Amp: what 25 developers actually say.

Sourcegraph · Terminal agent · Updated Nov 2026

25 Hacker News opinions from 24 developers, Nov 2025 – Nov 2026. 22 come from people who say they use it. Here is what they praise, what they complain about, and who switched.

25 opinions24 developersNov 2025 – Nov 2026Methodology

Approval

90%

±9.9 at 90% confidence

Last 90 days

–

No recent data

Rank

of 22 tools · CI-adjusted

Opinions

22 first-hand

Developers

1.0 opinions each

The short version

The case for Amp, in one paragraph.

Amp earns 90% approval across 25 opinions. Developers are divided on it.

Visit Amp ↗Compare with Claude Code →

Pick it if you

Want a polished terminal UI and quick setup
Like visible costs and automatic model selection
Will review generated code closely and want speed

Look elsewhere if you

Need low, predictable spending
Expect clean code without close review
Rely on handoff, which Amp removed

Where it wins and loses.

Mentions across 25 opinions. Bars are counts, not votes; a tool can be both praised and criticized on the same topic.

TopicComplaints ← → PraiseMentions

PraiseComplaint

0 topics shown · sorted by net sentiment · source: 25 opinions

View data table

Topic	Praise	Complaints	Mentions

How opinion has moved.

Monthly approval, Dec 2025 – Mar 2026. Bars show opinion volume; low-volume months are still included but faded.

62% Dec 202562% Dec 2025 low90% Mar 2026– last 90 days

Monthly approval, 0–100. Bars show opinion volume. Months with fewer than 3 opinions are left out.

View data table

Month	Approval	Opinions
Dec 2025	62%	4
Jan 2026	100%	5
Feb 2026	100%	4
Mar 2026	90%	5

What developers say.

Summaries are paraphrased. Each links to the original Hacker News comment.

Praise

4 selected

They loved Amp's UI/UX and it strongly inspired the UI of their own harness.

HN comment · Sep 15, 2026Read on HN ↗

Cost & pricing

The Amp coding agent alone is worth its $20 monthly price.

HN comment · Aug 30, 2026Read on HN ↗

Likes Amp's plug-and-play setup, visible costs, and auto-selected models.

HN comment · Jul 1, 2026Read on HN ↗

Amp's CLI was well designed, and that it once auto-picked the best model per task; hasn't used it recently.

HN comment · Jul 1, 2026Read on HN ↗

Complaints

1 selected

Cost & pricing

Finds Amp too expensive to use for anything useful, so stopped using it.

HN comment · Dec 8, 2025Read on HN ↗

Questions developers ask.

Is Amp worth it?

Most opinions are positive: approval is 90% across 25 opinions. Users like the UI, setup, and first-attempt accuracy. Cost is the main reason people leave.

Is Amp better than Claude Code?

In this sample, five opinions preferred Amp to Claude Code and two preferred it to OpenAI Codex. One person preferred OpenCode instead. The data does not include a count of people who chose Claude Code over Amp.

Why do developers stop using Amp?

Cost is the main reason. Users describe it as too expensive for useful work and say it burns through money quickly. Some also report bad code and regressions without close supervision.

The data behind this page.

Developers give Amp a 90% approval score across 25 opinions from 24 authors, and 22 of those authors speak from firsthand use. They praise the UI, quick setup, visible costs, speed, and first-attempt accuracy, and five opinions prefer it to Claude Code while two prefer it to OpenAI Codex. Cost is the main complaint: one user stopped because it was too expensive, and another said it burns through money quickly. Monthly approval stayed between 90% and 100% from January through September 2026, but the last four months have only one or two opinions each, so the recent trend rests on a thin sample.

Every number comes from public Hacker News comments collected through the HN Search API. A language model tagged each comment for sentiment, first-hand use, topic and switches; a human spot-checks samples. No vendor can pay for placement.

20,064 comments screened25 about Amp22 first-hand

Read the full methodology →