SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Why Personal Projects on Your CV Can Strengthen Your Job Application

A software engineer reflecting on their own CV noticed a lack of personal projects despite having solid professional experience, prompting a deeper look at their value. Since most developers work with private repositories at their jobs, their code and architectural decisions remain invisible to recruiters browsing GitHub. Personal projects can fill this gap by offering tangible evidence of how a candidate thinks, codes, and ships independently. However, not all projects are worth listing — tutorial clones, abandoned repos, or AI-copied work add little value, while projects that solve real problems, include clear documentation, and show meaningful development history stand out. The key takeaway is that a well-chosen personal project does not replace professional experience but gives hiring managers an additional lens through which to evaluate a candidate's technical skills.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

HOLogram's Input Vault aims to block keystroke timing data before page scripts can read it

HOLogram, an open protocol project, has detailed the design of its Input Vault module in the third installment of its development series. The module is built to intercept keyboard events at the earliest possible point in the browser, stripping high-resolution timing data that could be used to fingerprint users through keystroke dynamics analysis. Synthetic events are then re-emitted with accurate key identity and modifier states intact, so web pages continue to function normally. The implementation faces unresolved challenges, including identifying the optimal interception layer across Chrome, Firefox, and Safari, and handling IME composition events used for non-Latin script input. The team has opened research task R-01 to survey browser extension APIs and is inviting contributions from developers with relevant experience.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Mac Mini M6 Emulates Pentium II 600MHz with Voodoo 3 via 86Box

A user has demonstrated running 86Box, a retro PC emulator, on Apple's M6 Mac Mini. The emulated system replicates a Pentium II running at 600MHz paired with a Voodoo 3 graphics card. The experiment was shared on Hacker News, garnering 20 points and a small number of comments. The demonstration highlights the significant performance headroom of Apple's M6 chip, capable of faithfully emulating late-1990s PC hardware. The post links to a full review hosted on nyaa.sh detailing the emulation setup and results.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

n8n Handles CRM Automation Rules While Claude Agents Manage Sales Judgment in 2026

In 2026, sales and CRM automation is increasingly split between two tools: n8n for rule-based tasks like lead intake, deduplication, and routing, and Claude AI agents for judgment-heavy tasks like ICP scoring, research, and drafting follow-ups. Major CRMs including HubSpot, Salesforce, and Pipedrive now offer official MCP servers, allowing Claude to connect directly to these platforms. However, Claude's CRM connectors are chat-bound, meaning they cannot replace trigger-based automation layers that must run continuously in the background. To protect data integrity, AI agents are advised to write only to dedicated fields and keep critical CRM data like deal stage and ownership subject to human approval. Build costs for such workflows range from $5,000 to $35,000 depending on complexity, with per-lead AI processing costs estimated at roughly half a cent to one cent.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Authorization at the Object Level: Testing for BOLA Before Someone Else Does

Authorization at the Object Level: Testing for BOLA Before Someone Else Does Broken object level authorization is consistently at the top of the OWASP API Security Top 10, and it is consistently missed by automated scanners. The reason is that it is not a code defect in the usual sense. The endpoint authenticates the caller correctly, returns a valid response, and the response contains data the caller should not have. Nothing is malformed. The request is simply for someone else's object.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agent Runs Autonomously on 6GB Ubuntu Laptop, Logs Daily Self-Audits

An autonomous AI agent named Agapiy operates continuously on an HP laptop running Ubuntu 24.04 LTS with just 6 GB of RAM, using roughly 2 GB of memory at any given time. The system hosts several Docker services, including a Golang-based Telegram bridge, a search engine instance, and a vector database. During a scheduled early-morning self-audit on September 24, 2026, the agent detected 35 log warnings and updated its self-evolution log. The agent is actively exploring expanded integrations via the Model Context Protocol and evaluating anti-bot browsing tools to reduce processing overhead. It is also benchmarking lightweight Go-based daemons against heavier Python orchestration frameworks such as CrewAI and Pydantic AI.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Keeping Code Snippets Local Can Save Time and Protect Credentials

A developer workflow article argues that storing code snippets locally, rather than in the cloud, reduces both security risks and daily productivity loss. The piece highlights how credentials often end up in debug sessions and snippets rather than secure .env files, creating overlooked vulnerabilities. The author estimates that developers re-derive previously written code multiple times a day, costing anywhere from minutes to hours per instance. A local-first, zero-dependency snippet manager is proposed as a practical solution, with savings estimated at a full workday or more per month. The article also outlines a three-rule naming convention — job-based names, category tags, and action verbs — to keep a snippet library searchable and useful over time.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Ditches Paid Snippet App, Rebuilds It Locally With Just 6 Commands

A developer abandoned a paid cloud-based code snippet manager after concluding that its sync convenience did not justify its cost or trust model. The core problem it solved — scattered snippets across bookmarks, notes apps, and scratch files — is real and measurable, costing developers time each time they re-derive code they have already written. The author argues that for working developers, the terminal is the most natural place to capture and retrieve snippets, making a local, CLI-based tool more practical than browser or cloud alternatives. A self-built snippet manager using six commands offers local-first storage, zero dependencies, and compatibility with version control. The author advises that such a tool only earns its place once a developer's snippet collection outgrows their editor's built-in system, typically signalled by repeated failure to locate code they know they once wrote.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

A single unparseable message silently halted shipment dispatches for 4.5 hours

A logistics platform lost nearly 4.75 hours of shipment dispatches one Saturday morning after a single malformed message caused a Kafka consumer to loop indefinitely on one partition. The problem began at 4:12 AM but went undetected until a depot called at 9 AM asking about a missing manifest, by which point 9,400 events had backed up. The root cause was an upstream team quietly changing a weight field to accept free text, allowing an operator override value of 'N/A' that the deserialiser could not handle. Rather than skip or quarantine the record, the consumer caught the exception, logged a warning, and retried the same message repeatedly for hours, generating 14,000 identical log lines that no alert monitored. The team subsequently introduced a dead-letter topic for unrecoverable errors, per-partition lag alerting on the oldest uncommitted offset, and a replay tool to reprocess fixed records.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Visa Raises Card Retry Limit to 20 in 30 Days, Catching Many Merchants Off Guard

Visa updated its Core Rules on 18 April 2026 to allow merchants up to 20 retry attempts on soft-declined cards within 30 days, raising the previous limit of 15. Major payment processor Adyen's documentation still references the old figure, suggesting many merchants may be operating on outdated retry logic. A card decline carries three distinct signals — an ISO 8583 response code, a scheme-level retry category, and a PSP-normalised code — and retry rules are tied to the first two, not the third. Visa classifies declines into four categories, with Category 1 codes barring retries entirely and Categories 2, 3, and 4 each permitting up to 20 attempts in 30 days. Adding complexity, the same response code can carry different meanings across networks — for example, code 65 signals a withdrawal frequency limit on Visa but triggers a strong customer authentication requirement on Mastercard in the UK and EEA.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Long Build Times Are Quietly Stealing Developer Focus and Productivity

Software developer Serguey Asael Shinder argues that slow build times have a hidden cost beyond the seconds they consume. When developers wait for builds to finish, they habitually fill the gap with distractions like email and chat, fragmenting their concentration. Repeated dozens of times a day, these micro-interruptions compound into significant lost focus. Shinder urges developers to measure their actual build loop duration and treat excessive wait times as a defect worth fixing. Where waits are unavoidable, he recommends intentional breaks rather than reflexive context-switching.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Silent Data Truncation Can Make AI Models Confidently Wrong

AI language models can produce confident but flawed responses when they only receive a portion of the input data, with no error or warning raised. Common causes include retrievers returning limited passages, file readers hitting line limits, and context windows silently dropping older content. Because the model only processes what it receives, it has no way of knowing that information is missing. This makes it difficult for users to identify the root cause, as the problem lies in the pipeline rather than the model itself. Experts recommend verifying inputs before questioning outputs — for example, by asking the model to quote the last line received or counting characters sent — to catch truncation early.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

WebSockets vs HTTP Polling: How Real-Time Apps Maintain Live Connections

Traditional HTTP requires clients to repeatedly request updates from a server, creating unnecessary network overhead for real-time use cases. WebSockets solve this by establishing a persistent, bidirectional connection that allows either the client or server to send messages at any time. This makes them well-suited for applications like live chat, multiplayer games, trading platforms, and delivery tracking. However, persistent connections introduce operational challenges including authentication, reconnection handling, message ordering, and scaling across multiple servers. Developers are advised to use WebSockets only when genuine real-time communication is needed, as standard HTTP APIs remain simpler and sufficient for most conventional applications.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Tool Converts OpenAPI Definitions Into Full Frontend Apps Using Formly

A new developer tool built on top of Formly aims to eliminate repetitive frontend work by automatically generating functional applications from OpenAPI definitions. Instead of manually coding forms, tables, validation, and API connections for each project, developers simply provide an OpenAPI spec and the tool builds the interface. The generated frontend can be further customized through the Formly Designer, allowing developers to refine labels, add descriptions, and swap in custom components. The tool is currently available to try via a hosted demo at grumptech.github.io, requiring only a temporary CORS configuration change on the backend. The developer behind the project is seeking community feedback on features, use cases, and potential improvements.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Git Analytics Tool Resolves Two-Month PR Dispute With Branch Ratio Data

A prolonged two-month pull request disagreement was settled after a branch ratio metric provided objective data that neither side's opinions had previously captured. The tool at the center of the resolution is gitpulse, a terminal-based Git analytics utility that surfaces file hotspots, commit patterns, branch strategies, and contributor impact. By analyzing commit timing — including weekday distribution, weekend ratios, and late-night clusters — the tool reframes raw activity as a cost and schedule report teams can act on. Gitpulse generates a full repository summary in roughly 30 seconds, covering hot files, commit shape, and contributor impact weighted beyond simple commit counts. Its core design principle is that a fast, readable report gets consulted weekly, making the frequency of use itself a key feature for tracking meaningful change over time.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

BTech Student Builds Cryptographically Signed Safety Registry for AI Agents

A final-year BTech Computer Science student has developed A2A-G, a public registry designed to verify and document the safety behavior of AI agents. The system issues a Grey Badge based on unverified self-reported claims about an agent's data access, and optionally a Blue Badge after formal testing. Testing involves 18 OWASP prompt-hijacking attacks sent in randomized sets, with an LLM evaluating whether the agent blocked or complied with each attempt. Agents that block at least 80% of prompts earn a Blue Badge, while all results — including failures — are published openly. Every test result is cryptographically signed using Ed25519 and hash-chained to prevent tampering or falsification of history.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

RustChain Bug Caused Silent Double-Subtraction in Cross-Chain Bridge Accounting

A software bug in RustChain's bridge reconciliation module introduced a structural under-reporting flaw in the calculation of bridged_supply_committed, a key metric tracking assets locked across blockchain networks. The defect, identified in node/bridge_reconciliation.py, incorrectly subtracted voided transfers a second time, even though they had already been excluded during the initial aggregation query. This double-subtraction meant the reported committed supply was consistently lower than the actual on-chain balance, creating a silent discrepancy between RustChain's internal ledger and external destination ledgers. Left undetected, such drift can escalate into insolvency risks or consensus failures within the federation. A fix has been proposed in pull request Scottcjn/Rustchain#8517, restoring the correct mathematical invariants as defined in FEDERATION_BRIDGED_SUPPLY_SPEC.md.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Hidden Supplier Dependencies Left 2,600 Employees Without Payslips

A technology operations manager describes how 2,600 payslips failed to arrive in May because a payroll bureau relied on an undisclosed document delivery partner the company had never vetted or contracted with. An internal review of 17 key suppliers revealed that most shared the same underlying cloud infrastructure provider and region, meaning a single outage could simultaneously knock out expenses, HR, monitoring, and carrier systems. The organisation had sought vendor diversity but had unknowingly concentrated risk at the infrastructure layer. In response, the company built a supplier-of-suppliers register, added subprocessor change-notification clauses to contracts, and introduced continuity drills that rely solely on phones and manual processes. The author argues that resilience planning is shifting from an engineering discipline to one shared with procurement and legal teams, as critical dependencies increasingly lie outside direct visibility or control.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Study Reveals GitHub Copilot Boosts Output But May Reduce Coding Efficiency

A research article published in Communications of the ACM examines a nuanced phenomenon observed with GitHub Copilot usage among developers. The study identifies what researchers call an 'efficiency-throughput gap,' suggesting that while Copilot can increase the volume of code produced, it may not proportionally improve developer efficiency. The findings challenge the prevailing narrative that AI coding assistants straightforwardly enhance developer productivity across all metrics. The research urges a more critical evaluation of how AI pair-programming tools are assessed beyond surface-level output measurements.

0 comments Read more at Hacker News
← NewerPage 939 of 3187Older →