Link parkin’: Amp
Why Amp?
Amp is a frontier coding agent for your terminal and editor, built by Sourcegraph.
- Multi-Model: Sonnet, GPT-5, fast models—Amp uses them all, for what each model is best at.
- Opinionated: You’re always using the good parts of Amp. If we don’t use and love a feature, we kill it.
- On the Frontier: Amp goes where the models take it. No backcompat, no legacy features.
- Threads: You can save and share your interactions with Amp. You wouldn’t code without version control, would you?
Amp has 2 modes: smart (unconstrained state-of-the-art model use) and free (free of charge, using fast basic models).
Want to go much deeper? Watch our Raising an Agent podcast that chronicles the first few months of building Amp, and see our FIF.
This post was started before a recent edition of Hamel Husain’s newsletter arrived in my transom today (apologies, can’t find a Web page version to link to). He’s positive on Amp as well:
Even though Cursor is my daily driver, I frequently try new coding agents. One that’s stood out for me over the last few months is Amp. Amp is the Omakase experience of coding agents: the team maniacally experiments with the frontier of coding models, and tweak the tools, system prompt, etc accordingly for you. For example, over the past week they moved the default model from Gemini 3 to Opus 4.5. (BTW all of this is backed by rigorous internal evals, which allows them to move quickly).
Typical of Husain, his email contains much more detail, including caveats for working with Amp. For example, the version of Amp that he uses is priced on a per-token basis. This can get expensive if you’re not careful. Also, he reported that the Amp CLI had problems with iTerm, but works well in Ghostty. Hopefully it’ll be made available somewhere easy to reference.
The field of terminal user-interface coding agents is getting crowded, but Amp is worth watching.
Built using