Sabitlenmiş Tweet
I share practical backend engineering insights:
• System design breakdowns
• Clean architecture patterns
• Dev tools that save time
• Real-world coding tips
Follow if you want to build scalable systems.
English
Phoenix.dev
363 posts
Phoenix.dev
@0xphxdev
Senior Full-Stack Engineer AI tools • Coding tips • Dev workflows Build faster. Save hours 🚀
🌍Internet, ~/root/h Katılım Aralık 2020
373 Takip Edilen284 Takipçiler
Vibe Coding 101:
> Learn how to code
> Learn how to code
> Learn how to code
> Learn how to code
> Learn how to code
> Learn how to code
> Learn how to code
> Learn how to code
Never skip the core fundamentals.
English
@thesayannayak Haha, right? It's like the new buzzword everyone’s throwing around now.
English
@lochan_twt Hackathons can be tough, but it's all about learning, right? Just keep at it!
English
Your service depends on concrete classes.
Now you can’t test or swap implementations.
That’s a Dependency Inversion problem.
Fix:
Depend on abstractions.
Bad:
UserService → MySQLRepo
Good:
UserService → IUserRepo
Now you can:
• mock easily
• switch DB
• scale safely
English
In my experience, Claude is best for agentic work and long-running tasks.
Some people say GPT Codex is better, but this is just my personal opinion.
Feel free to share yours.
Ayushi☄️@iyoushetwt
is there any AI model that can replace Claude ?
English
@svpino I also follow this pattern now... Every module keeps 90%+ test coverage, testing as much as possible.
English
The funny thing is, I'm writing more tests than ever since I've been writing more code with AI.
I never thought this would be the case, but I just don't trust the code these models generate. Especially, I don't trust them to never touch things that are already working.
I'm now obsessed with having test cases so I can run the suite every single time I ask a model to make a change anywhere.
English
@justbyte_ It depends on the project size. If the authorization is complex, I would rather choose a clerk; if it's simple, then a custom one. Only for authorizing custom or Supabase are okay.
English
@0xhashlol @nextjs Umm, remote caching is less used. Mainly, I am not familiar with it. I use TurboRepo, but okay, I will try remote caching. Thanks for the suggestion.
English
Next.js 16.2 is a major improvement. I kindly request the @nextjs team to consider enhancing build speeds for monorepo codebases as well.
Next.js@nextjs
Turbopack: What's New in Next.js 16.2 • Up to ~375% faster updates with Server Fast Refresh • Improved Web Workers support • Subresource Integrity (SRI) support • Tree shaking for dynamic 𝚒𝚖𝚙𝚘𝚛𝚝() • 200+ bug fixes nextjs.org/blog/next-16-2…
English
Gemini 3.1 Pro vs Opus 4.6 thinking vs GPT 5.4 via API vs Grok 4.20
Harshith@HarshithLucky3
Gemini 3.1 pro in Gemini App vs Gemini 3.1 Pro in Google AI Studio same prompt
Indonesia
With @Grok, we keep the honest versions and kill the bad transformers (I believe they are called “Decepticons”)
I,Hypocrite@lporiginalg
This is fine.
English
@0xphxdev @Angula_007 @thatechiegurl @calixtaa_draws @theRedEra_ @WiselyRsa @iammaeve_ @JacobEsthe44188 @IAM_Jacci @iamdKamisama @oxbashir yesss pretty good 2 cups of coffee ☕️ and you? 🫠
English
@thatechiegurl @Stevenchyyy @urspeaks @calixtaa_draws @Angula_007 @JacobEsthe16697 @theRedEra_ @WiselyRsa @Linhlinhnguyenx @443Shots @nakedw1090 @0xphxdev Yeah I got suspended temporarily
English
Phoenix.dev retweetledi
GPT-5.4 mini is available today in ChatGPT, Codex, and the API.
Optimized for coding, computer use, multimodal understanding, and subagents. And it’s 2x faster than GPT-5 mini.
openai.com/index/introduc…
English