Claude Code: what developers actually say
We read every Hacker News comment from Nov 2025 – Nov 2026 that gave a real opinion on Claude Code. 2,397 of the 3,026 opinions come from people who say they use it themselves.
Claude Code scores 54% approval across 3,026 opinions from 2,204 authors, and 2,397 of those opinions come from people who used it directly. Praise centers on autonomy (280 positive to 88 negative) and speed (125 to 44), while reliability is the weak spot at 107 positive to 296 negative. Approval has slipped over time: 68% in November 2025, a low of 42% in April 2026, and 51% across the last 90 days. In comments that compared tools, more developers preferred Codex (181 to 62) and OpenCode (73 to 42) instead, while Claude Code came out ahead of Cursor (61 to 42) and GitHub Copilot (40 to 18).
Pick it if
- Want an autonomous agent that keeps working with little steering
- Prize fast turnaround over a polished first draft
- Plan to review architecture yourself and treat output as a starting point
Look elsewhere if
- Need dependable, repeatable runs on production code
- Run into usage limits often or need predictable spend
- Want to design systems rather than patch them
- Prefer a GUI over a terminal workflow
Strengths and weaknesses
Each row counts opinions that mentioned the topic. Bars to the right of the line are praise, bars to the left are complaints.
How opinion has moved
What developers say
Praise
Complaints
Summaries are paraphrased. Each links to the original comment.
Claude Code compared
Best Claude Code alternatives, according to developers who switched →
FAQ
Is Claude Code worth it?
Approval sits at 54% from 3,026 opinions, with 2,397 from direct users. Developers rate it highly for autonomy and speed but complain most about reliability, usage limits and cost. It is a reasonable pick for exploratory or supervised work, less so where you need consistent, repeatable output.
Is Claude Code better than OpenAI Codex?
In comments that compared the two, 181 developers said they preferred Codex and 62 said they preferred Claude Code. Codex is also the most cited alternative overall, ahead of OpenCode with 73 mentions. The data favors Codex on preference, though Claude Code still leads against Cursor (61 to 42) and GitHub Copilot (40 to 18).
What do developers dislike about Claude Code?
Reliability draws the most negative feedback, at 107 positive to 296 negative, followed by usage limits at 64 to 164 and cost at 128 to 185. Some users describe the tool as slow and bloated, and report that limit hits block even simple requests. A few say output is clean but poorly architected, which pushes review time up.