@bytervin

Aku MLE yang lagi deep dive di dunia AI & data engineering. di akun ini bakal share: - hal-hal yang aku pelajari tiap hari - mini tutorial & code snippets - opini soal AI yang (mungkin) unpopular follow kalau kamu juga suka ngulik tech 🧵

@claudeai This is good! I'm from Indonesia and I hope I get selected for this

We're launching Claude Community Ambassadors. Lead local meetups, bring builders together, and partner with our team. Open to any background, anywhere in the world. Apply: claude.com/community/amba…

@OpenAI wow I think I want to try use codex right now

GPT-5.4 Thinking and GPT-5.4 Pro are rolling out now in ChatGPT. GPT-5.4 is also now available in the API and Codex. GPT-5.4 brings our advances in reasoning, coding, and agentic workflows into one frontier model.

@Alibaba_Qwen very impressive

🚀 Introducing the Qwen 3.5 Small Model Series Qwen3.5-0.8B · Qwen3.5-2B · Qwen3.5-4B · Qwen3.5-9B ✨ More intelligence, less compute. These small models are built on the same Qwen3.5 foundation — native multimodal, improved architecture, scaled RL: • 0.8B / 2B → tiny, fast, great for edge device • 4B → a surprisingly strong multimodal base for lightweight agents • 9B → compact, but already closing the gap with much larger models And yes — we’re also releasing the Base models as well. We hope this better supports research, experimentation, and real-world industrial innovation. Hugging Face: huggingface.co/collections/Qw… ModelScope: modelscope.cn/collections/Qw…

@HuggingModels wow this is very useful for my OCR Project

Meet a specialized AI detective for Indonesian text. This BERT model doesn't just understand Indonesian, it identifies and classifies key entities like names, organizations, and locations within sentences. It's a game-changer for processing Indonesian language data.

@YutaWeb33 Hi

tips I use daily: - give a specific role - show example output - set constraints ("respond in JSON") - write "STRICT RULES" for non-negotiables simple but game changer.

Most underrated skill in AI development: prompt engineering. the difference between a lazy prompt and a structured one is like talking to an intern vs a senior engineer.

@hiarun02 to replace themselves in the future. but right now, AI can't debug it's own hallucinations. So they still need engineer to babysit the thing they're building the irony writes itself

AI will replace engineers. Cool. Then why are AI companies hiring engineers?

model I recommend to start: llama-3.3-70b solid for summarization, analysis, and extraction. groq.com 🔥

just tried Groq for LLM inference response time is sub-second. literally faster than me typing the prompt. the crazy part: it's free. if you're building a side project with LLMs on a budget, try Groq before paying for OpenAI.

@XcodingwaliGyrl no, writting code on paper only tests your memory, not your problem solving. a better developer isn't someone who memorizes syntax it's someone who knows what to build and why. spend that time reading other people's code instead. you'll learn 10x more.

does writing code on paper actually makes you a better developer? 🤔

@SNagarani1419 python because this language can act as a generalist

Be honest, which is the easiest programming language to learn??

@gxjo_dev probably yes, because some companies in some countries start to stopped hiring juniors. But the one who survive are the ones who learn to work with AI

Did AI kill junior developer jobs already ?

@Bhavani_00007 from the premium services and ads (maybe)

just found out iLovePDF is a million-dollar company. How is it even making money? 😭

@NanouuSymeon the hardest part is not about writting code. It's deciding that's good enough

What’s the hardest part of being a developer that nobody talks about?

@akramcodez antigravity is good for free

Be honest, which AI coding tool do you use most of the time?

@Adidotdev for me 16 GB is enough

as a Developer, how much RAM is enough for you ? 1) 8 GB 2) 16 GB 3) 24 GB 4) 32 GB 5) 64 GB 6) 128 GB+

@GoogleAI Awesome can reach out 77.1% on ARC-AGI-2

Introducing Gemini 3.1 Pro 🚀 3.1 Pro represents a major step forward in core reasoning. It scored 77.1% (more than doubling 3 Pro’s score) on ARC-AGI-2, the benchmark that evaluates a model's ability to solve new logic patterns and work through challenges it hasn’t encountered before. This demo illustrates how the model can go beyond the prompt. Instead of rendering a video or static graphic, 3.1 Pro codes a full environment, integrating generative audio and providing UI controls.

@PythonPr the indentation

What is The Difference ❓Comment The Output