Your AI learning roadmap:
PATH #1: general AI (foundational knowledge that most neglect)
- what an LLM is & how it works
- types of LLM's & how they differ
- how LLM's are trained (pretraining vs. finetuning)
- token economy
- context windows
- effective context management
- prompt engineering
- RAG, embeddings & vector databases
- MCP
- awareness of hallucinations & bias
PATH #2: creative
- understanding multimodal capabilities of diff LLM's (e.g. gemini's video analysis)
- creating style profiles for consistency
- generating text & copy via. LLM's
- image generation (sora, midjourney, flux)
- video generation (kling, runway, higgsfield)
- speech/sound generation (elevenlabs, sesame, suno)
- tool-chaining
PATH #3: AI automation
- how AI automation differs from regular automation
- workflow automation tools (n8n, make, zapier)
- basic understanding of data structures (JSON)
- triggers, actions, credentials & API's
- error handling, data manipulation
PATH #4: vibe-coding
- best LLM's for coding (power vs. cost)
- best AI IDE's & how to use them (cursor, windsurf, replit)
- breaking down projects into step-by-step goals (an extension of prompt-engineering)
- a basic understanding of coding is beneficial
- handling errors & using AI to fix them
- front-end tools (v0, tailwind, shadcn)
Credit - Link
