SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World

Programming

LatestTopDiscussed
0
ProgrammingHacker News ·

Strot compiles an LLM's intent once so later decisions can run locally

A developer has shared Strot, a project posted on Hacker News as a Show HN, that aims to capture an LLM's intent a single time and then run decisions locally. The project is hosted on GitHub under the strot-ai organization. The post gathered 16 points and 4 comments on Hacker News at the time of listing. The available source provides only the project link and engagement figures, so details on how the compilation works are not confirmed here.

0 comments Read more at Hacker News
0
ProgrammingDEV Community ·

Designing AI Study Notes That Learners Can Actually Use

The article argues that an AI note tool producing a summary, flashcards, and a quiz from one document should give each output a distinct purpose and let users verify its claims. It uses an invented offline recording specification to illustrate how summaries must keep critical conditions, such as the difference between recording, uploading, and processing. Flashcards work best as separate questions that each cite a specific source detail, rather than one card that restates the whole passage. The author also cautions that fixed card counts can encourage repetitive filler, so the number should depend on the material. Suggested controls include editing questions, correcting answers, and removing duplicates, along with scenario questions that test understanding.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Committed Android Baseline Profiles Can Silently Do Nothing

A baseline profile can be checked into a project and bundled into the APK yet still have no effect on startup speed. A common cause is generating it against a debug build rather than a release-shaped build, since R8 renames classes and the profile then references symbols absent from the shipped app. The profile also goes stale when new screens or navigation destinations are added without regenerating it, leaving those code paths unoptimized. Because none of this fails the build, developers are advised to verify installation and compilation state directly with adb commands like dumpsys package compiler-stats. Regeneration should be treated as a release-checklist item, with before-and-after startup measurements used to confirm the profile is working.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Automating .NET Assembly Obfuscation via Visual Studio CLI Post-Build Steps

A developer at Rustemsoft describes how manual GUI-based obfuscation of .NET assemblies can be skipped by mistake, leaving code exposed in CI/CD pipelines. The guide explains wiring Opaquer's command-line interface into Visual Studio post-build events so Release builds are protected automatically. It covers limiting parallel project builds to avoid file-lock errors, and running obfuscation with options for private member renaming, string encryption, and control-flow obfuscation. A batch script is provided for CI/CD use, along with a note that setup projects need obfuscation before packaging. The Command Line Interface requires the Pro or Enterprise edition, as the free edition includes only the GUI.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to design a web search engine in a system design interview

A web search engine prompt is deliberately oversized, hiding three distinct systems: a crawler, an indexer, and a serving layer. Strong candidates scope the problem early and treat it as top-K retrieval rather than crawling. The core structure is an inverted index that maps each word to the documents containing it, turning queries into lookups. Because the index is too large for one machine, it is sharded by document and each query fans out to all shards, with results merged afterward. The rest of the interview covers how the index is built, how fan-out stays fast, and how matching pages are ranked.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

SQL Interview Question: Finding Departments Paying Above the Company Average Salary

A practical SQL interview problem asks for departments whose average salary exceeds the company-wide average. The article supplies two sample tables, one of employees with department IDs and salaries, and one of departments with names. The company-wide average is given as ₹67,500, and the IT department's average of ₹90,000 is the only one above it. The suggested solution uses a common table expression to compute each department's average and a subquery for the overall average, then filters the results. The author also poses a follow-up question about whether HR would qualify if a new ₹120,000 employee joined it.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Designing failure tests from what users must never see

The author, who builds free open-source alternatives to subscription work software, proposes starting test design from the wrong thing a user could see rather than from the code's behaviour. The suggested exercise picks one visible claim, such as an operation shown as complete, and identifies the evidence needed to justify it. Dependencies like the database, cache, or mail provider are removed one at a time, with the user-facing result and the operator-facing log checked separately. As an example, their status page tool Lamp promises that a failed read never shows an all-clear, and its tests verify this and check that a failed page serves a labelled last known state. A stress test also revealed a scheduled-update delay of 238.5 seconds, so the promise was revised from 60 to 240 seconds, with the old and new promises kept in the release record.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

RepoTrace Captures Failing Commands as Replayable, Testable Reports

RepoTrace is an open-source MIT-licensed Python tool that records failing commands as structured JSON reports stored by default under .repotrace/reports. Users can replay those recorded commands later, including on another machine, to check whether a failure still occurs and compare results against the original trace. The tool also renders reports as Markdown and generates self-contained pytest regression tests that embed the command and its expected output. It is built with Python 3.12+, a Typer-based CLI, Pydantic v2, and Rich, and is available on PyPI and GitHub. The project's authors position it as a way to reduce "works on my machine" debugging and to support CI pipelines that collect failures automatically.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

PSEC-99: A 3D Browser Game Built in C# and WebAssembly Without WebGL

OpenSilver developers created PSEC-99, an unofficial 3D tribute to the 1982 TI-99/4A game PARSEC, running in the browser. The game is written in C# and compiled to WebAssembly, and it renders its entire 3D scene into a single int[] pixel buffer rather than using WebGL or a game engine. That buffer is displayed on an HTML canvas through the new WriteableImage control in OpenSilver 3.4 Preview. Both the game and the OpenSilver framework are open source under the MIT license, and the game runs on .NET 10 WebAssembly, including on phones.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

VoiceLedger Turns Spoken Job Notes into GST Invoices Using Open-Weight Models

VoiceLedger is a prototype app designed to help independent contractors such as plumbers, electricians, and home bakers create tax invoices quickly. Users record a short voice note in English or Hinglish after completing a job, and the app transcribes it with ElevenLabs Scribe. It then matches the client against a MongoDB Atlas database, extracts materials and labor into a structured format using open-weight models like Gemma 2 or Llama 3.3, and computes GST totals through deterministic JavaScript logic. The resulting invoice can be edited, exported as a PDF, shared via WhatsApp, and includes a UPI payment QR code. The project is positioned as a micro-SaaS subscription for small service businesses.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Invoice AI Turns Phone Photos of Invoices into Small Business Insights

Invoice AI is an open-source tool built for a Hacktoberfest challenge that helps small shopkeepers analyze their finances. Users photograph invoices with their phone, label each as a sales or purchase invoice, and upload them. The system uses PaddleOCR to extract fields like vendor details, dates, subtotals, taxes, and totals, then validates totals and flags ambiguous entries for user review before they feed into analytics. A FastAPI backend and Streamlit dashboard present metrics such as vendor spending, monthly trends, and sales counts. An interactive assistant powered by the open-weight Gemma 4 model, accessed through Ollama Cloud, answers business questions using the confirmed data.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Designing a CDN: a system design walkthrough for interviews

A CDN design is framed not as a simple file server but as a globally distributed read-through cache that serves requests from nearby edge servers. The core challenges are routing each user to a healthy nearby edge and keeping cached content fresh and invalidated correctly. Because geography and caching are essentially the same problem, the design aims to answer most requests from a close cache hit while keeping cache misses cheap and safe for the origin. The walkthrough follows the structure a candidate typically draws in a roughly 45-minute interview round, covering requirements, estimates, data model, high-level design, and deep dives. It also highlights a frequent follow-up scenario: a sudden viral surge that causes a cache-miss storm against a single origin.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

DOE Emergency Grid Orders in 2026 Surpass All of 2025, Compiled in New Dataset

The U.S. Department of Energy has issued 66 section 202(c) orders, extensions and amendments in 2026 so far, compared with 19 across all of 2025. A compiled dataset of 85 entries was assembled from DOE's published pages, with accompanying Python scripts whose tests check the counts. The analysis finds that short emergency orders arrive in clusters, and that orders keeping named plants available are typically renewed back to back, often on the day the previous order lapses. One plant, H.A. Wagner, had a gap of 141 days without an order, while J.H. Campbell has been under continuous order since May 2025. The author notes that DOE's 2026 list omits one amended order that ERCOT's own page includes, and has published the data under CC BY 4.0.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Hosted Checkout Breaks Purchase Analytics; Webhooks Restore Attribution

A developer building subscriptions for Belle AI found that a standard GA4 purchase event on the thank-you page recorded almost no sales. Hosted checkout sends users to the payment provider's domain, so they often never return to the site, and revenue appeared in the payment dashboard but not in analytics. To fix this, the developer stored first-touch UTM parameters and referrer data in localStorage, which are preserved across later visits. At checkout, that attribution data and GA4 identifiers were passed into the order metadata, which the payment provider returns in its webhook. The webhook handler then sends a purchase event to GA4 via the Measurement Protocol, reusing the stored client and session IDs so sales are credited to the correct traffic source.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

LLM API list prices steady in mid-October as Chinese models trade at about a fifth of US prices

The TANDIU weekly data note for October 13 to 19, 2026 tracks official list prices for 69 AI models and reports that no vendor changed an official list price that week. Chinese models are priced in RMB, so their dollar equivalents shifted only with the exchange rate. The blended average is about ¥56.48 per million tokens for US models versus ¥10.79 for Chinese models, a gap of roughly 5.2 times. DeepSeek V4.1 Flash drew attention on social media, with its input and output tokens costing about 1/13 and 1/17 of Claude Opus 5.5's at list price. The same period's funding tracker logged over 150 rounds, including Ramp's $1.85B Series G at a $60B valuation.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why a two-node Proxmox cluster freezes, and how a QDevice vote fixes it

A two-node Proxmox cluster can lock itself up because losing either node leaves the survivor with exactly half the votes, which is not a majority. Without quorum, the cluster filesystem at /etc/pve goes read-only, blocking guest starts, config changes and backups even though the surviving node is healthy. Adding a QDevice, a small external voter that can run on a Raspberry Pi or similar always-on Linux box, provides a third vote so the survivor keeps quorum. The QDevice must run on separate hardware, since placing it on a cluster node would remove two votes if that node fails. Unstable links, such as Wi-Fi with high jitter, can also make a cluster appear broken by causing nodes to repeatedly drop out and rejoin.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Choosing Between Vercel and Cloudflare for Low-Cost Website Hosting

A solo developer working on a micro-SaaS product shares lessons on keeping hosting costs minimal while avoiding unexpected cloud bills. The piece compares traditional VPS setups, which demand manual server management, with serverless platforms that handle operations and scale automatically. It argues that early-stage founders should default to serverless to preserve their time for building and validating products. The author also notes that workloads needing persistent processes, heavy computation, or custom binaries are better suited to VPS or container hosts. The article then turns to a comparison of Vercel and Cloudflare Workers, as the author's preferred platform choice.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Stateless Crypto Store Reads Payments Directly From BNB Chain, Skipping a Backend

A developer has built an e-commerce checkout with no backend or database, relying entirely on public blockchain data. The store checks the merchant's USDT balance on the BNB Smart Chain through a read-only contract call made from the customer's browser. The page captures a baseline balance when an order starts, then polls every 20 seconds and marks the order paid once the balance increase matches the quoted amount within a 0.1% tolerance. Orders expire after 60 minutes, and small random amount suffixes plus manual owner verification are used to handle concurrent buyers. The author presents the design as eliminating dropped-webhook reconciliation problems, at the cost of a manual confirmation delay before goods are shipped.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

LatePay Uses TabPFN to Flag Freelance Invoices Likely to Be Paid Late

LatePay is an open-source tool that scores outstanding freelance invoices by their risk of late payment using TabPFN, a tabular foundation model. Each invoice receives a risk band from HIGH to LOW, with a suggested action such as chasing payment, requesting upfront deposits, sending a reminder, or keeping standard terms. Users can score invoices individually or upload a CSV for batch scoring, and missing client history falls back to a new-client baseline. The model is trained on a simulated dataset of 4,000 invoices rather than real payment records, so all outputs are labeled as simulated and are intended only for prioritising follow-ups, not as financial advice or credit scores. The project is built with a Python/Flask backend and a static demo that runs without a server, with code available on GitHub.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Unix Timestamps Explained: Seconds vs. Milliseconds, Leap Seconds, and the 2038 Problem

A DEV Community guide breaks down Unix time, which counts elapsed seconds since the 1970 epoch and ignores timezones and calendar quirks. It highlights a common microservice bug where 10-digit second-based values are mixed with 13-digit millisecond values, leading to absurd dates like the year 58,742. The article explains how leap seconds can cause out-of-order events in high-frequency systems and how NTP smearing is used to mitigate this. It also covers the origins of the 1970 epoch in early Unix development and warns that 32-bit systems will overflow in January 2038. The piece closes with practical guidance on storage and migration planning.

0 comments Read more at DEV Community
← NewerPage 330 of 3080Older →