How to Automate Flutter App Releases Using Shorebird and GitHub Actions

Developer Samuel Adekunle published a guide on September 10 explaining how to automate Flutter app releases using Shorebird and GitHub Actions. The approach allows developers to push updates directly to users without going through the App Store review process. Shorebird is a code-push solution for Flutter that enables over-the-air updates to live applications. By integrating it with GitHub Actions, teams can set up continuous delivery pipelines that streamline the release workflow. The tutorial is aimed at Flutter and mobile developers looking to reduce release friction and deployment time.

How to Automate Flutter App Updates Using Shorebird and GitHub Actions

Mobile developers often face delays of days or weeks waiting for App Store and Play Store reviews before updates reach users. A technical guide published on DEV Community outlines how to use Shorebird, a code-push tool, combined with GitHub Actions to automate Flutter app releases and patches. The setup involves creating a Shorebird API key, storing it as a GitHub repository secret, and configuring CI workflows using three official Shorebird GitHub Actions for installation, releases, and patches. Separate workflow configurations are required for Android and iOS, each needing signing credentials stored securely as GitHub Secrets. Once configured, the pipeline allows developers to push updates to users within minutes, bypassing traditional app store review cycles for eligible changes.

Why NBA play-in odds and playoff odds are two very different numbers

A developer building an NBA playoff-odds simulator discovered that the NBA's play-in tournament structure makes its probability markets behave differently from those in other major sports. Unlike baseball or hockey, NBA teams finishing 7th through 10th in each conference enter a play-in round, and only two of those four teams advance to the playoffs. Prediction market Kalshi explicitly separates play-in qualification from playoff qualification, meaning the two outcomes can almost act as opposites for certain teams. To model this accurately, the simulator runs full season schedules, seeds conferences, plays out the three play-in games, and then simulates the complete playoff bracket. Key refinements — including sport-specific scoring exponents, per-season team rating variation, and accounting for two unscheduled NBA Cup games — significantly improved the model's accuracy against market prices.

NHL 2026-27 Monte Carlo Model Flags Big Gaps With Betting Markets Before Season Starts

A developer building Monte Carlo playoff simulators for the NHL identified four key rules that distinguish hockey from other sports: a three-outcome points system, a conference wild-card qualification structure, overtime strength dampening, and a new 84-game regular season starting in 2026-27 under the 2025 CBA. Running 20,000 simulated seasons in early September 2026, the model projected notable differences from prediction market prices, with Boston appearing 27 percentage points undervalued and San Jose 25 points overvalued compared to Kalshi contracts. However, the developer cautions that September model outputs carry limited weight because the simulator relies solely on prior-season performance regressed toward the mean. It has no knowledge of off-season trades, signings, injuries, coaching changes, or goaltending updates that prediction markets have already priced in. The article serves as both a technical breakdown of NHL-specific simulation challenges and a candid warning about the limits of early-season probabilistic models.

OzBrain

Discussion | Link

Developer builds Dixon-Coles football prediction model for eight under-modelled leagues

A software developer built a football prediction model using the Dixon-Coles statistical method, applying it to eight leagues that rarely receive modelling attention, including MLS, Liga MX, and Colombia's Primera A. The model generates 1X2, Over/Under 2.5, Both Teams To Score, and correct-score probabilities from a single consistent framework. Dixon-Coles improves on basic Poisson models by correcting for the real-world tendency of matches to end 0-0 or 1-1 more frequently than independent probability predicts, using an extra parameter called rho. The developer fitted league-specific parameters using three seasons of historical match data sourced from ESPN's public scoreboard API, with results weighted by recency to reduce the influence of older fixtures. Key practical findings included that rho and home advantage vary meaningfully by league, and that querying the API by date range rather than single day avoids false data gaps for lower-activity leagues.

US Pulls Racist 'Mr Singh' Self-Deportation Ad After Widespread Backlash

The United States government has removed a controversial advertisement that featured a character called 'Mr Singh', which was widely condemned as racist. The ad, which promoted self-deportation, drew sharp criticism from Indian-American and Sikh advocacy groups who called it deeply offensive. Democratic leaders, including California Governor Gavin Newsom, also voiced strong opposition to the advertisement. The backlash prompted authorities to pull the ad following mounting public pressure from multiple communities and political figures.

Cognition's SWE-2

Discussion | Link

Global climate record: August this year joint-warmest month with July '23

Developer Tests Frontier LLM's Ability to Extract Secrets from Obfuscated App Builds

A developer ran an experiment to assess how difficult it would be for an outsider to recover secrets from a hardened application build, using a frontier large language model as the investigator. The app's configuration values are XOR-encoded, with working keys derived at runtime from a 32-byte master key that ships inside the binary but is never stored as a readable string. The developer acknowledged upfront that this constitutes obfuscation, not cryptographic protection, since the binary contains all material needed for decoding. Key hardening measures included removing readable strings and decorrelating encoded values to prevent clustering by shared prefixes. The experiment focused not just on whether the LLM could recover secrets, but on how to critically evaluate its findings and gauge confidence in a failed recovery attempt.

Developer Cuts Docker Image 87% From 442 MB to 56 MB With Simple Optimizations

A software developer reduced a Python Flask microservice's Docker image size from 442 MB to 56.6 MB — an 87% reduction — by systematically redesigning its Dockerfile. The key changes included switching from the full python:3.12 base image to the leaner python:3.12-slim variant, removing unnecessary packages like curl, git, and vim that have no place in a production container. Additional improvements involved consolidating RUN instructions to avoid redundant image layers, reordering COPY commands to preserve Docker's layer caching, and adding a .dockerignore file to exclude test files and local environments. The hardened image also runs under a non-root system user instead of root, and replaces Flask's built-in development server with the production-grade Gunicorn WSGI server. The author notes that Debian slim was chosen over Alpine Linux due to better compatibility with Python packages that rely on C extensions.

Daily Laptop Use Causes Rare 'Toasted Skin' Condition in Boy

A boy developed a rare skin condition known as 'toasted skin syndrome' after using a laptop on a daily basis. The condition, medically termed erythema ab igne, is caused by prolonged exposure to moderate heat from the device's underside. While laptops typically do not reach temperatures high enough to cause burns, sustained low-level heat can still damage skin over time. The case highlights an underappreciated health risk associated with everyday laptop use, particularly when the device is placed directly on the body.

Princess Diana's 1994 'Revenge Dress' Expected to Fetch Up to $300,000 at Auction

Princess Diana's famous black silk gown, designed by Christina Stambolian, is set to go under the hammer with an estimated price range of $150,000 to $300,000. Diana wore the dress in 1994 on the same night then-Prince Charles publicly admitted to his extramarital affair, giving the outfit its iconic status. The gown previously sold at auction in June 1997, just weeks before Diana's death in Paris. The dress has since become one of the most recognizable pieces in royal fashion history.

Last year's row in mind, govt says won't reduce med PG cut-off

One Developer Spent a Full Day Using Only Expo-Built Apps — Here's What He Found

A developer challenged himself to spend an entire day using only apps built with Expo's open-source tooling, covering everything from sleep tracking and news to cooking and car washes. The experiment revealed that Expo-powered apps handle a surprisingly wide range of daily tasks, including weather, air quality, recipes, and even Tesla car controls. However, a significant gap emerged around messaging, as no viable Expo-built messaging app exists that could replace everyday tools like iMessage. The developer resorted to paying a celebrity via Cameo — which runs on Expo — to deliver a message to his wife, highlighting the platform's creative but uneven coverage. Overall, the test showed Expo apps can sustain a full day's routine, while also pointing to clear opportunities for developers to fill underserved categories.

Pezeshkian arrives today, will meet Modi ahead of Summit

Leaders set to fine-tune incubator network, startup fund, supply chain cooperation framework

Metrice launches zero-dependency post-quantum P2P mesh networking protocol

Metrice is a new decentralized peer-to-peer mesh networking protocol built entirely on Node.js core libraries, requiring no external npm dependencies. It integrates post-quantum cryptography using the NIST FIPS 203 ML-KEM-768 standard alongside Ed25519-based node identities for secure, IP-agnostic addressing via virtual .mesh domain namespaces. The protocol features a Tor-like three-hop onion routing architecture with uniform 2048-byte cell padding to resist traffic analysis and deep packet inspection. Nodes behind firewalls or CGNAT can maintain connectivity through persistent reverse TCP tunnels to public relay nodes, with dynamic role escalation allowing edge nodes to bridge segmented network segments. Additional security measures include SSRF defenses, loop-free gossip bridging, and an embedded in-memory SSH-2 server.

Claude Task Master CLI Automates Full Pull Request Lifecycle for Developers

Claude Task Master is an open-source command-line tool designed to automate the entire pull request workflow, from planning and code writing to merging. The tool reads a repository, generates a task list, executes changes in isolation, and opens pull requests automatically. If CI checks fail, it updates the code and pushes new commits without human intervention; it can also fetch and apply GitHub review comments on its own. Additional features include session state persistence, multi-profile support, and external integrations via a REST API, MCP server, and signed webhooks. Available under the MIT license, it can be installed through PyPI or Docker and works with Claude Code or any Anthropic-compatible API.

VernLLM: Open-Source TypeScript Framework Built to Handle Large-Scale LLM Calls

VernLLM is an open-source TypeScript framework designed to manage high-volume calls to large language models. The tool offers features including adaptive AIMD rate-limiting, advanced circuit-breakers, and multi-provider fallbacks. It also supports configurable retries with budget controls and built-in observability for monitoring call performance. Released under the MIT license, VernLLM is publicly available on GitHub and as an npm package at vern-llm.