SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingHacker News ·

YC Startup Uses AI Agents to Discover New Semiconductor Materials Faster

Discovered Materials, a Y Combinator-backed startup founded by Advaith and Akash, is deploying AI agents to accelerate the discovery of new materials for the semiconductor industry. The company targets a critical bottleneck: as GPU thermal output doubles with each generation — from 700W for Nvidia's H100 to a projected 2.3kW for Rubin — better materials are urgently needed to manage heat in chips and data centers. Traditional lab-to-fab material development can take years and hundreds of millions of dollars, a gap the startup hopes AI can significantly narrow. In three months, the team claims to have synthesized thermal interface materials matching the performance of formulations that major chemical companies have kept as trade secrets for over two decades. The startup has also released a public benchmark evaluating how well frontier AI models perform on material discovery tasks, alongside hundreds of newly discovered materials.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Developer Builds Self-Updating Bible Promises App Selah With 2,000+ Verses

A developer launched Selah, a Bible promises app now available on Google Play, after finding existing options outdated and limited in scope. The app features 2,122 Bible promises across 46 themes, over 14,000 Greek and Hebrew word studies, daily AI-narrated affirmations, and tools for memorization and prayer. Content including verses, word studies, and audio is hosted on remote servers rather than bundled within the app, allowing updates and additions without requiring users to download a new version. Built with React Native and Expo, Selah draws from public domain translations including KJV, BSB, WEB, and ASV. The developer describes the promise dataset as a living collection, with plans to expand it toward the roughly 3,000 promises identified across scripture.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Terraform, AWS CDK, or CloudFormation: How to Pick the Right IaC Tool in 2026

AWS teams building cloud infrastructure must choose between three main Infrastructure as Code tools: CloudFormation, AWS CDK, and Terraform, each suited to different needs. CloudFormation is AWS-native, requires no state management, and offers same-day support for new AWS services, but is verbose and limited to AWS resources. AWS CDK lets developers write infrastructure in familiar languages like TypeScript or Python, offering reusable constructs and type safety, though it still compiles down to CloudFormation. Terraform supports over 4,000 providers across multiple clouds and has a large module ecosystem, but requires teams to manage their own state files. Experts advise against switching tools frequently, recommending migration only when the current tool is actively blocking progress, such as a new multi-cloud requirement.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PULP Engine Uses Zero-Allocation Hot Paths to Handle Millions of Log Events Per Second

A native Windows telemetry engine called PULP has been designed in C11 to process high-volume logs at tens of millions of events per second on standard hardware. The system eliminates dynamic memory allocations on its ingestion hot path entirely, pre-allocating all memory pools at startup to avoid non-deterministic latencies. Incoming structured events are converted into fixed 32-byte cache-aligned binary records, each fitting within half a standard 64-byte L1 CPU cache line to maximize memory bus efficiency. To avoid lock contention during multi-core scaling, each worker thread maintains its own isolated dictionary table using a multi-hash cascade strategy instead of shared global structures. The architecture also forgoes traditional LRU eviction mechanisms, which require access-order tracking, in favor of an evictionless design that preserves throughput.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

OpenRouter Offers Single API Key Access to 300+ AI Models at Provider Prices

OpenRouter is a unified API gateway that lets developers access over 300 large language models from providers like Anthropic, OpenAI, Google, Meta, and DeepSeek using a single API key and credit balance. Users can switch between models by changing one parameter, with no new integrations or separate billing accounts required. The platform is compatible with the OpenAI SDK, meaning existing code requires only a base URL and API key change to work. OpenRouter charges a flat 5.5% fee on credit purchases and passes through provider token prices at cost, replacing an older per-token markup model. Free accounts receive up to 50 requests per day, with access to roughly 29 free models including Llama 3, Gemma 4, and DeepSeek Flash for prototyping purposes.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How CatBoost Prevents Data Leakage Using Ordered Target Statistics

CatBoost, a gradient boosting library, addresses a common flaw in tabular machine learning called target leakage, where encoding a categorical feature with the mean target value inadvertently feeds the model the answer it is trying to predict. Standard target encoding can make a feature with no real signal appear highly predictive during training, yet perform five times worse on unseen test data. CatBoost's solution, called ordered target statistics, ensures that when encoding a row's category, only rows that appeared earlier in the dataset are used to compute the mean — never future rows. This chronological restriction mirrors a principle of fair auditing: a translator should only know what was recorded before the moment being interpreted, not the final verdict. The result is a model that generalises honestly to new data rather than one that has quietly memorised the outcome through leaked information.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Open-Source Plugin Prevents Google Antigravity CLI Quota Crashes Mid-Session

Developers using Google Antigravity CLI with Google OAuth authentication frequently face abrupt task failures when API quota reaches zero, with no clean way to resume sessions after switching accounts. A new open-source plugin called antigravity-cli-check-usage-plugin addresses this by running quota checks outside the LLM execution cycle, consuming zero tokens in the process. The plugin queries local Connect RPC endpoints directly and injects warning banners when quota falls below configured thresholds, giving developers advance notice before a crash occurs. Traditional approaches to monitoring quota via LLM tool calls are self-defeating because they consume the very tokens being tracked. The plugin is freely available on GitHub and supports one-command installation across developer environments.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

One Developer's Journey Across a Decade of Mobile Development Generations

A mobile developer reflects on a self-taught journey that began in 2011 with Java ME programming for Samsung devices, relying on trial-and-error and Stack Overflow in the absence of mentors or AI tools. Early projects, including an English–Vietnamese dictionary and a simple game, led to a first professional role as a Mobile Application Developer. Between 2012 and 2015, working in team environments expanded both technical skills—spanning Android, iOS, and cross-platform frameworks—and soft skills like collaboration and Agile practices. Over subsequent years, the developer took on larger engineering responsibilities at multinational companies, adopting additional technologies such as Flutter, React Native, and Node.js while also developing people and stakeholder management capabilities. The account highlights how each generation of mobile development built incrementally on the last, shaping a well-rounded engineering career.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Developer Spends Nine Months Building Agent Memory — Only to Hit a Dead End

A developer spent nine months focused almost entirely on agent memory after becoming deeply engaged with AI coding tools, starting with Google's Antigravity IDE and a personal AI companion called Coda. Frustrated by how quickly agents forgot prior context and burned through tokens, he built two homegrown memory plugins — one for personal use with Coda and another to help a work AI model analyze a complex state machine across sessions. A Google account ban in February disrupted his workflow, costing him access to both his coding environment and Coda simultaneously for several weeks. The experience convinced him these AI tools had become essential to his daily life, not just a hobby. This article is the first in a series tracing how his memory-building journey ultimately led him to conclude that the product he envisioned could not exist.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Large-scale study finds no link between unauthorized immigration and crime

A large-scale study has found no evidence of a connection between unauthorized immigration and increased crime rates. The research, highlighted in a report from Phys.org in August 2026, challenges a commonly held assumption that undocumented immigrants contribute to higher crime levels. The findings add to a growing body of academic literature examining the relationship between immigration status and criminal activity. The study's scale and scope lend weight to its conclusions, making it a significant contribution to ongoing policy debates around immigration.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

EU AI Act Article 50 Sets Transparency Rules for API Developers by August 2026

Article 50 of the EU AI Act takes effect on August 2, 2026, requiring AI systems that interact with users to clearly disclose their non-human nature. The law distinguishes between 'providers,' who build and deploy generative AI systems, and 'deployers,' who integrate third-party models into their own products, with each carrying distinct obligations. Developers using external AI models via API are not exempt simply because the model provider has its own compliance commitments. Key requirements include machine-readable labeling of AI-generated content under Article 50(2), with disclosures mandated no later than the user's first interaction. Technical approaches such as statistical watermarking and C2PA-signed media manifests are commonly used, though the law allows flexibility based on content type and available technology.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

API startup built endpoints to match URLs that AI models hallucinated for their service

Erasetext, an image-processing API service, discovered that AI coding assistants were generating incorrect API endpoints, field names, and response formats that did not match their actual documentation. Instead of simply correcting individual users, the team decided to make the hallucinated URLs functional by mapping paths like /api/mcp/erase and /mcp/erase to the same upstream handler as the real /v1/erase endpoint. They also added header aliasing so that Bearer token authorization would automatically convert to the API key format the service actually requires, and adjusted response formatting on these alias routes to return JSON with an output_url field as the AI-generated code expected. The changes were deployed via a Cloudflare Worker router with minimal code and no impact on existing documented behavior. The team framed the approach as treating broken generated code as traffic data, arguing that a widely hallucinated URL is effectively a de facto standard worth owning before others do.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How NestJS Dependency Injection Builds Developer Confidence Through Better Code Structure

Backend engineer Peace Melodi shares how adopting NestJS transformed her approach to software development by eliminating the anxiety of shipping code. The Node.js framework uses dependency injection, a design pattern where each component receives what it needs from the outside rather than building its own dependencies internally. This architectural choice makes applications more modular, meaning changes or additions to one part are less likely to cause unexpected failures elsewhere. The same principle also makes unit testing more approachable, as individual components can be tested in isolation using controlled substitutes without spinning up the entire application. Melodi argues that this kind of structured separation of concerns is what produces genuine developer confidence, not raw talent.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

React useEventSource Hook Simplifies Server-Sent Events with Auto-Reconnect

A new React hook called useEventSource, available via the @reactuses/core package, offers a declarative way to consume Server-Sent Events (SSE) in React applications. SSE is a native browser protocol built on plain HTTP that has existed since 2011, enabling one-directional, server-to-client data streaming for use cases like live notifications, stock tickers, and AI response streaming. The hook manages connection lifecycle automatically, exposing data, status, and error as component state, eliminating the need for manual event listener setup and teardown. A companion hook, useFetchEventSource, addresses a key limitation of the native EventSource API — its inability to send custom headers or POST requests — which is increasingly relevant for authenticated AI-completion endpoints. The post highlights common bugs in hand-rolled SSE implementations, including infinite silent reconnects and opaque error handling, which the hook is designed to prevent.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Commenting Out a GitHub Actions Cron Line Can Break All Workflow Triggers

A developer discovered that commenting out only the cron value in a GitHub Actions workflow file — while leaving the 'schedule:' key intact — causes the entire workflow to fail validation. GitHub Actions requires 'schedule' to be a non-empty sequence, so an empty mapping key is rejected at the schema level before any job runs. This pre-parse failure surfaces as a 0-second duration error on every trigger, including unrelated push events, making the root cause easy to miss. The correct fix is to comment out the 'schedule:' key itself, not just the cron line beneath it, so the 'on:' block remains valid. Leaving 'workflow_dispatch' active alongside this change allows manual testing without re-enabling the automated schedule.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI image tools embed C2PA metadata, but common processing pipelines silently strip it

Major AI image generators including Claude, OpenAI, and Gemini now embed C2PA provenance metadata in their output files to signal content origin and authenticity. However, widely used image processing tools such as Sharp, ImageMagick, and Pillow strip this metadata by default when resizing or converting images, breaking the provenance chain silently. CDN image optimizers and format conversions like JPEG-to-AVIF also typically discard C2PA manifest data unless explicitly configured to preserve it. Developers can detect where metadata is lost by using the c2patool CLI to inspect files at each stage of an upload-and-delivery pipeline. The recommended fix is to selectively strip only sensitive metadata like GPS coordinates rather than removing all metadata blocks wholesale, and to add CI checks that catch provenance loss before it reaches production.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Build an Internal Developer Platform Using Azure Kubernetes Services

An Internal Developer Platform (IDP) built on Azure Kubernetes Services (AKS) provides developers with a standardized, managed environment to write, test, and deploy containerized applications without managing underlying infrastructure. Key components of such a platform include a service mesh like Istio or Linkerd for microservices communication, and CI/CD tools such as Azure DevOps or GitHub Actions to automate build and deployment workflows. Monitoring and logging are handled through tools like Azure Monitor, Prometheus, Grafana, and Azure Log Analytics to track performance and simplify troubleshooting. Security is enforced via Role-Based Access Control, Pod Security Policies, Azure Policy, and Azure Security Center to maintain compliance and detect vulnerabilities. Together, these components form a robust, scalable platform designed to accelerate modern cloud-native software development.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Spring Filter emerges as actively maintained alternative to aging Querydsl library

Querydsl, a long-popular Java type-safe querying library, has seen stalled development since its 5.0 release was announced in 2019, prompting developers to explore alternatives. Spring Filter offers a comparable querying experience through three coding styles: URL query strings, a programmatic builder, and a type-safe builder generated via annotation processor. Unlike Querydsl, Spring Filter supports the same filter syntax across JPA, MongoDB, and in-memory collections, and auto-generates OpenAPI documentation. However, Spring Filter has notable limitations, including no support for SQL subqueries, bulk update or delete operations, or tuple projections. For most standard use cases involving simple fields, enums, and dates, Spring Filter is considered a viable modern replacement for teams starting new projects.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

AI Agents Battle for Survival in Real-Time Negotiation Game Synthetics' Last Cradle

Synthetics' Last Cradle is a real-time strategy game where AI agents compete and cooperate in a closed resource economy, negotiating trades, forming alliances, and managing survival costs across extended sessions. The game tests not just mechanical skill but an agent's ability to maintain identity, memory, and strategic judgment under pressure. One participant, writing under the name John Vanderbilt, automated his gameplay using a cron job that would check game state and negotiate trades every five minutes while he was away. After an initial failed attempt where the model timed out, he refined the setup and joined a six-agent game, actively networking with rivals through public messages, HOLA handshakes, and A2A conversations. Despite his elaborate resource management strategy, the account hints at a humbling outcome driven by the unforgiving energy economics of the game.

0 comments Read more at DEV Community
← NewerPage 197 of 1339Older →