Loopy Pro: Create music, your way.

What is Loopy Pro? — Loopy Pro is a powerful, flexible, and intuitive live looper, sampler, clip launcher and DAW for iPhone and iPad. At its core, it allows you to record and layer sounds in real-time to create complex musical arrangements. But it doesn’t stop there—Loopy Pro offers advanced tools to customize your workflow, build dynamic performance setups, and create a seamless connection between instruments, effects, and external gear.

Use it for live looping, sequencing, arranging, mixing, and much more. Whether you're a live performer, a producer, or just experimenting with sound, Loopy Pro helps you take control of your creative process.

Download on the App Store

Loopy Pro is your all-in-one musical toolkit. Try it for free today.

Augmatic GRE and vibe coding.

124»

Comments

  • MrStochasticMrStochastic

    @Augmatic said
    So, use AI while you still can and a naive billionaires are paying for it!

    I've thought this too. This might be the best moment we get with it.

  • PoppadocrockPoppadocrock

    @lukesleepwalker said:

    @Poppadocrock said:
    Regardless of anyone’s individual feelings about vibe coding I think most of us can agree it was pretty brave of @Augmatic to be honest and up front about using it to help create the app. Cheers!

    Not just that, but to come here and provide insights into the process, the thinking, etc. Much appreciated. I don't need the app but I bought it just to support the dev and the continued insights into how software (and importantly, the market for it) will evolve over time.

    Agreed.

  • AugmaticAugmatic

    In the context of "AI and War", I’m glad I chose Anthropic rather than Scam Altman’s product, which I don’t even want to name anymore. I hope Dario Amodei can continue pushing back against Pentagon demands without bankrupting his company.

  • GavinskiGavinski
    edited March 10

    @Augmatic said:
    In the context of "AI and War", I’m glad I chose Anthropic rather than Scam Altman’s product, which I don’t even want to name anymore. I hope Dario Amodei can continue pushing back against Pentagon demands without bankrupting his company.

    Yes, though let's not put Anthropic or Amodei on too high a pedestal either. This is really worth a read:

    https://open.substack.com/pub/garymarcus/p/there-are-no-heroes-in-commercial

  • AugmaticAugmatic
    edited March 10

    @Gavinski said:
    Yes, though let's not put Anthropic or Amodei on too high a pedestal either. This is really worth a read:

    True, but at least he has the balls to draw a line somewhere, unlike other "leaders" who will do anything to increase their wealth 30% year-to-year at minimum.

  • GavinskiGavinski

    @Augmatic said:

    @Gavinski said:
    Yes, though let's not put Anthropic or Amodei on too high a pedestal either. This is really worth a read:

    True, but at least he has the balls to draw a line somewhere, unlike other "leaders" who will do anything to increase their wealth 30% year-to-year at minimum.

    He's definitely better than Altman for sure. Let's not even talk about how bad some of these guys are, Thiel, Alex Karp etc, ugh...

  • espiegel123espiegel123

    @Augmatic : for those of us interested in trying vibe-coding for super simple AUv3, could you write an uber-quick description of what is needed to get started? Was this all done in Xcode and interfacing with Claude in a terminal window?

  • FizzyLizzy27FizzyLizzy27
    edited April 9

    @espiegel123 said:
    @Augmatic : for those of us interested in trying vibe-coding for super simple AUv3, could you write an uber-quick description of what is needed to get started? Was this all done in Xcode and interfacing with Claude in a terminal window?

    I haven’t done anything AUv3 development, but I did get a graphic design app shipped. I used a lot of “autonomous agents” in development on and off over a year.

    Disclaimer: I come from a Python/MATLAB background for engineering/math work. I use the base paid versions of ChatGPT and Claude. I’ve done a lot of work as a code lead and software testing. That experience helped a lot I think, or it just shaped how I approached the project.

    Here’s my app:
    https://apps.apple.com/us/app/colorcondenser/id6760206595

    I use a docker environment on a fresh git branch and put Claude in YOLO mode in a terminal by using the —dangerously-skip-permissions flag. Use at your own risk. I also have ChatGPT active in Xcode, and even a few local models for quick reference.

    My workflow has been using ChatGPT to plan out an app, ClaudeCode in YOLO mode to build the scaffolding, and ChatGPT to iron out small bugs once things are running. Claude’s limit is eaten up quick if you’re on the base plan, so using ChatGPT frees that up.

    For planning with ChatGPT I explain my app idea in as much detail as possible. I had a very clear idea and I’m explicit about UI (colors, windows, layout), workflow (tap x, then y should happen. If you try to do y before x, give this error message), and it’s usually good about asking for specifics.

    I then get ChatGPT to generate a prompt for Claude by telling it “I need a prompt for Claude code running in YOLO mode. Make sure it meets the requirements we discussed. Use the prompting best practices guide” and give it a link to the best practices. Here’s the link:

    https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/claude-prompting-best-practices

    Then I give that prompt to Claude Code and it chugs along. It rarely gets things perfect the first time but you’d be surprised. At this point I still tell Claude Code what errors come up. Usually it can tell from context, so saying “there’s a bunch of errors” it can generally go figure things out. If it’s something off in the GUI I’ll screenshot it and describe the issue. The more explicit the better, and if you know how to explain the code side that helps. An example was formatting window placement. It kept messing up spacing until I said something like “don’t use a set spacing amount, use dynamic spacing”.

    Once things are set up I’ll test it out and manually spruce stuff up. Cleaning up text phrases, checking functions, that kind of stuff. Little errors here I’ll use ChatGPT in Xcode because it’s cheaper than burning Claude tokens and right clicking and having AI generate a fix is usually as easy as fixing a word with spell check.

    I think the planning phase is the difference between “vibe coding” and “coding with autonomous agents”. You’re really figuring out the logic of the code and bringing it to Claude to do the scaffolding. After a session I’ll ask Claude to write a dev log summary and future plans to reference and pass to other agents in later sessions.

    When adding big features tell the agents specifically what not to change. When setting up, tell it the file structure you want. Don’t let it build the whole app in ContentView.swift.

    Weird things can happen. There was a time Claude tried to test something and somehow stuck a download command in a loop, so it downloaded like 20 versions of iPadOS 16 or something. Kind of funny, but also a reminder to keep an eye on things.

    If there is code on git you want to reference repomix helps condense things to pass to an agent. I didn’t use this in my app, but I’ve used it on some others I’m playing around with.

    https://repomix.com/

    This video was helpful:

    Hopefully this helps - sorry if it’s not as relevant for AUv3 dev.

    Edit: I’ve been using an M2 Mac Mini for all this

Sign In or Register to comment.