has released an open-source tool called RunTheNumbers that backtests the rent-versus-buy decision across 241 US metropolitan areas. The project uses pinned historical data to ensure reproducible and consistent results over time. It is hosted on GitHub and allows users to compare the financial outcomes of renting against buying in specific local markets. The tool aims to provide data-driven clarity on a question that is often debated with conflicting anecdotal evidence. It was shared on Hacker News, where it attracted early community attention.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

How Data Types Control Memory, Meaning, and Math in Programming

Data types are fundamental instructions that tell a computer how to interpret binary data, how much memory to allocate, and which operations are valid for a given value. Without a data type, the same sequence of bits — such as 01000001 — could mean the integer 65 or the character 'A' depending on context. Java defines eight primitive types, ranging from 8-bit bytes to 64-bit doubles, each suited for different kinds of values and memory requirements. Choosing the wrong integer type can lead to overflow bugs, as seen when adding 1 to Java's maximum integer value wraps around to the most negative number. Floating-point types like float and double use the IEEE 754 standard to represent decimals across a wide range, but introduce precision trade-offs that programmers must account for.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Game Developer Shares Passion for Unity, Godot, and Indie Game Design

A game developer has introduced themselves on DEV Community, highlighting their work across Unity, Godot, and C++ for performance-focused projects. They describe their approach as iterative and prototype-driven, focusing on refining systems that prove effective. Their areas of deep interest include shaders, procedural generation, and the nuanced quality known as 'game feel.' Outside of building games, they actively play indie titles and study the design decisions behind them. They expressed openness to connecting with others in the game development community to discuss tools, engines, and mechanics.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

macOS Screen Sharing Flaw CVE-2026-65400 Lets Attackers Gain Root, Install Crypto Miner

A critical authentication bypass vulnerability, CVE-2026-65400, has been discovered in macOS Screen Sharing, affecting macOS Tahoe before 26.6.1, Sequoia before 15.7.9, and Sonoma before 14.8.9. Attackers can connect remotely via TCP port 5900 and exploit an authentication state management flaw to establish a session without valid credentials, requiring no user interaction. In multiple incidents reported to NCSC-NL, threat actors leveraged this access to gain root privileges and install a Monero cryptocurrency miner. Apple has released patched versions of all three affected macOS branches to address the vulnerability. Users and administrators are advised to apply updates immediately, disable unnecessary Screen Sharing, and block TCP/5900 from public internet access.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

GeoServer Zero-Day SQL Injection Draws Mass Probes Within Hours of Disclosure

A critical SQL injection vulnerability in GeoServer's jsonArrayContains filter was publicly disclosed on August 14, 2026, allowing attackers to inject malicious arguments into backend database queries on PostGIS, Oracle JDBC, and H2 data stores. Security firm WatchTowr observed hundreds of exploit attempts from a small number of source IPs within hours of the public disclosure. The flaw can potentially escalate to remote code execution depending on the server's database configuration, though RCE has not been confirmed across all affected deployments. No patches were available at the time of publication, and no successful compromises have been publicly confirmed despite the volume of probing activity. Administrators are advised to isolate GeoServer from public internet access, apply WAF rules to block malformed filter inputs, and enforce database least-privilege principles until an official fix is released.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AmnesiaStealer Malware Uses Fake GitHub Pages to Hijack macOS Browser Sessions

Jamf Threat Labs has identified AmnesiaStealer, a multi-stage Rust-based infostealer targeting macOS users, published in a report dated August 13, 2026. The malware spreads through a fake GitHub ClickFix page that tricks users into pasting a Terminal command, which downloads and executes a malicious shell script. Once installed, it steals login passwords, Keychain data, Apple Notes, Telegram content, Chromium browser profiles, and cryptocurrency wallet information, sending it all to a remote command-and-control server. A secondary module called stream_module uses Chrome DevTools Protocol to silently take over live browser sessions, enabling remote control of tabs, cookies, mouse, and keyboard input. The malware also attempts to establish persistence by creating a root LaunchDaemon, though its exploit of CVE-2020-9771 for TCC bypass was found to fail on current macOS 26 without Full Disk Access.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Critical SAP Commerce Cloud Flaw Under Active Exploit Three Days After Patch Release

A critical unauthenticated remote code execution vulnerability, CVE-2026-58231, affecting SAP Commerce Cloud's Data Hub Adapter was patched by SAP on August 14, 2026. Attackers began attempting to exploit the flaw just three days after the patch was released, with activity detected by the Defused honeypot. The vulnerability allows external attackers to abuse a default authentication client and bypass input validation in the Data Hub Adapter, potentially enabling arbitrary code execution without any user interaction. SAP has issued Security Note 3771065 with a fix, and administrators are urged to update, restrict Data Hub Adapter access to trusted IPs, and isolate the component from the internet. Over 4,200 IPs flagged by Shadowserver represent product fingerprints rather than confirmed compromised or unpatched systems.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AWS Explains Elastic Beanstalk, Lightsail, and Lambda for Managed Compute

AWS offers several managed and serverless compute options that reduce the operational burden compared to raw EC2 management. Elastic Beanstalk is a PaaS layer built on EC2 that automatically provisions infrastructure when developers upload their code, while still allowing direct server access. Lightsail provides simplified, flat-rate virtual private servers pre-loaded with popular software stacks, designed for small businesses and non-technical users who need a straightforward hosting solution. AWS Lambda is a serverless compute service that runs code in response to triggers without any server provisioning, charging only for actual execution time rather than idle capacity. Each service targets a different level of abstraction and control, making the right choice dependent on a team's scaling needs, technical expertise, and operational requirements.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Writer Discovers His Online Doppelganger Is Entirely Fictional

A person named Sean Byrne investigated another individual sharing his name online, only to conclude that this second Sean Byrne does not actually exist. The piece, published on his personal website, details the process of researching and verifying the identity of someone with an identical name. The investigation raises questions about fabricated online personas and digital identity. The article garnered modest attention on Hacker News, receiving 8 points and 2 comments.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Semantic Caching Can Cut LLM API Costs by Up to 70% Using Embeddings

Semantic caching is an optimization technique for LLM-powered applications that uses vector embeddings to detect meaningfully similar queries, rather than relying on exact string matches like traditional caching. By recognizing that questions such as 'What is your refund policy?' and 'Can I get my money back?' are functionally identical, the system can serve cached responses instead of making redundant API calls. This approach can reduce LLM API costs by 30–70% and cut response latency from seconds to milliseconds without requiring changes to prompts or model configuration. In production environments, vector databases such as Redis with vector search, Pinecone, Qdrant, or pgvector are recommended to ensure the cache scales reliably. The technique is considered especially effective for applications where users frequently ask the same questions in different ways.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Connection Pooling Explained: Why Apps Reuse Database Connections Instead of Recreating Them

Every time an application queries a database, it must establish a secure connection through authentication and memory allocation, a process that can take 50–100 milliseconds. Recreating this connection for every user request is slow and resource-intensive, making it impractical at scale. Connection pooling solves this by maintaining a fixed set of open, reusable connections — for example, capping at 20 connections even when 1,000 users send requests simultaneously. Queries routed through pooled connections execute in just 1–2 milliseconds, since the costly setup work is already done. Without pooling, a sudden traffic spike can exhaust database memory and trigger a full system crash.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Build a Lightweight Delivery Failure Alert System Using Node.js and Lambda

A logistics team can detect delivery notification failures using two small, independently deployed programs — a Node.js metrics publisher and a separate scheduled poller, such as an AWS Lambda function. The two components share an explicit contract: for each closed time window, the poller receives an event time, delivery attempt count, and failure count. Keeping the alert evaluator in a separate deployment domain from the service it monitors prevents a rollback of either component from silently replaying old incidents or misreading stored state. The design also distinguishes between zero failures, zero attempts, and missing data, treating each as a distinct signal to avoid false assurances. This polling-based approach is recommended only when the alert surface is small, a few minutes of detection latency is acceptable, and a reliable metrics endpoint and webhook destination are already in place.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Qwen 3.8 27B Language Model Now Available on Hugging Face for Developers

Qwen 3.8 27B, a large language model with 27 billion parameters, has been released on the Hugging Face platform as an update to the Qwen model series. The model is designed for natural language processing tasks including text generation, language translation, and sentiment analysis. Developers can access it via the Hugging Face Transformers library using Python, making integration into existing workflows relatively straightforward. Practical applications include chatbot development, automated content creation, and customer support systems. However, deploying the model demands substantial computational resources, and cloud-based tools or containerization solutions like Docker are recommended to manage infrastructure requirements.

0 comments Read more at DEV Community
← NewerPage 85 of 1300Older →