CodeBucks⚡

35K+ YouTube subs | Software Engineer who ships AI apps | Scaling AI first development Agency | DM for AI SaaS development

[CodeBucks⚡ @code_bucks]

Your portfolio gets ~10 seconds to make an impression.

And most developer portfolios look the same, so I recorded a full tutorial building one that doesn't:

🔸day/night hero with live video backgrounds
🔸real-time visitor weather widget
🔸Smooth scroll animations
🔸an editorial design system (no shadows, no bold, one accent)

It's developed using next.js, tailwind CSS, and framer motion.

Full live portfolio and tutorial links are in the first comment 👇

(It's quite long so I can't upload it here on X)

[CodeBucks⚡ @code_bucks]

@soni_sahil_0 Yes 💯 library selection is much important for long term projects.

[CodeBucks⚡ @code_bucks]

"the llm said so" is not a reason.

> don't make decisions just because the llm said so
> don't believe it blindly just because the llm said so
> don't pick that framework or library just because the llm said so
> don't add extra complexity to your code just because the llm said so

every day i see people, including my clients and interns, trusting whatever the llm spits out.

use it for research. don't let it make your decisions.

my best work, apps or otherwise, came when i made the calls, not the llm.

people built great things before llms. now they just build them faster. the core is still the same: it's you and your skills.

someone who already knows a niche will always get more out of an llm in that niche than you ever will.

learning still matters. reading the right stuff and actually using your brain matters more than ever.

use llms blindly and you're just creating garbage you'll have to clean up later.

[CodeBucks⚡ @code_bucks]

@krl_grn That's the true beauty of open source. T3 code doesn't have to be everything for everyone. It's a solid starting point and anyone can take it in a completely different direction.

[CodeBucks⚡ @code_bucks]

Not sure why people are hating on T3 code. Most of the tweets I saw were just hate without any meaningful feedback or criticism.

I think 80% of people who tweet about AI aren't even using those apps/features.

[CodeBucks⚡ @code_bucks]

@Wv3EcACHtizRjZK It's not about what features of t3 code people are missing, I'm talking about the people who hate/criticize aren't exactly mentioning what they didn't like it's just pure hate that's it.

[CodeBucks⚡ @code_bucks]

@owenbjennings @buzzdotxyz would love to know whether it supports claude or openAI's subscription or not.

It doesn't make sense to pay API pricing if you already have subscription.

[CodeBucks⚡ @code_bucks]

@dannypostma This is 💯 true, I'm actually looking for someone who can do the marketing for my apps.

[CodeBucks⚡ @code_bucks]

@soni_sahil_0 Yes 💯

[CodeBucks⚡ @code_bucks]

Frontend without AI is fine.
Frontend without shadcn UI is pain.

[CodeBucks⚡ @code_bucks]

@thdxr @grok please list down all those acquision announcements.

[CodeBucks⚡ @code_bucks]

@soni_sahil_0 💯

[CodeBucks⚡ @code_bucks]

Anthropic models got so much ahead of others after release of Fable 5.

even latest models can't compete with the recent versions of sonnet and opus 5.5.

[Cognition @cognition]

We've evaluated Gemini 4 Carbon and Gemini 4 Argon on FrontierCode 2.0, our coding benchmark that measure's a model's ability to go beyond the spec.

Gemini 4 Carbon scores near SOTA with a score of 40% and a cost of $4.04 per task, behind Claude Opus 5.5, GPT-6 Astra, and Claude Sonnet 5.5.

Gemini 4 Argon scores 26.3% with a cost of $6.41 per task, ahead of Muse Wonder 1.0, but behind Claude Haiku 5.5 and Grok 4.7.

See the leaderboard with full scores: https://t.co/PPHwljDuaN

[CodeBucks⚡ @code_bucks]

"the llm said so" is not a reason.

> don't make decisions just because the llm said so
> don't believe it blindly just because the llm said so
> don't pick that framework or library just because the llm said so
> don't add extra complexity to your code just because the llm said so

every day i see people, including my clients and interns, trusting whatever the llm spits out.

use it for research. don't let it make your decisions.

my best work, apps or otherwise, came when i made the calls, not the llm.

people built great things before llms. now they just build them faster. the core is still the same: it's you and your skills.

someone who already knows a niche will always get more out of an llm in that niche than you ever will.

learning still matters. reading the right stuff and actually using your brain matters more than ever.

use llms blindly and you're just creating garbage you'll have to clean up later.

[CodeBucks⚡ @code_bucks]

@TCCardoza openAI should distill fable and opus to get better taste 😂🙌

[CodeBucks⚡ @code_bucks]

someone used CLAUDE to get their childhood game running for their kid.

A 1991 DOS game. A 3-byte patch.

I want more AI demos like this.

[CodeBucks⚡ @code_bucks]

@flutterwhat Well thanks for explaining that /s part.

[CodeBucks⚡ @code_bucks]

someone used CLAUDE to get their childhood game running for their kid.

A 1991 DOS game. A 3-byte patch.

I want more AI demos like this.

[CodeBucks⚡ @code_bucks]

@flutterwhat Do you think he should’ve preserved the authentic 1991 experience of it not working?

I don't think putting the 3 bytes back can help retro computing economy recover.

[CodeBucks⚡ @code_bucks]

@mehulmpt it's actually unexpected to see that google beating opus 5.5. openAI is taking it's sweet time to release their next frontier model.

[CodeBucks⚡ @code_bucks]

Here is the full post: https://t.co/Opp7jIEEKc