SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

BIO Baochip Explained: How FIFOs and Events Enable Host-Processor Communication

The BIO (Baochip/Dabao) processor features four FIFOs accessible as registers x16–x19 on the BIO side and via BIO_BDMA_SFR_TXFx/RXFx on the host side, enabling bidirectional data passing. BIO-side reads block until data is available, while host-side reads always return the last known value without blocking, which can cause race conditions in certain use cases. To handle collisions, separate FIFOs can be used for input and output, and the BIO_BDMA_SFR_FLEVEL register allows checking how many items are queued. A shared 32-bit events register lets BIO processors signal the host by setting or clearing bits, with the host able to poll BIO_BDMA_SFR_EVENT_STATUS for status updates. Together, FIFOs and events provide a synchronization mechanism for coordinating tasks between the BIO cores and the host processor.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Cloudflare Agents Week Highlights Shift Toward Production-Ready AI Agent Infrastructure

Cloudflare held its Agents Week from August 3–7, 2026, releasing a series of infrastructure primitives aimed at developers building production AI agents. The announcements covered runtime environments, identity and access controls, payment handling for agents, and real-time observability tools. The event coincided with similar moves from Amazon Bedrock and LangChain, both of which launched or expanded managed agent runtime offerings around the same period. The broader industry pattern points to a market split between agent frameworks that developers control and the underlying infrastructure they no longer want to manage themselves. Analysts and developers are advised to focus on durable execution, identity scoping, and policy layers rather than committing to any single vendor's full stack.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How a 1913 Russian Math Experiment Became the Foundation of Google and Modern AI

In January 1913, Russian mathematician Andrey Markov presented a probability model to the Imperial Academy of Sciences in St. Petersburg, derived from analyzing letter sequences in Pushkin's poem Eugene Onegin. His framework, later called a Markov Chain, predicts future states based solely on the current state, ignoring all prior history — a principle known as the memoryless property. Decades later, in 1998, Stanford PhD students Larry Page and Sergey Brin applied this concept to build PageRank, the algorithm that powered Google's search engine by ranking websites according to the probability a random user would land on them. The same mathematical foundation underpins many modern AI and machine learning systems used today. The article, authored with AI assistance and published on DEV Community in August 2026, traces this 113-year journey from a poetry-counting experiment to world-changing technology.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Build a GDPR-Compliant Audio Moderation Pipeline for EU Marketplaces

A developer guide on DEV Community outlines a framework for building GDPR-compliant audio moderation systems for EU-based marketplace apps. The approach prioritizes selecting a speech-to-text API based on data residency, retention controls, and provider portability rather than model quality alone. The recommended architecture uses an internal transcription adapter that isolates vendor-specific logic, ensuring moderation and review workflows remain independent of any single provider. Audio, transcripts, and audit records are each assigned explicit retention rules to avoid accidental data handling defaults. A TypeScript code example demonstrates the pattern without relying on any commercial SDK, treating the transcription endpoint as swappable configuration.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Builds Browser-Based Hash Calculator Using Web Crypto API and Vanilla JS

A developer built a client-side file integrity tool after a silent file-upload failure revealed corruption during transfer, with no trusted hash utility on hand. The tool leverages the browser's native Web Crypto API for SHA-256, SHA-384, and SHA-512, while a custom vanilla JavaScript implementation covers MD5, which the API deliberately excludes due to security concerns. To handle large files without crashing the browser, the tool reads data in 1MB chunks rather than loading entire files into memory at once. A debounce mechanism and case-insensitive hash comparison feature were added to improve usability and prevent performance issues during rapid input. The project was built with AI assistance, though the developer notes that human review remained essential to catch algorithmic errors the AI introduced.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Inside Java .class Files: How JVM Bytecode Structure Works

A technical deep-dive explains the internal binary structure of Java .class files generated by the javac compiler. Every valid .class file begins with the magic number CAFEBABE, followed by version bytes that identify the Java release — for example, major version 52 corresponds to Java 8. The article details how the JVM requires big-endian byte ordering for multi-byte numbers, which differs from the little-endian format used by most modern processors. A central component called the constant pool stores all program constants — including class names, method names, and type descriptors — referenced by the bytecode. The piece is part of a series building a Brainfuck-to-JVM bytecode compiler in Node.js, and uses xxd to inspect raw bytes of a minimal Hello.class file.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Supply Chain Attack Compromised 100,000 WordPress Sites via Poisoned Cloud Storage

Over 100,000 WordPress sites were simultaneously compromised in a coordinated supply chain attack targeting a DigitalOcean Spaces bucket used by BdThemes plugins' BigOpti component. Attackers gained write access to this third-party cloud storage and served malicious JSON responses containing an unsanitized display_id parameter, which executed XSS payloads the moment any logged-in administrator opened their dashboard. A script called w2.js then automated three actions: connecting to command-and-control servers, creating rogue admin accounts, and planting hidden web shells in WordPress directories. A secondary payload used a Base36 hashing algorithm to derive valid admin credentials deterministically from a site's hostname alone, eliminating the need for attackers to maintain a database of compromised targets. Traditional File Integrity Monitoring tools flagged no issues because the core plugin files on disk remained unmodified — the malicious code was delivered dynamically over the network at runtime.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

ADK 2.x Drops Need for Monkey Patch in Gemini Live Biometric Scanner Project

A developer building a biometric security scanner with the Gemini Live API removed a 187-line monkey patch file after upgrading from google-adk 1.x to version 2.6.3. The patch had been necessary because Gemini 3.1 deprecated the media_chunks field that older ADK versions used to send real-time audio and video, causing the two to be incompatible. The scanner works by streaming webcam and microphone data from a browser to a FastAPI backend via WebSocket, which then forwards the feed to Gemini 3.1 Flash Live to detect finger gestures and trigger registered tools. ADK 2.x now handles media routing natively, detecting the model generation and dispatching audio and video blobs using the correct typed keywords. The project runs locally or on Cloud Run, and both the original and updated versions are available side by side in the public repository.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Budget Homelab in Japan 2026: Best Mini PCs, Setup Guide and 20+ Self-Hosted Services

A comprehensive guide published on DEV Community helps Japan-based users build an affordable home server using budget second-hand mini PCs such as the ThinkCentre M720q, OptiPlex Micro, and Raspberry Pi. The guide covers minimum hardware requirements — including an 8th-gen Intel CPU, 16 GB RAM, and an SSD — and recommends keeping power consumption under 20 watts for cost efficiency. It walks readers through two installation paths: Ubuntu Server 24.04 LTS for a straightforward setup, and Proxmox VE for those wanting virtualisation. Over 20 self-hosted services are listed alongside Docker-based deployment instructions and a step-by-step Cloudflare Tunnel configuration to expose services without opening router ports. Reference prices are provided for Japanese second-hand marketplaces such as Mercari, Yahoo Auctions, and Hard Off, with a caution that costs vary by condition and specs.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Build an Offline AI Health Analyst on a MacBook Using Llama-3 and MLX

A new tutorial demonstrates how Mac users can analyze their Apple Health data entirely offline using Meta's Llama-3 language model and Apple's MLX framework, eliminating the need to upload sensitive health information to the cloud. The setup runs a 4-bit quantized Llama-3-8B model locally on Apple Silicon chips, taking advantage of the Unified Memory Architecture for fast on-device inference. Users first export their health data from the Apple Health app, then parse metrics such as heart rate and step count from the XML file using Python and Pandas. The processed data is fed into the local model, which generates actionable health insights without any data leaving the device. The approach requires a Mac with an M1, M2, or M3 chip and a few open-source Python libraries to get started.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Claude Code 2.1.229 adds SSE keepalive pings to fix gateway stream timeouts

Anthropic released Claude Code version 2.1.229 on August 12 UTC, introducing SSE keepalive pings designed to prevent idle-timeout disconnections during extended model-thinking pauses. The fix targets gateway streaming responses routed through Google Vertex AI and Amazon Bedrock upstreams. Teams using enterprise LLM gateways or direct provider connections are advised to upgrade the Claude Code client and verify the fix by testing a workload that previously failed with a stream idle timeout error. Anthropic clarified that the patch addresses only streaming liveness and does not resolve unrelated issues such as expired credentials, provider quotas, or buffering reverse proxies. Platform and developer-experience teams are cautioned to test one variable at a time to accurately attribute any improvements in streaming reliability.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Kanban Was Born in Japan, Yet Japanese Offices Still Swear by Excel

Toyota engineer Taiichi Ohno developed the Kanban system in the late 1940s and 1950s, drawing inspiration from American supermarket restocking methods to create a visual, pull-based workflow that minimised waste on factory floors. The concept was later adopted by the global tech industry and became a cornerstone of modern agile project management. Despite this Japanese origin, many traditional Japanese workplaces continue to rely heavily on Microsoft Excel for nearly all organisational tasks, from scheduling to project tracking. Excel's unstructured, rules-free flexibility makes it a deeply entrenched cultural tool that resists replacement by more opinionated systems like Kanban boards. A firsthand account from an English school in Japan illustrates this tension, where a sprawling, overloaded Excel file managed all student and staff data until the system began to struggle under its own weight.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Handle Speech-to-Text API Timeouts for Large Audio Uploads in EdTech

Developers building edtech support queues face a critical design challenge when handling large audio file transcriptions via speech-to-text APIs. Rather than relying on a single API call, the recommended approach treats transcription as a multi-stage ingestion workflow with distinct checkpoints for file acceptance, upload completion, transcription result, and output quality. Each stage requires its own state and timestamp to avoid ambiguous failures, such as duplicate transcripts caused by retrying with a new job ID after an unclear client-side timeout. An idempotency policy keyed to a stable recording ID is essential to prevent duplicate work from reaching downstream classifiers. File-size limits should be treated as application-level policy decisions, with rejected recordings retained locally and routed through an approved asynchronous path rather than silently compressed.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Build a Reliable Semantic Search RAG System for SaaS Support Tickets

A property-management SaaS can convert support tickets into structured, cited answers by using portable model contracts for embeddings and chat completions while keeping retrieval logic within the application. The core pipeline involves chunking support documents, generating embeddings, retrieving the closest matches, and prompting a chat model to answer strictly from those passages — with the application owning citation IDs and output schema. To prevent duplicate processing, each ticket revision should carry a deterministic key and the database write must be idempotent, mirroring discipline used in financial ledger systems. Output schemas should stay narrow — using enums for categories, a boolean escalation flag, and an array of chunk IDs — with the application mapping validated results to queues rather than letting the model define workflow actions. Vendors must be evaluated against applicable privacy regulations such as GDPR and US state laws, as a technically valid JSON response does not itself establish legal compliance.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

AI Governance Explained: Key Concepts and Steps to Put It Into Practice

AI governance refers to the frameworks, policies, and processes organizations use to ensure artificial intelligence systems are developed and deployed responsibly. It covers areas such as accountability, transparency, fairness, and compliance with emerging regulations. Operationalizing AI governance means embedding these principles into day-to-day workflows, tools, and decision-making structures. Organizations are increasingly urged to assign clear ownership of AI risks and establish monitoring mechanisms for deployed models. As AI adoption grows, structured governance is seen as essential for managing legal, ethical, and reputational risks.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Path Traversal Attacks: How Hackers Exploit Directory Flaws to Access Server Files

Path traversal, also known as directory traversal, is a web security vulnerability that allows attackers to read or write arbitrary files on a server running an application. By manipulating file path parameters with sequences like "../", an attacker can navigate outside the intended directory and access sensitive files such as credentials or operating system data. In severe cases, write access can enable attackers to modify application behavior and gain full control of the server. While many applications implement defenses such as stripping traversal sequences or enforcing base directory rules, these protections can often be bypassed using techniques like URL encoding, nested sequences, or absolute path references. Both Unix and Windows systems are susceptible, making path traversal a critical vulnerability that developers must address with robust input validation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How PDF Chunking and Metadata Design Improves B2B Catalog Semantic Search

A developer building a Node.js RAG service for B2B catalog PDFs outlined an architecture that prioritizes search quality over ingestion speed by processing chunking and embedding asynchronously after upload. The core challenge is that product details such as names, dimensions, and compatibility notes are often scattered across headings and pages, making coherent retrieval difficult without careful structure. The proposed design hashes uploaded files, writes immutable document versions, and only exposes catalog data to search queries once ingestion is fully complete, preventing partial or conflicting results. Metadata is deliberately split into two roles: filter metadata to narrow search candidates, and evidence metadata to support citations and audits after results are returned. The author argues that the most durable part of the system is not the parser or language choice, but the evidence record schema that enables retries, reconciliation, and traceable answers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Missing DMARC record was silently blocking registration emails on Gmail and Outlook

Registration confirmation emails from the domain wpmm.jp were failing to reach users on Gmail and Outlook, landing in spam or going missing entirely. Investigation revealed the domain had SPF and DKIM configured but lacked a DMARC record, which major providers like Gmail factor into mail delivery decisions. A DMARC TXT record with a permissive 'p=none' policy was added to DNS, enabling aggregate reporting without risking legitimate mail being blocked. Alongside this, the sending code was updated to properly handle failed send attempts and include required headers such as Message-ID and Date, which are known spam signals when absent. Following the fix, Google began delivering aggregate DMARC reports confirming legitimate mail passes DKIM alignment, giving the team active visibility into email deliverability.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Essential Docker Commands Every Developer Should Know for Daily Use

A practical guide to core Docker commands covers daily essentials such as docker run, exec, logs, ps, and build for working with real containers beyond basic examples. Key flags for docker run include -it for interactive shell access, --rm to auto-remove stopped containers, -e for environment variables, and -v for bind-mounting host directories. The docker ps and docker inspect commands help monitor running containers, while docker stats enables real-time CPU and memory tracking without entering a container. Docker logs, particularly with the -f flag, is highlighted as the go-to first step when diagnosing misbehaving containers, since applications are expected to write output to stdout or stderr. The docker exec command allows developers to run commands or open a shell inside an already-running container, making it useful for debugging in production or staging environments without restarting services.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Product Builder: The Hybrid PM-Developer Role Gaining Traction in 2025

A new job role called 'AI Product Builder' is emerging in tech hiring, describing a single professional who handles both product management and software development using AI coding tools. The role is made possible by advances in AI code assistants, better context and prompt management, and automated quality checks that reduce the traditional idea-to-deployment cycle. However, its success depends heavily on the maturity of the existing codebase, including documentation quality, extensibility, and automated testing coverage. Organisations with well-structured, documented codebases are better positioned to let such hybrid roles thrive, while those with younger or messier code should limit the scope of tasks assigned to these individuals. Demand for the role is expected to grow, but companies must invest in the right engineering foundations to make it viable.

0 comments Read more at DEV Community
← NewerPage 174 of 1337Older →