Phoenix.new – Remote AI Runtime for Phoenix

Ah man I'm really happy to see this and excited to try it out.

As an Elixir enthusiast I've been worried that Elixir would fall behind because the LLMs don't write it as well as they write bigger languages like Python/JS. So I'm really glad to see such active effort to rectify this problem.

We're in safe hands.

This last few weeks I've been going hard on LLMs to put together a new prototype project. I've exclusively been using Claude Sonnet 3.7 within Zed (via github copilot) and it' fantastic.

From time to time it tries to do something a little old-school, but nothing significant really. It's very capable at spitting out entire new features, even in liveview.

Over all the experience has very productive, and at least on-par with my recent work on similar sized python and nextjs applications.

I think because I'm using mostly common and well understood packages it has a big leg up. I also made sure to initialise the phoenix project myself to start with so it didn't try to go off on some weird direction.

Assuming you are on the $20 Zed plan? Has the 500 prompts/mo been sufficient for you? I'm debating between the Zed and Claude $20 plans-- no doubt I'd get better value from Zed's?

I'm on the free plan and have been using it via GitHub copilot instead, as the current project is a work one and they pay for that.

Before this I did a small project and I hit the 50 free tier limit through Zed by the time I was about 90% done. It was a small file drop app where internal users could create upload links, share them with people who could use them to upload a file. The internal user could then download that file. So it was very basic, but it churned out a reasonable UI and all the S3 compatible integration, etc.

I had to intervene a bit and obviously was reviewing everything and tweaking where needed. But I was surprised at how far I got on the 50 free prompts.

It's hard to know what you really get for that prompt limit though as I probably had a much higher number of actual prompts than they were registering. It's obviously using some token calculation under the hood and it's not clear what that is. All in all I probably had about 60-70 actual prompts I ran through it.

My gut says 500/mo would feel limited if I was going full "vibe" and having the LLM do basically everything for me every day. That said, this is the first LLM product I'm considering personally paying for. The integration with Zed is what wins for me over Claude, where you'd have to pay for API credits or use Claude Code. The way they highlight code changes and stuff really is nice.

Bit of a brain dump, sorry about that!