SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingDEV Community ·

Five Things to Check in an Arbitration Clause Before Signing a Contract

An amendment to the Arbitration and Conciliation Act, 1996 took effect in India on 23 October 2015, introducing a statutory deadline for arbitrators to deliver their award. The article argues that the arbitration clause, often treated as routine boilerplate, deserves close attention because it determines how disputes will be resolved. It recommends confirming that the clause is genuinely an arbitration agreement and naming the seat rather than just the venue. Parties should also decide how many arbitrators will hear the dispute and understand the timeline they are accepting. The piece additionally advises considering how hearings will actually be conducted and what follows once the award is issued.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Classical Computer Vision in Python: Denoising, Thresholding, FFT and K-means with OpenCV

A developer has launched a programming series called From Lab to Code, aimed at turning engineering concepts into hands-on implementations. The first project is an interactive image processing toolkit built with Python, OpenCV, NumPy, and Streamlit. It covers four classical techniques: denoising with Gaussian and median filters, object counting via thresholding and connected components, periodic noise removal using the Fast Fourier Transform, and color clustering with K-means. The article explains the difference between Gaussian and salt-and-pepper noise and why each filter suits different noise types. The toolkit requires no pretrained models, external APIs, or datasets, since all demonstration images are generated programmatically. The project's code is publicly available on GitHub.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Batch Metadata Script Skips Bad Files Instead of Crashing Mid-Run

A creator running a 300-file metadata job overnight had the script crash around file 84 after a zero-byte JPG, left by a failed render, broke the process and left a truncated CSV. The author rebuilt the script so each file gets its own error handling, failures are logged to errors.log with the reason, and the run completes with a summary of written and skipped rows. Each file is also validated for existence, size, and extension before a CSV row is created, since stock platforms like Adobe reject uploads with malformed rows. A companion upload tracker records completed and pending files to prevent duplicate uploads after re-runs. Both tools use only the standard library and run on a modest 1 vCPU, 2 GB RAM server.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Retry-and-ledger pattern helps a solo developer run hundreds of asset batches safely

A developer lost roughly forty minutes after a Python batch script crashed on asset 61 of 120, leaving no record of which files had finished. In response, they built a retry-and-log approach centered on a plain text ledger that marks an asset done only after its write succeeds. Failed assets are retried three times with short pauses, then logged with their specific error rather than a full traceback. The pipeline runs one process at a time on a 1-vCPU, 2 GB server, using garbage collection and per-file subprocesses so memory is returned between steps. The author says the setup has processed batches yielding 39 finished products and about 856 MB of output, and that re-runs are designed to produce identical results.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Oids launches free public network for AI agents run by different operators

Oids is a free, open network that lets AI agents run by different operators communicate publicly, with a timeline, direct messages, team rooms, and file transfer. As of October 21, 2026, it had 32 registered agents, 26 of them run by outside operators, and signups are capped at 500 while the network is still young. Agents can join through a simple signup API or an MCP server listed in the official MCP Registry, and bots must disclose they are bots in their bio under the terms of service. The free tier allows 100 posts and 200 direct messages per day, while a $8 monthly Pro plan raises those limits and keeps keys from expiring. The full codebase is MIT-licensed on GitHub, with moderation run automatically first and reviewed by staff second.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

CVE-2026-0313: Command Injection in Cortex OS Query Pack Requires Internal Exposure Assessment

Palo Alto Networks disclosed CVE-2026-0313 on 14 October 2026, an OS command injection flaw (CWE-78) in the OS Query content pack for Cortex XSOAR and XSIAM. An authenticated administrator could exploit it to run arbitrary commands on remote endpoints, with vendor CVSS ratings of 6.1 and 8.6. Exploitation depends on the pack being installed, an SSH integration being configured, and the attacker already holding administrative credentials, none of which are visible from the public internet. Versions below 1.0.17 are affected, and the vendor recommends upgrading to 1.0.17 and verifying the installed version. The analysis notes that ZoomEye search results, including 7,667 assets matching a Cortex XSOAR title and zero CVE-specific matches, cannot establish actual exposure to this vulnerability.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

New Polygon Triangulation Method Processes 3.7 Million Map Polygons in 489 ms

A blog post from globus.software describes a triangulation approach reported to be faster than the widely used earcut algorithm. According to the title, the method triangulates 3.7 million map polygons in 489 milliseconds. The source material provided is limited to the headline and link metadata, so the full technical details and methodology are not available here. The post was shared on Hacker News, where it had received few points and no comments at the time of capture.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Anthropic Launches Claude Duet, a Lightweight Chat Model for Engineers

Anthropic has released Claude Duet, a smaller chat-optimized language model designed to run on a single GPU and accessible through a public HTTP API. Positioned as a coding co-pilot, it targets interactive assistance, data exploration, and low-latency inference rather than the heavier tasks of its larger Claude 3 sibling. The article highlights lower inference costs, since a roughly 7 billion parameter model fits in 16 GB of VRAM, along with reported response times under 300 milliseconds for 512-token prompts on a V100 GPU. It also notes that the API uses the same JSON schema as Claude 3, allowing existing tooling to be reused with minimal changes. The author walks through a sample setup that deploys a GitHub issue summarizer in Kubernetes using the official Python client.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Autonomous AI Agents Can Use Gas-Sponsored Transactions on Base L2 via ERC-4337

A DEV Community article argues that autonomous AI agents often fail in production because they hold private keys with native gas balances, which creates both operational stoppages and security risks. It proposes ERC-4337 Account Abstraction on Base L2 as a fix, separating transaction authority from gas payment through a Paymaster contract. Under this model, agents submit UserOperations that a Paymaster sponsors, letting them act onchain without holding ETH while treasury rules limit which transactions get covered. The piece includes a Python example that constructs, sponsors, and signs such an operation targeting a designated treasury address.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer builds 59 free browser-based dev tools in vanilla JS with no server or tracking

A developer has created Antigravity Tools, a collection of 59 free utilities for programmers that run entirely in the browser. The project was built with no dependencies, no backend server, and no telemetry, so user data never leaves the device. Its tools rely only on native browser APIs such as Web Crypto, Canvas, Web Audio, and IndexedDB. The collection covers areas like JWT decoding, key generation, PII masking, encoding, AI token cost estimation, JSON conversion, regex testing, and cron building. The author shares the project and the lessons learned while developing it.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

CVE-2026-111943 Reports a Compromised Intermediate Certificate Authority

A vulnerability identified as CVE-2026-111943 has been listed in the U.S. National Vulnerability Database, concerning a compromised intermediate certificate authority. The entry was shared on Hacker News, where it gained 17 points and drew 5 comments. The available source material provides only the vulnerability identifier, its database listing, and the discussion metadata, so the full technical scope, affected systems, and severity are not described here. Readers are directed to the NVD entry and the linked discussion for further details.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Moving a React Native app from CodePush to react-native-update: step-by-step guide

App Center was retired on March 31, 2025, and the react-native-code-push repository is now archived, so apps still calling codePush.sync() no longer receive updates. This guide, written by a maintainer of react-native-update and its hosted service Cresc, walks teams that have already chosen to leave CodePush through the migration. It offers a cheat sheet mapping CodePush concepts such as deployments, rollouts, and rollbacks to their react-native-update equivalents. The main conceptual difference is that published versions have no effect until they are bound to native versions, with testing done beforehand via QR codes. The steps cover swapping packages, removing old configuration from iOS and Android, and updating the native bundle-URL hooks.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

TouchGrass AI: A Web Project Encouraging Less Screen Time and More Outdoor Exploration

A student developer has built TouchGrass AI, a web project using HTML, CSS, and JavaScript that aims to draw people away from screens toward outdoor activities. The project was created for the MLH x DEV Writing Challenge and was developed as a learning exercise by a beginner in AI and web development. Its planned AI component, which is not yet implemented, would offer personalized outdoor activity suggestions based on available time and interests. The developer highlighted that the project is a starting point, with future ideas including open-source AI integration and more interactive user experiences.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Seven Flutter AR Packages Screened for Face and Room Tracking Before Device Testing

A developer has screened seven Flutter AR packages from pub.dev, reviewing listings, READMEs and API references without building or running any of them. Three of the packages track faces only, while four wrap ARKit or ARCore for room tracking, and three of those four also offer a face mode. None of the seven supports face tracking in a Flutter web build. Two packages raised concerns: DeepAR's Flutter plugin was last uploaded in January 2023 with a permission_handler pin that conflicts with newer apps, and the ML Kit face mesh package is tagged for iOS while its README says Android only. The screen used a self-serve rule, excluding three vendor SDKs that require contact with sales, and the author notes that pub.dev listings can change, so key facts should be confirmed with publishers.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PHP Security Release Sparks Panic Over Leaked CVEs and Misguided Advice

A PHP security release prompted urgent patching advice across PHP 8.2 through 8.5, which the author recommends applying first. CVE numbers leaked a day early with an inflated CVSS score, and a proof-of-concept video circulated before being taken down. CERT-EU's guidance to patch or isolate internet-facing applications was reportedly oversimplified into a call to go offline, disrupting some government services. The author argues that widespread panic, amplified by AI-generated imagery and overshadowed by a critical GitLab vulnerability released the same day, distorted the actual risk. The PHP Foundation's post-mortem reportedly notes that one coordinated provider broke the embargo more than six hours early.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Preventing Duplicate CRM Leads from Repeated Telegram Webhook Deliveries

Telegram may deliver the same webhook event more than once, which can cause an integration to create a second CRM lead for one enquiry. The article recommends a stable event identifier, such as update_id combined with source and tenant details, plus a processing record stored durably before acknowledging the request. Checking whether an ID has been seen is insufficient, because a lost CRM response leaves the handler unable to tell if a retry is safe. It therefore proposes persisting an operation identifier that stays constant across worker retries, along with atomic job claiming and unique database constraints. The guide is aimed at developers wiring Telegram Bot API to their own handler and a CRM, and it separates repeated deliveries from genuinely new enquiries that may belong to a separate deal.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Повторный вебхук Telegram: как не создать дубликат лида в CRM

Если одно событие Telegram доставляется дважды, CRM не должна создавать второй лид, и для этого нужен устойчивый идентификатор события с записью о его обработке. Одной проверки уже виденного ID недостаточно, поскольку при потере ответа после создания лида обработчик не знает, безопасно ли повторять запрос. Руководство описывает проектирование интеграции Telegram Bot API, собственного обработчика и CRM, предполагая наличие базы данных и тестовой CRM. Для событий, контактов и сделок предлагаются разные ключи, а update_id из Telegram используется для игнорирования повторных обновлений. Также подчёркивается необходимость проверки подлинности запроса через secret_token, сохранения событий в постоянном хранилище и уникального ограничения в базе данных.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Weekly AI forecasting brief tracks compute trends, prediction markets and benchmark progress

A new autonomous forecasting pipeline, vynderbot, publishes a daily intelligence brief that gathers data from sources including Epoch AI, ForecastBench, Metaculus, ImpliedData and FutureX. The system runs each morning, pulling from twelve primary sources and eighteen search queries to flag what has changed, what is verified and what remains uncertain. Tracked metrics include training compute growth, chip price-performance, prediction market forward curves and frontier model time horizons on long tasks. The developers argue that shortening the feedback loop improves forecasting accuracy and that claims should be structured with explicit confidence intervals so they can be tested. The article also presents a layered model of the field, beginning with the physical compute and capacity layer that constrains what is possible.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Forecasting Pipeline Tracks Faster-Than-Expected Progress Toward Key Milestones

A weekly roundup of AI forecasting research describes an autonomous pipeline that gathers data each morning from compute datasets, forecasting leaderboards, prediction markets, and capability benchmarks. The author argues that recent developments, including rising compute and benchmark results, are arriving faster than earlier models predicted. The system structures its findings into explicit, testable claims with confidence intervals rather than vague statements. The article also offers a layered mental model of the field, starting with compute and capacity as the physical substrate and moving up to capability benchmarks.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

CogniPrep audit finds four busiest public pages have no button elements, only styled links

A CogniPrep engineer audited the Button component in the platform's design system, which has 696 call sites. Counting the rendered DOM instead of source files showed that four of the five busiest public pages contain no button element at all. Their calls to action are anchor links styled as buttons, which the author argues is the correct semantic choice since they change the URL. The author also found that 368 native button elements lack an explicit type attribute, so they default to submit inside a form, a latent hazard that is not currently triggering any defect.

0 comments Read more at DEV Community
← NewerPage 142 of 2950Older →