Marián Mrva

@peterhajducek Projekt !== produkt. Vibe kód matlaniny niesú održatelné dlhodobo.

15 hodin mi zabralo s Codex vytvorit a deploynut funkcnu HR appku na hiring. Ma Google prihlasovanie, Slack notifikacie, verejnu cast pre kandidatov, odosiela a prijma emaily od kandidatov a ma kompletne funkcny backend s rolami a opravneniami. Hype is real.

@aidenybai Long task agents

tbh i still don't understand the hype around terminal coding agents it's just objectively worse UX

@leerob @iannuttall @AnthropicAI @claudeai Wheeen?

@iannuttall @AnthropicAI @claudeai Cursor will support it as well 😄

This is a petition to get @AnthropicAI / @claudeai code to support the .agents/skills standard, just as these have: - amp - codex cli - factory/droid - gemini cli - gh copilot - kimi cli - opencode - vscode - etc Like this post if you want them to support .agents/skills ✊

@karthikponna19 Codequality

i am a Vibe Coder, scare me with one word

@DynamicWebPaige @antigravity @GoogleAIStudio @AnthropicAI @OpenAI @cursor_ai

⚙️ Really love the Toolkit for @Antigravity extension (available on the OpenVSX marketplace). It keeps track of your token consumption for the @GoogleAIStudio Gemini models, @AnthropicAI's Opus and Sonnet models, and @OpenAI models - and shows you hour-by-hour stats for each.

@AliGrids Left

Left or right? -which would you ship?

@askwhykartik Figma-to-code… show me one tool that do that with 100% accuracy

Frontend-only jobs are dying. AI builds landing pages in 10 minutes. Figma-to-code? Automated. Responsive design? One prompt. If you're ONLY doing frontend, you're replaceable. Go full-stack. Learn backend. Add AI to your skillset. That's survival.

@mds I wonder how does it look like after 1 year of addding lines 😀

read more about this here floatprompt.com

Next time you use AI to work with a large amount of information: - transcript - notes - documents - whatever Use this as your first "system" prompt of sorts to help you process the information better. raw.githubusercontent.com/mds/floatpromp…

@ericzakariasson How do you switch cursor to this full screen view ?

generate diagrams and visualize code with this command cursor.com/link/command?n…

@abhi1thakur Do you store any sensitive informations there? Before you fix issue with ignoring AI, maybe it shouldn’t be part of the repo:)

WTF!?

@wesbos I got a billion dollar idea. Let’s verify what LLMs are generating before pushing to repository.

What the HECK is going on with tech? In the last week: Multiple cloud outages, x DMs totally broken, antigravity doesn't work, my watch is showing me 15 year old cal events, mac OS is a mess, email is spammed to hell and every nerd on here is talking like new AI is the second coming

@housecor No need to provide context manually

@housecor True. I’m positively surprised how Cursor index npm dependencies even though they are not public. Definitely better than few months ago

The strange thing about coding in 2025: I reference docs less, and prompt more. Example: I need to configure ESLint to disallow using console.log. In the past I would have Googled, referenced docs, copy/pasted the relevant config, then tested. Now, I just tell Claude to do it.

@rauchg @shadcn @turborepo @DrizzleORM * paid partnership with Vercel xD

Green flags in a React codebase ✅ Next.js ✅ Tailwind ✅ @shadcn ✅ 𝚋𝚞𝚗.𝚕𝚘𝚌𝚔𝚋 / 𝚙𝚗𝚙𝚖-𝚕𝚘𝚌𝚔.𝚢𝚊𝚖𝚕 ✅ @turborepo ✅ @drizzleorm

@eiselems @catalinmpit Outsorce security side to service - Clerk

@catalinmpit What are you regularly using for Auth?

Perfect stack for 99% of the apps: - SPA: React, TanStack Router, TailwindCSS - Backend: Bun, Hono, Postgres - Misc: TypeScript, Zod, Vite - Marketing: Astro I don't know why we keep complicating stuff.

@ryolu_ @crsmoore @cursor_ai @grok what is composer-1 and why should I care?

@crsmoore @cursor_ai composer-1

the best way to code right now: • @cursor_ai 2.0 • composer-1 for live coding • gpt-5 high for plan mode • gpt-5-codex for long-running cloud agents • best n-models for hard / ambiguous tasks

@housecor @AlGMacD Do you use preview environments? If so, what kind of Devops service ?

@AlGMacD That’s an interesting take. Never seen that done but can see how it could work. I generally prefer embedding QA with dev so they can collab on tests and approve before merge

Many companies believe they need traditional long-lived environments like Dev, QA, UAT, and Staging. But maybe not. There are many alternatives to consider. Could we eliminate our Dev environment by coding on our local machine and using CI to validate the build before merge? Could we eliminate our QA environment by embedding QA with development so they can collaborate and approve *before* merge? Could we eliminate our UAT environment by automatically generating a preview deploy for each branch so stakeholders can review before merge? Could we eliminate our Staging environment by “testing in prod” via feature flags, canary deploys, dark canaries, and blue/green deploys?

@housecor This is trade-off of an open source. You win some you loose some, it’s always balance.

Each time I open up an old JS project, I get dozens of breaking changes when I update npm packages. It's wonderful that all these open source projects are actively supported, but the tradeoff is every app that uses npm quickly "rots".

@housecor The main struggle with local mocks are mutations 💩

If your app depends on many services supported by other teams, how do you handle that during dev? Do you mock them? Do you run them locally? Do you call each service's staging environment? Do you call the corresponding environment? (Call dev from your dev, QA from your QA…)

@housecor - creating mock with msw - using mock for testing Edge cases and for automatation tests - test against real env before creating PR