Solveit – A course and platform for solving problems with code

Original Article ↗ Hacker News Discussion ↗

What Solveit Is (Course + Platform + Method)

  • Described as a 5‑week course teaching a problem‑solving methodology (coding, writing, sysadmin, research) plus access to a custom AI-enabled environment.
  • Creators emphasize it is not a “learn the tool” course but a structured way to think, iterate, and learn with or without AI.
  • Several participants summarize it as “AI‑assisted literate programming” or an “intelligent notebook” that can go from exploration to full apps.

Human-in-the-Loop Philosophy

  • Strong focus on small, fast iterations, deep understanding, and reflection; explicitly framed as the opposite of “vibe coding” and one‑shot agentic workflows.
  • AI is presented as an optional helper for learning and feedback, not as an autonomous code generator; some users report using the AI less over time.
  • Emphasis on preserving human agency and avoiding dependence and “slot-machine” patterns of waiting for large AI dumps of code.

Platform Features (as Described)

  • Combines chat with an LLM, a notebook-like interface, Monaco editor, a persistent Linux VPS with URL, terminal, and Claude Code‑style tools.
  • Novel pieces claimed: turning any Python function into an AI tool, referencing live variables in prompts, context editing (editing AI’s answer directly), metaprogramming the environment, and real‑time collaborative notebooks.

Pricing, Scope, and Fit

  • Course costs about $400 for 5 weeks, including platform access for the duration plus a short tail; no usage quotas.
  • Time expectation: ~4 hours homework + 3–4 hours videos per week. Recordings available for asynchronous participation.
  • Creators say it’s not just for juniors; mention experienced engineers, academics, and senior leaders in the first cohort.

Enthusiastic Feedback vs Skepticism

  • Multiple first‑cohort participants report that Solveit changed how they program and learn, helped them ship real projects, and improved understanding of their code and domains.
  • Others see it as an overhyped coding course with AI “training wheels,” question the need for 5 weeks to learn a tool, or call it “a grift” and “consultant‑like.”
  • Some argue the platform is essentially “Jupyter + chat” and not revolutionary; others say the integration and workflow are uniquely effective.

Communication, Marketing, and Trust Issues

  • Many readers say the original article was unclear, burying that this is primarily a course; creators later add a clearer TL;DR.
  • The testimonial page (many quotes per person) and a wave of positive comments from low‑history accounts lead to accusations of astroturfing; moderators intervene but note this may be genuine enthusiasm from a tight community.
  • Several commenters suggest the team needs better language, positioning, and product/marketing communication, especially for people with AI fatigue or limited time.