daily.dev · Popular · 实时热榜

A developer reflects on how shifting from GUI tools to the terminal reshaped their understanding of how software actually works, covering filesystem navigation, process management, Linux permissions, basic networking debugging, and reading error output. The real payoff wasn't typing speed but developing investigative habits: checking what's running on a port with ss or pgrep, inspecting file permissions with ls -l and stat, testing DNS and HTTP with dig and curl, and understanding pipes. The pie

Some of the best conversations on daily.dev start in the comments and then want somewhere quieter to go. A follow-up question, a “how did you set that up?”, a “we should build this together”. Now you can take it one-to-one. Here’s what’s new Messages, right in daily.dev : Your inbox lives at https://daily.dev/messages . Open it from the Messages icon in the header or the sidebar. Message anyone from where you meet them : Hit Message on a profile, the mail icon on a user’s hover card, or Message

Gemini 4 Argon is now available in the Gemini app for Google AI Pro and Ultra subscribers, offering improved reasoning, clearer and less repetitive responses, and better progress updates on longer tasks. The update also brings Google Antigravity's agentic coding interface to the Gemini macOS app, a new chat-to-tasks toggle for background task completion, and connections to over 40 apps including Instacart, OpenTable, Adobe, and Shopify.

A dark software factory is where agents take over the complete software development and delivery. So where are the humans? Outside the loop. In software, a dark software factory means a fully autonomous SDLC pipeline. Agents plan, write, review, test, and push code straight to production. No human ever reads the individual pull requests and everything is set to autonomous. You give the system a high-level goal or specification, and the agents handle the rest. Instead of human approvals, they rel

Anthropic just published a piece about rebuilding their internal version control system from scratch, and one number stopped me mid-scroll: in May 2026 they were merging ~2,500 code changes a day. By September, over 30,000. Same company, four months apart, 10x throughput. That’s not people typing faster. That’s what happens when most of your code changes start coming from Claude agents instead of humans. I went down the rabbit hole of understanding what they built, a system called Scion, and it

A developer built KillSwitch, an esoteric stack-based programming language deliberately designed to confuse LLMs and coding agents, exploiting known weaknesses like letter counting, safety guardrail triggers, prompt-injection-style phrases, and noisy narrative syntax. Benchmarking several major models (Claude Opus 5, Gemini 4 Argon, Fable 5.1, open-weight MiMo-V2.6-Pro) on simple tasks in this novel language found Opus 5 performed best with a score of 66.8, suggesting models can handle simple ta

Alibaba's Qwen team released Qwen3.8 Flash Next, an experimental preview model with a technical report described as an early look at the architecture planned for Qwen 4. The design mixes a large mixture-of-experts network (125B total parameters, ~6B active per token) with hybrid attention: most layers use a compact linear-memory mechanism (Gated DeltaNet) while every fourth layer uses sparse attention for more precise lookups, plus residual gating to preserve signal across layers, an n-gram look

Omarchy 4.0.5, the Arch Linux distribution built around Hyprland, now launches the default web browser faster by replacing the slow xdg-settings call with a custom helper. David Heinemeier Hansson, using Claude Code, wrote omarchy-cmd-default-browser, which reads mimeapps.list files directly and resolves the default browser in about 10ms instead of the 72ms taken by xdg-settings due to its chain of POSIX shell scripts. The helper checks Hyprland-specific config first, falls back to xdg-settings

TLDR: The biggest news this week is a soundness bug in Rust’s borrow checker that’s been compiling clean since 2017 (and possibly 2015), letting safe code trigger use-after-free via a malicious Drop impl. Separately, two independent write-ups on the same Head Start compiler patch show up to 54% faster cargo check on real projects by overlapping compilation across dependency chains. Rustls shipped four security fixes including an uncapped memory allocation bug, and Rust 1.99.1 landed to fix three

A cartoon-framed opinion piece warns that AI assistants generating SQL statements remove the human pause that used to prevent destructive UPDATE or DELETE mistakes. The author (via a recurring cartoon character, Cardinal) proposes three cheap safeguards: give AI assistants read-only database access, wrap data-changing operations in explicit transactions with a row-count check before committing, and always write the WHERE clause before the verb. The piece argues AI can speed up writing queries, b

A software engineer recounts completing a design system migration that unified styles across React, HTMX, and Jinja stacks, replacing the unsupported JoyUI library with a custom Tailwind + ShadCN/BaseUI component system. The approach started with small tooling (a DESIGN.md guide, linters, minimal component library), followed by a big-bang migration PR, a month of iterative polish, then a second large push using AI coding agents. One agent session ran 6-7 hours and dropped @mui/joy imports from 5

PostgreSQL 19's general availability has slipped from October to 19 November 2026, with a second release candidate on 12 November. Thirteen release-note items covering 11 features were pulled after beta 1, including SQL/PGQ graph queries, UPDATE/DELETE FOR PORTION OF, online checksums, and ALTER TABLE MERGE/SPLIT PARTITION. Peter Eisentraut, who committed several reverted features, attributes the reverts to a mix of causes: unresolved design problems, a long tail of edge-case bugs surfaced faste

1,170+ free, open-source animated icons, now available beyond React. AnimateIcons brings the same collection of animated icons to React, Vue, Angular, and vanilla JavaScript, with support for Svelte, Solid, and Astro. Add motion and polish to your interfaces without creating animations from scratch. Choose your framework and use the same icons across your frontend stack. Browse the collection: https://animateicons.in Open source: https://github.com/Avijit07x/animateicons

React Native 0.88 ships with no user-facing breaking changes, the third such release in a row. Key highlights include switching event dispatching to a Web-based EventTarget model by default, variable font support via fontVariationSettings on Text and TextInput for both platforms, SceneDelegate lifecycle support on iOS (needed ahead of Apple's iOS 27 SDK requirement), ArrayBuffer support for Java TurboModules, and experimental Swift Package Manager hardening with CocoaPods eventually being phased

freeCodeCamp.org published a free 3-hour YouTube course teaching system design using a five-step framework for structuring thinking and tackling vague requirements under interview pressure. The course walks through building high-level architectures for real-world systems like Spotify, Booking.com, and WhatsApp, covering bottleneck identification, time management, storage and traffic estimation, and trade-off discussions such as WebSockets versus HTTP or when to add load balancers. It also explai

An explanation of how Uber's real-time map handles millions of GPS location updates per second by splitting the system around read patterns rather than writes. Covers an edge ingestion layer that validates, deduplicates, and adaptively samples pings before they hit a database; H3, Uber's open-sourced hexagonal geospatial index, which turns 'find nearby drivers' from a full scan into a bounded lookup; a fast in-memory store (Redis-like) for live rider map queries paired with a durable Cassandra-f

Engineers using AI fall into three camps: refusers who distrust it, overreliers who stop thinking critically, and multipliers who use AI to amplify their judgment. The piece offers concrete workflows for using AI to onboard into unfamiliar codebases (via Claude Code's /init, architecture mapping, testing strategy discovery, and setup checklists), to learn new languages/frameworks by mapping concepts from a known language, to build hands-on learning projects instead of watching tutorials, and to

An investigative breakdown examines Cloudflare's acquisition of Deno, revealing that the Deno runtime will get only one more year of bug fixes before development ends, with Deno Deploy shutting down 6 months after that. Commit data shows contributions collapsed over 90% between June and September 2026, with most remaining work done by two people, well before the October 9th announcement. The piece traces warning signs: a 2025 layoff of roughly half the staff that was never disclosed, shrinking f

LazyState, a Swift macro library from Point-Free, has graduated from beta and reached version 1.0. It solves a gap left by Apple's new @State macro: while Apple's macro fixed eager initialization for inline, static default values, it did not address state that must be dynamically constructed from data passed in by a parent view. LazyState provides a @LazyState property wrapper that lazily creates such state exactly once per view identity, avoiding the common workaround of making state optional a

An experiment with two Spring Boot instances using HikariCP against PostgreSQL shows that default pool sizing (minimumIdle = maximumPoolSize = 10) leaves 20 idle connections open during quiet periods, consuming about 57-58 MiB of PostgreSQL cgroup memory. Setting minimumIdle to 1 with a short idleTimeout let idle connections retire down to 2, cutting PostgreSQL cgroup memory to roughly 28-29 MiB. A control profile confirms that starting with a low minimumIdle alone doesn't help — only enabling i

A detailed benchmark study measures the same Benchmarks Game programs across every Go release from 1.2 to 1.27 on identical hardware. Go 1.27 is found to be 11.5% faster than Go 1.23 and roughly 41% faster than Go 1.2, with most recent gains coming from runtime changes (Swiss-table maps, math/big, cgo, GC, memory allocator) rather than numeric code speed, which hasn't changed since Go 1.7's SSA compiler backend. Using an energy model, Go 1.27 is estimated to use 11.5% less energy than Go 1.23. T