SanDiegoNick

What makes a serious home AI setup great? What are you integrating together. Are you testing? Benchmarking?

@nyk_builderz I use deep research with Chat GPT, dump into .MD, then feed Opus for designing PRDs, requirements and testing, then sprints for incremental features to pass coding agents. I passed a 13-sprint plan to codex and apart from secrets input for APIs, it absolutely crushed it.

@Nick__Losee Imo Claude has to be the brain and orchestrator - codex is good for implementation but lacks reasoning

I’m running both Claude and Codex heavily, and one thing is clear: Usage limits are now a product surface, not a billing detail. If your workflow depends on one model, you don’t have an AI stack yet. You have a single point of failure. What’s your fallback stack today?

@Palak3312 Ask for UI from codex and you’ll still be turning back to Claude

Unpopular opinion: $20 on Codex goes a long way, but $20 on Claude is gone before you even notice.

@HowToAI_ @HowToAI_ What do you think are the implications of pointing agents at this tool with arduino guides/textbooks for reference? Custom Computer Vision use cases would be cool

I just found a website that lets you create, program and test electronic hardware. It even has pre-built projects to experiment and learn before building them in real life. 100% free.

@techhdive @bekacru I appreciate this take

design taste matters, but design curiosity might matter more. engineers who think they lack taste just haven’t spent enough time noticing why something feels off. with ai, you can now describe a problem, generate multiple solutions, and refine your judgment through iteration. it is something you can learn much faster now than before.

I never realized how rare good design is in software. If you’ve good design taste, you’ve incredible edge with AI

@Bhavani_00007 Claude: this sprint plan should have an MVP by week one and 4-6 for the remaining features Claude code: 😏

No Claude, this isn’t a 3-month project. We will finish it today☺

@devoznunes Free Speechify + Audible = VoxCast

Drop your startup 👇 I’ll pick the best and promote them for free to my 9k+ members community

@staysaasy @staysaasy I was there a bit ago, now noticing it more and more. What’s your next move?

It takes some real mfuckin courage to drop an obvious AI slop document on your manager's manager's desk. Respect to the person who just did it to me. This isn't gonna go particularly well for them, but respect all the same.

My wife asked why there’s not a claudette

@ashpreetbedi @ashpreetbedi I studied Systems Engineering and 100% agree… handoff to agents comes with more peace of mind if you build and give it SE design documentation, requirements and models for them to run with

New post: Systems Engineering Coding agents have lowered the barrier to writing code, but they haven't lowered the requirements of production software. Agentic software is just software. The agent replaces business logic. Everything else is the same. ashpreetbedi.com/articles/syste…

How are Businesses on Enterprise accounts mitigating risks for things like cowork and what this is?

@BawsaXBT Yes. Just depends if you’re building for others or self In production systems, updates shouldn’t get pushed immediately to Prod. Updates might sit on Dev server for days/weeks before being tested thoroughly and features are smooth. Regardless things might slip through cracks….

if you're new to vibe coding, this post is for you: i've been vibe coding for 4 months now. my only method of testing? testing locally. if it looks good, it works, cool. push to production. but, not once did AI tell me: "hey, you should probably stop doing that." so i finally setup a staging environment this month. > separate vercel + supabase > separate branches > separate database here's what it's changed for me: > localhost isn't testing. it's a demo for yourself. > staging is where you actually catch breaks. > shipping to real users with no middle step? that's gambling, not building. vibe coding gets you to "it works on my machine." but it doesn't get you to "it works for everyone." tldr; build a staging environment asap.

@NitinthisSide_ Genuinely liking the graphics for added info/subject overview

@Nick__Losee Sureee

🧵 Day 1/30 — #SystemDesign Most systems don’t fail because of bad code. They fail when traffic increases and architecture can’t handle it. Let’s start with the foundation of scalability: Sync vs Async systems.

In multiplayer games, a lobby feels a lot like load balancing in disguise: thousands of players hit “join,” and the system spreads them across matches, regions, and server capacity so one server doesn’t get crushed while others sit underused. @NitinthisSide_

Had an incredible conversation today about what unlocks with these coding agents for manufacturing and operations. - “Outdated” hardware Drivers and interfaces can be recreated. - Integrations can be built and automated into SCADA type dashboards - Notifications‼️ & Signals

@HeyYoGB What’s your opinion on Ruff and Vulture for dead code?

claude code tip: never paste your whole codebase and ask it to "fix this" instead: • describe the one thing that's broken • tell claude what you expected to happen • tell it what actually happened context wins every time. vague requests lose every time.

@Shekswess Designed a PRD and 13 sprint plan for codex and apart from secrets input for APIs into Supabase, it crushed it.

Codex is running non-stop for 6 hours without any intervention of mine. I love it

@quionie @diegobernabest So refreshing. Mine is predictable lol, I like your wow factor.

@diegobernabest your “About” is really cool! creative

if you vibe coded your own personal website / portfolio (not a business), share them here with me i like to see people’s tastes