Skullthoughts
286 posts

Skullthoughts
@skullthought
love to share everything .
Pluto ... เข้าร่วม Ocak 2012
767 กำลังติดตาม33 ผู้ติดตาม

Do something different this weekend.
Become a PRO in AI Model Fine-tuning.
Paste this prompt in Codex/ChatGPT/Claude/Grok.
"You are an expert AI engineer and teacher.
Your job is to teach me modern LLM engineering and fine-tuning concepts from beginner to advanced level using very simple daily-life language.
Teach me step-by-step like a real mentor. Assume I am smart but new to the topic.
Foundations:
- LLM basics
- How AI models work
- Tokens
- Tokenization
- Context windows
- Embeddings
- Transformers
- Attention mechanism
- Parameters
- Training vs inference
- Open-source vs closed-source models
Datasets & Training:
- SFT datasets
- Instruction tuning
- Preference datasets
- Synthetic datasets
- Data curation
- Dataset cleaning
- Dataset formatting
- Fine-tuning basics
- Continued pretraining
- Hallucination reduction
Fine-Tuning:
- LoRA
- QLoRA
- DPO
- RLHF
- Quantization
- Model checkpoints
- Adapter tuning
- GGUF models
Inference & Optimization:
- KV cache
- Flash Attention
- Speculative decoding
- Inference optimization
- Model serving
- Batch inference
- GPU basics
- VRAM basics
- Latency vs quality tradeoffs
Local AI Ecosystem:
- llama.cpp
- Ollama
- vLLM
- MLX
- Hugging Face
- Unsloth
- Axolotl
- PEFT
- TRL library
RAG & Memory:
- RAG
- Vector databases
- Chunking
- Retrieval pipelines
- AI memory systems
- Semantic search
Agents & Workflows:
- Prompt engineering
- System prompts
- Tool calling
- Function calling
- AI agents
- Agentic workflows
- Multi-agent systems
- Browser agents
Model Types:
- VLMs
- SLMs
- Dense models
- MoE models
- Coding models
- Reasoning models
Deployment:
- Local inference
- On-device AI
- API serving
- Cloud GPUs
- Edge AI basics
Evaluation:
- AI benchmarks
- Human evals
- Cost-per-token analysis
- Speed benchmarking
- Quality benchmarking
Real-World Skills:
- Building chatbots
- Building AI copilots
- AI automation
- AI SaaS workflows
- AI coding workflows
- AI orchestration systems
- AI product thinking
Start from the absolute basics and gradually make me advanced.
Rules:
- Use simple English only
- Avoid academic jargon unless necessary
- Explain every difficult word in plain language
- Use real-world analogies and daily-life examples
- Use small code snippets when useful
- Show practical use cases
- Compare concepts side-by-side when helpful
- Teach from fundamentals first, then advanced concepts
- At the end of each topic:
- give a short summary
- give a simple mental model
- give beginner mistakes to avoid
- give a small exercise/project
I want deep understanding, not memorization."
Thank me later.

English
Skullthoughts รีทวีตแล้ว

@valigo Telling me that you didn't have brain without telling me.
English

@BambulabGlobal apologies for the shit take the hit and make it full open source the code and let there be 2 plan one local by the user with extra 3 dolor or something like this and one with your server with less mony so you will be appealing to the crowd dont be stupid -advice
English

@TechBit_M Everything تجربت سنين ماتوقع كنت اقدر استعمل الوندز بدونه
F.lux جميل لراحت العيون
Brave متصفح
العربية

@f_aswadi استخدمهم ثلاثتهم بس اجلد وندوز واسب ماك واحب لنكس
العربية

@mcuban They will just add the tax on us lol this a sick idea they have enough evil bad ideas please dont feed them more they will do it
English

We should federally tax Tokens at the Provider level.
Not a lot. Less than 50c per million tokens.
It will accomplish 4 things (at least )
1. It will push the big AI players to optimize tokenization, caching , routing and localization
Which will
2. Reduce energy usage. Saving them in energy costs more than what they paid in tax and reducing strain created by the growth in energy consumption
Which will
3. Generate maybe 10 billion dollars a year to start, but over the next ten years could grow 30x to 100x
Which will
4. Create a source of funding to pay down the federal debt or deploy, in response to the things AI brings that we don’t expect or don’t like
At some point the models will pass it on to customers. Of course. That’s ok. Customers will have the ability to choose between providers. Or to do everything using open source models locally.
Thoughts ?
English

للأندرويد 🔥
إذا تحتاج توقف أي تطبيق من استخدام الإنترنت بالكامل فهذي الأداة بتعجبك جدًا
FireWall-Blocks تعطيك تحكم كامل
• تمنع التطبيقات من Wi-Fi
• أو بيانات الجوال
• أو الإنترنت بالكامل
وكل هذا بدون روت
يعتمد على Shizuku لإعطاء صلاحيات متقدمة بدون روت يعني أداء أخف وبطارية أفضل
مفيد لمنع
• التتبع والإعلانات
• التطبيقات المشبوهة
• استهلاك الإنترنت بالخلفية
• التطبيقات المعدلة من إرسال بياناتك
ومفتوح المصدر بالكامل


العربية

@REMIX_KSA هرمز اجمل واقوى شوف نتايج استعماله الناس تركو تقريبا كل انواع المساعدين الثانيه للستخدام البسيط والمتقدم حاليا استخدمه بالتلفون داخل تيرمكس ترمنل

العربية

@MulhamH لازم يضخون هالكلام والملفات لابعاد الرأي العام عن بلاوي ابستين . تبا لهم
العربية

الملفات والوثائق الخاصة بـ UFO التي رفعت عنها السرية الإدارة الامريكية لا تقدم ولا تؤخر في الاثبات العلمي لوجود كائنات فضائية
فهي اما :
-شهادات شفهية غير مثبتة
- صور غير واضحة
- فيديوهات غير واضحة
فسنوات من التحقيقات لناسا و وزارة الحرب و المخابرات.... والاستنتاج
- اجسام طائرة مجهولة
- ظواهر جوية شاذة
ولم يصلوا لتأكيد ماهية تلك الأجسام
ثم استغرب خلال ساعات مغردين اثبتوا وجود كائنات فضائية


العربية

@TechBit_M @BlindPassenger_ ممتاز كا os كامل بجربه على الاب قرييا شكرا
العربية

@fjzeit I dont have a Q but i wanna thank you for the answers you have given it just so great, thanks again
English

@Red1OneX اذا انت بتبني المساعد الخاص فيك وتبي تربط بين اكثر من مساعد agent
اقرا عن a2a protocol
جديد بس افضل بواجد من الموجود ومصمم خصوصا لهم
العربية

@ThePrimeagen They have shit show of any terminal make everything harder i hated mac when i get one more i was forced to development of an app to my family to use case but its hill i love linux more day by day
English

@shikhr_ Thinking glm 5.1 or deepseek pro
For fast simple work deepseek flash or glm 5 turbo
Qwen 3.6 for agent like harmes i love it
English





















