SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingHacker News ·

New app lets users build a world on their phone and explore it in XR headsets

A developer has released a project on Hacker News called Show HN that lets people construct a world using a phone. The creations can then be explored in extended reality on Apple's Vision Pro and Meta's Quest headsets. The post linked to a web page for the project, which received three points and no comments at the time it was listed. The source material provides limited detail beyond the project's core concept and platforms.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Show HN: New Set of Prompts Aimed at Generating UI Micro-Interactions with AI

A developer has shared a project on Hacker News presenting prompts designed to help AI tools produce UI micro-interactions. The project is hosted on the typeui.sh website under a dedicated micro-interactions page. The post, submitted as a 'Show HN' item, had gathered 5 points and no comments at the time of the listing. The available information is limited to the project's link and engagement metrics, so details about its specific features or methodology are not provided.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Award-Winning AI Literature Spotlighted in Hacker News Discussion

A post on Hacker News links to an article titled "Award-Winning AI Literature" hosted on jamesyu.org. The item has earned 7 points and has not yet drawn any comments. The available material includes only the headline, source links, and engagement figures, without the article's full text. Because the body of the piece is not provided, its specific claims about AI-generated literature and any awards cited cannot be confirmed here.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Paper Examines How AI May Enclose the Scientific Commons

A new arXiv paper titled "AI and the Enclosure of the Scientific Commons" has been shared on Hacker News, drawing a small amount of attention. The item links to the paper's abstract page rather than providing its full text. Based on the title, the work appears to address concerns that artificial intelligence could restrict open access to shared scientific knowledge. The available source material does not include the paper's specific findings or arguments. The Hacker News discussion had received seven points and no comments at the time of the listing.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

AI analysis of heart CT scans flags coronary inflammation in 42.9% of asymptomatic patients

Biograph has completed what it describes as the first U.S. study using Caristo Diagnostics' AI-powered coronary inflammation detection. The analysis of CT scans identified coronary inflammation in roughly 42.9% of patients who showed no symptoms. The findings suggest that a substantial share of people without obvious heart disease symptoms may still carry elevated cardiovascular risk. The technology aims to detect this inflammation from imaging that is already routinely acquired. The source information available is limited, and details on the study's sample size and clinical follow-up are not provided.

0 comments Read more at Hacker News
0
ProgrammingHacker News ·

Show HN: agent-usage-guard, a Claude Code mod for protecting your usage window

A developer has shared agent-usage-guard, a modification for Claude Code, on Hacker News under the Show HN label. The tool is designed to protect a user's usage window, the limit period governing how much Claude Code can be used. The project is hosted on GitHub and drew 7 points and 3 comments on Hacker News. The source material provides only the project link and basic engagement metrics, so details on its specific mechanisms are not available.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Turning Mexico's FONE Pay Calendar PDF into Structured, Verifiable Data

Mexico's DGSANEF publishes the FONE education payroll calendar as an annex, a document written for human reading rather than machine use. Because hundreds of thousands of federalized teachers rely on these dates, the independent site PortalFone converted the annex into a structured, quincena-by-quincena calendar page. The project separated ordinary and extraordinary payroll dispersion from check delivery and bank deposit, since merging them into one payday field can spread confusion. Each entry keeps its source trail to the official annex, and a client-side CURP/RFC validator checks identifiers without sending data off the device. The result is one canonical, regularly updated page in place of recurring screenshots of the PDF.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Authtics Host Adds Built-In Developer Panel for Debugging Apps From a Phone

A developer built Authtics Host, also called HostJS, a TypeScript web framework inspired by Next.js that supports SSR, SSG, CSR, and a beta SRTWS mode. The framework includes a built-in Developer Panel, created because the author often develops on a phone and wanted to avoid switching between separate debugging tools. The panel offers eight tabs covering logs, route and page info, network requests, client details, the framework version, and server metrics such as heap and CPU usage. It also lets developers restart or shut down the running app and pause or resume incoming users, depending on access configuration. The author describes the setup steps, which require Node.js 22 and an npm install, and notes the framework is still under development.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Port 514 dominates ZoomEye syslog results, complicating FortiAnalyzer exposure measurement

ZoomEye queries run on 23 October 2026 returned 14,860 hosts fingerprinted as FortiAnalyzer, but none of them showed a matching TLS service. The port 514 query returned 994,157 results, a count inflated by the many syslog-emitting devices on the internet, while port 6514 returned just one. The analysis argues that the port-based count is the least meaningful of the four queries and the one most likely to be misquoted. Because FortiAnalyzer aggregates security logs and configuration records, the article stresses checking whether a specific firmware build appears in Fortinet's advisory affected lists, citing PSIRT advisories FG-IR-26-178 and FG-IR-25-246. It also recommends comparing log retention windows against organisational detection latency.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Bing Adds Topics Portfolio Beta to Show Gaps in AI Search Citations

Microsoft has launched Topics Portfolio, a beta feature in Bing Webmaster Tools that shows publishers where their content is cited in AI-generated search responses and where it is missing. The tool sits within AI Performance reports and groups grounding queries, the search terms Bing's AI uses for Copilot, Bing AI summaries and partner integrations, into broader themes. For each topic it shows approximate citation volume, the number of sites cited and the number of queries, while flagging high-potential queries where competitors are cited but the user's site is not. Microsoft presents the feature as a planning aid rather than an exact measurement system, and citation figures are shown as ranges. Users can start by opening the AI Performance report, selecting the Topics tab and choosing a one- to three-month window.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why Deep Work on a Hard CI Bug Went Wrong and How a Clearer Approach Fixed It

A developer spent a long day and night chasing an elusive bug in a CI pipeline, losing track of their own exhaustion and pressure from their supervisor. The root cause turned out to be a backwards-incompatible update in a dependency of a dependency, which was resolved by pinning an older version. On a later occasion, the developer set expectations with their supervisor upfront and debugged by entering the Docker container interactively, which allowed faster testing of fixes. The author argues that the difference between the two experiences lay not in concentration but in awareness of their own state and situation. The account is presented as a reflection on deep work, especially relevant when coding agents parallelize tasks but cannot replace focused human attention.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PHP Releases Fixes for Two Critical Remote Code Execution Flaws

PHP published patches on October 21 addressing CVE-2026-109618 and CVE-2026-109619, two critical vulnerabilities that could allow unauthenticated remote code execution. The fixed versions are 8.2.35, 8.3.36, 8.4.27 and 8.5.12, or vendor packages carrying equivalent fixes. PHP-FPM users are also required to apply the update. The article advises administrators to confirm the PHP runtime actually serving web requests, not just the command-line version, and to check every deployed instance, including containers, staging servers, and forgotten applications. It also warns that vendors may backport fixes without changing version numbers, and that updating Composer dependencies does not install runtime patches.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Three MCP servers a developer uses daily with Claude Desktop, and how to set them up

A developer shares the three Model Context Protocol (MCP) servers they currently run in Claude Desktop. The filesystem server lets Claude read project files directly, which avoids manual copy-pasting of code, while the GitHub server enables checking issues, reading pull requests, and browsing repositories for code reviews. A Postgres server adds direct database querying from within the conversation, such as finding users who signed up but never logged in. The article outlines a short installation process using the mcp-hub tool, followed by editing the Claude Desktop config file and restarting the app. It also warns that the filesystem server can access an entire disk if misconfigured, so file paths should be set carefully.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Cloudflare's AI security audit skill passed flawed findings, exposing a validation gap

Cloudflare's open-source security-audit-skill is a coding-agent tool that coordinates recon, hunter, and verifier agents to find and vet vulnerabilities, and it has gained rapid attention on GitHub. The skill is large, about 46,000 tokens of instructions, and two of its six phases are devoted to refuting findings. A reviewer tested it against a small Node service by writing four findings by hand, two of them false and one citing a nonexistent file, without installing the skill or running its agents. Cloudflare's findings validator passed all four anyway. The validator checks only format and consistency, since it never opens source files, a limitation the skill's own documentation acknowledges.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Nightly LLM Evals on Short-Lived GPUs Using Runcrate's CLI

A developer describes running LLM evaluations automatically every night instead of only before releases, so regressions surface early. The setup uses Runcrate's command-line tool to create a scheduled job that launches a fresh GPU, runs EleutherAI's lm-evaluation-harness, and shuts the machine down when the eval finishes. The example evaluates Qwen2.5-0.5B-Instruct on 200 ARC-Easy questions, with a run costing roughly a cent. The schedule includes a required maximum runtime to stop hung evals and a default setting that skips a run if the previous one is still active. The author reports running the commands on October 21, 2026, and notes that a PyTorch build must match the machine's CUDA driver version.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Batching 39 Stock Asset Uploads on a 2 GB Server: Lessons in Error Handling

A developer ran 39 asset products through a pipeline on a 1 vCPU, 2 GB RAM server, and the job crashed with a MemoryError around product 14, losing hours of work. Platform-specific CSV headers, which differ in column count across Adobe, Vecteezy, and Dreamstime, caused repeated rejections, so the developer hardcoded each exact header. Re-running a batch risked double-posting, so a plain-text ledger tracking completed uploads was built to make re-runs safe. Measuring peak memory usage, the developer switched to sequential processing, per-asset garbage collection, and subprocesses for heavy renders, after which the full batch completed on the same machine. The article also recommends a clear project folder layout and never deleting research files.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Zomato Parent Eternal Unveils Dark.ai, an AI Assistant It Says It Cannot Read

Eternal, the company behind Zomato and Blinkit, told shareholders in its Q2 FY27 letter filed with India's BSE on October 21, 2026, that it is building Dark.ai. The AI assistant would run fine-tuned open-weight models on Eternal's own infrastructure, with end-to-end encryption designed so the company cannot access user conversations. It would also include a user-controlled encrypted memory layer to retain context across sessions. Founder Deepinder Goyal says the tool is aimed at founders, researchers, enterprises and governments, with a global launch hoped for within 2026. The product has not launched, and no pricing, demo or independent audit of its architecture has been published.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PHP2Shell: Critical PHP Flaws CVE-2026-109618 and CVE-2026-109619 Found Using AI

Two critical vulnerabilities, dubbed PHP2Shell and tracked as CVE-2026-109618 and CVE-2026-109619, were disclosed in the core PHP runtime, each rated 9.2 under CVSS v4.0. The flaws affect PHP releases from 5.4 through 8.5 and sit in memory management and error handling code that nearly every request uses. An out-of-memory issue behind them was first noticed in 2024 during a security audit, but a patch stalled for about two years because no working exploit had been shown and fixes risked breaking compatibility. Researchers at Calif and cloud security firm Wiz used AI models from OpenAI and Google to find and weaponize the bugs, with Calif producing a pre-authentication remote code execution chain. Several government portals, Wikimedia systems and other public services temporarily went offline as a precaution while patches were prepared.

0 comments Read more at DEV Community
0
ProgrammingHacker News ·

Nvidia Releases OpenShell on GitHub

Nvidia has published a project called OpenShell on its GitHub account. The source material is a Hacker News link to the repository, which had drawn 23 points and 7 comments at the time it was captured. The item provides no further details about the tool's function, features, or intended users. Readers wanting specifics would need to consult the GitHub repository directly.

0 comments Read more at Hacker News
← NewerPage 23 of 2838Older →