Ask HN: Best learning path to build real apps with AI tools in 2025?

I’m working on MakersHub.dev, a community platform for makers learning to build with AI tools. As I talk to more beginners, I keep hearing the same question: “Where do I actually start?”

The landscape has changed dramatically. Tools like Cursor, Claude Code, and Replit, are making it possible to build functional apps without years of traditional study. But there’s still a gap between “I can generate code” and “I can build and ship a complete app with auth, a database, and payments.”

For someone starting from absolute zero today—never used a terminal, doesn’t know Git, hasn’t written a line of code—who wants to build real web/mobile apps (think: a SaaS with Stripe integration, user authentication, and a database), what’s the most effective learning path? Specifically, I’m curious about:

• What foundational skills are non-negotiable? (terminal basics, Git, understanding APIs, etc.)

• What fundamentals should you learn deeply vs. what can you pick up on-demand with AI assistance?

• Should beginners start with a specific stack (Next.js, Supabase, etc.) or focus on concepts first?

• How do you balance learning through AI tools vs. understanding what’s happening under the hood?

• Realistically, how many hours/months does this take in the AI-assisted era? (Are we talking 100 hours? 500? More?)

• What’s the minimum viable knowledge to go from idea to deployed app?

Would love to hear from both experienced developers and recent learners who’ve made this journey.