Ask HN: What Are You Learning?

Hacker News Discussion ↗

Programming, Systems & Tools

  • Many are learning new languages and ecosystems: Rust, Go, Haskell, Erlang, Elixir, Zig (considered but sometimes rejected), Swift/SwiftUI, Jetpack Compose/Android, React Native, Deno, Nix, Docker, Node, ncurses, Jetpack Compose Multiplatform, and game engines like Godot and Compose-based games.
  • Several focus on systems and performance: graphics programming, CUDA/JAX, eBPF, PCB design with embedded Rust, ESP32 IoT, Casey Muratori’s low-level performance course, precision frequency synthesizers, secure networking with MikroTik/VLANs, FPGA‑like “BitGrid” ideas.
  • Web dev is seen by some as increasingly commoditized and threatened by AI; advice trends toward becoming more “architect-shaped,” using AI as a tool, or specializing (e.g., WebGL, data visualization).
  • Rust inspires both enthusiasm (strong typing, tooling, new ways of thinking) and frustration (borrow checker, “impossible” patterns vs C). Alternatives like Zig are discussed but sometimes ruled out on licensing/ecosystem grounds.
  • Nix is explored for reproducible dev environments and container builds, with recognition that it adds complexity and might need higher-level abstraction.

AI, Data & Statistics

  • Multiple people are learning LLMs, especially RAG, building small agents or document chat tools with LangChain, Streamlit, OpenAI APIs, and embeddings.
  • Data engineering remains attractive for systems/design reasons; getting started advice includes Airflow + simple ETL pipelines.
  • Several are relearning statistics, often pivoting to Bayesian approaches (Jaynes, McElreath, Downey, bayesrulesbook). Frequentist methods are called historically messy; modern, simulation-based curricula are praised.
  • One ongoing stats book project aims to demystify the field; notebooks and concept maps are shared.

Math, Physics & Formal Foundations

  • People are working up from basic math to calculus and real analysis, often after years away from school, sometimes for career changes (e.g., into engineering).
  • Resources include community college courses, MIT OCW, various textbooks, concept maps, and Anki; some use LLMs heavily for step-by-step explanations, others warn about their math unreliability.
  • Classical mechanics, Lagrangian/Hamiltonian formalisms, proof-based math, type theory, real analysis, complex analysis, and Lean theorem proving are common study paths.

Hardware, Making & DIY

  • Many are learning welding (with serious safety concerns noted), toilet and plumbing rebuilds, fence building, mini‑split installation, woodworking, FreeCAD, analog synthesis, electronics (breadboards, NEETS, repair resources), PCB design, and ESP32 production.
  • DIY vehicle projects (Frankenbikes, recumbent/trike builds, four‑wheel contraptions) and home hardware hacks (OLED over I2C, RC plane flight computers) feature strongly, with communities and YouTube channels as key resources.

Languages, Music & Creative Skills

  • Language learning spans Spanish, Finnish, Mandarin, Japanese, Chinese, French, and English; tools include Anki decks, Dreaming Spanish, graded news archives, and custom scrapers.
  • Many are studying instruments and music production: piano (with apps for MIDI feedback), jazz guitar and theory, drums, band synthesizer work, hymn accompaniment, and DAW/composition.
  • Writing skills: fiction revision, game narrative, cybersecurity storytelling, communication/negotiation (Difficult Conversations, Getting to Yes, Nonviolent Communication), and self-help/psychology are frequent themes.

Career, Business & Personal Development

  • People are learning cloud platforms (AWS, Azure), options trading, bookkeeping/corporate finance, marketing/SEO, indie SaaS growth, reselling, GTM for first-time founders, and job-market reentry.
  • Several explore ADHD/ASD, time-tracking, contentment, leadership, and social engineering (out of fascination, not malicious intent).
  • A recurring meta‑theme: balancing breadth vs depth, using side projects as learning vehicles, and leveraging (but not fully trusting) AI to accelerate self-directed learning.