Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Code & Curiosity by Maximilian Schwarzmüller

Code & Curiosity by Maximilian Schwarzmüller artwork

Don't sleep on the "PI" (coding) agent!

Mar 12, 202615 minAudio
  • Apple Podcasts

There are many AI (coding) agents. Claude Code, Codex ... and many more. They're all great. And they all can actually do more than just coding. That's why the "PI" agent stands out!

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

I gotta talk about yet another coding agent. And I know, I know, I already created videos and courses on Cloud Code and on Codex because they are both amazing, and you also have, of course, Cursor and GitHub Copilot, and I got a course on that too, below. But today, I wanna talk about the Pi coding agent also because it's so much more than just a- a coding

agent. Now, again, all these tools will get you there. There is no wrong or right choice. And I totally get that this can all feel like the framework wars, the JavaScript framework wars from 2019 again, where every week, a new shiny tool is coming out. And yeah, to some extent, it kind of is like that, I guess. But then again, just like back then, if we're very honest, it doesn't matter that much which one you choose. This is also not a sponsored video,

and I don't have a course on that. I just genuinely think this Pi coding agent is a tool you also might wanna look into. Now, unlike Codex and Cloud Code, or OpenCode, for example, this doesn't come with a subscription you can get. With OpenCode, of course, you can also use it without a subscription subscription, like your Codex subscription, With the Pi coding agent, you only have the option of paying per

use or of using another subscription. Uh, for example, here, if I started, I installed it on my system, I'm using it with my Codex subscription, which I could use with the Codex app, the Codex CLI, but which I can also use here. I think you can also use it with the Anthropic, the Cloud Code subscription, but, uh, you maybe heard that they don't like that, and account banned as far as I know. Now what's so special about this Pi thing? Why would you use it instead of the Codex CLI? Couple of reasons. For one,

Pi, this agent, is super lean and simple in a- a positive way. It has a very minimal system prompt, and it only comes with a few tools. Essentially, as far as I know, if that hasn't changed, it only comes with read, write, edit, and a bash tool. And the bash tool, of course, is the powerful one because if you have a bash tool, you essentially

have access to everything because you can, of course, control your entire system, your entire machine with just the bash, with just the command line in the end, because you can invoke a bunch of other And as it turns out, and as I also covered in another video, CLIs seem to be, at least right now, the main thing you wanna expose to your agents, to your coding agents because they are really good at using CLIs, even ones they haven't seen before.

And of course, through CLIs, through command line interfaces or tools, uh, written by other people, your agents can do a bunch of stuff on your computer. They can send HTTP requests. They can r- spin up scripts and execute scripts. They can parse JSON data. They can do all kinds of stuff. And that's kind of the philosophy of this Very minimal, but give it access to the most powerful it can essentially do everything. And combined with a very minimal lean system prompt

that's baked in, you get an agent that has a- a context window that's not cluttered up flexible to do whatever you want it to do, so to say. And their whole philosophy really is that instead of packing a lot of stuff in there, you get an agent that is super extensible. You can install a thing that's called Extensions. We'll get back to that. You can use Agent Skills. So here, I'm talking

about this official standard, which is kind of a standard, I guess. Uh, certain tools implement it differently, course, behind Agent Skills is that you have skill directories and skill MD files, and you have prompts in there, or extra context, however you wanna name it, and those are loaded on demand, not eagerly, but instead lazily on demand depending on which task your agent is working on. Uh, for example, here in this running Pi

session here, I have a bunch of skills loaded, skills I set up, uh, where I played around with certain skills, and certain skills that come in really handy, skill, for example. And if I were to take a look at that, Markdown file with a name, a description. A description is, of course, super important with those skills because the end, what gets an agent to actually activate a a skill. Um, it then loads the rest of this file only if, uh, it decides that the skill is relevant for the given

task based on this name and description here. And then, well, again, that's just some extra context, where, in this case, for my code research skill, agent is interested, like Pi, but also Codex if it loads this skill, how to do code research, how I want it to do that, that it should use the MCPorter tool by Peter Steinberger to use other MCP servers like the DeepWiki MCP server, which is a- an MCP server that can be used to explore GitHub repositories and learn more about them, and some

other tools it should use to do research. And that's another important thing here, talking about MCPs. The Pi agent has no support for MCP. The reason for that is that MCPs tend to really fill up your context window because there's a lot of information about the available MCP tools and resources that must be put into the context window for the AI to be aware of it. And we-... that, that, that team or the person, uh, behind Pye doesn't want. That, that's the reasoning here.

It's an opinion shared by many other people, like myself too, kind of solutions like MCP Search for that, but still, no MCP support here, and you don't need it if you use something like this MC-Porter tool. And when I say use, I mean I'm telling (laughs) the AI how to invoke this MC-Porter tool on the fly when it wants to actually work with the MCP so that this is not something that needs to be installed window. So, you get the idea. It's really minimal and lean,

entire, uh, story or the main thing of this Pye coding agent. I already mentioned it, one other great thing about this agent, though, is its extensibility, and that's not just skills. It is first and foremost, I think, extensions. And the idea behind extensions is coding agent has, uh, first-party support for extending it, for hooking into different parts of the agent, hooking into different steps of the

loop, so when the agent is working, and allowing you to do all kinds of stuff and extending this agent in all kinds of way. And you could, for example, add MCP support if you wanted to. Now, I got a bunch of extensions, uh, here. For example, I added a plan mode through an extension. You can do that. Uh, there is no plan mode built in by default, but it's so extensible that you can add one that blocks the agent from using write or edit tools if it's still in,

um, plan mode. It- this extension allows you to set up a shortcut that allows you to switch into that plan mode. It allows you to update the UI, the terminal UI, to show the user that you're in plan mode. You can also add extra slash commands, like, uh, slash plan, which switches me in plan mode, gives me this indicator here, and now, for example, certain tools would be, uh, blocked, and I can switch out of it again. So that's the kind of stuff you can do with extensions, and there also is

an kind of official extensions marketplace, because this Pye agent also has the idea of allowing you to package up your extensions or skills and share them with others. And other people, like you and me, we can install those packages to, install extensions built by other people. And there are things like sub-agents, which you can add, uh, through a package that includes an extension for adding the sub-agents, or the web access package, which is great, which gives the agent extra tools for doing web research and fetching website

content efficiently. So, you get the idea. It's really extensible. You get a very minimal core, extend it in any way you want, add any skills and any extensions you want, and that is, uh, on its own, already really interesting, because of course, with Cloud Code and with Codex and all these other tools, you get pre-built tools that are way more powerful out of the box, but also less extensible. With Pye, it's the opposite idea. You get a minimal core, and you can then

convert it into anything you want. And you can do that globally, but also on a per-project basis, because all these skills, all these extensions can be installed globally or per project, and that makes it so versatile, and that's why it might be worth a look. I've been using it over the past weeks in addition to Cloud Code and Codex. I'm switching all the time, uh, also because these tools are evolving so quickly, and I really like Pye. Now, here's the interesting thing, though. You are not limited to

just using Pye for coding. It is called a, a coding agent, and that is first and foremost what you will probably use it for, and what I've been using it for, but that is not something you're limited to. So for example, I did install the web access package here, uh, and wired that up to my Gemini API key to this Pye agent, uh, full access to, to web research based on the Gemini API, and therefore, I could run

this agent outside of any project, outside of any coding project. And I could give my agent a task like this: Do some research on the last seven days of the Apple and NVIDIA stocks, fetch their prices, and do an analysis on their seven-day performances. And if I do that, it'll go ahead and do that, and it'll figure out a way of fetching price data for these stocks. It will do a, a web search, maybe visit their investor relation pages, we'll see, and it will

then also, of course, do the actual calculations, uh, spin up a temporary script We'll see what it does. For example, here, it activated a web research skill I added where I give it more details on how it should conduct web research, and then it tries to do that. Now, facing a problem here, because I don't have Python installed, or not the Python executable like this, it would have been python3, and it figured that out too, and then wrote a script where it, um, well, fetched some data from a website with help

temporary script, which it executed here, and looks like it got some stock prices here. And then it will very likely also write a little script that allows it to do calculations based on that and, uh, calculate the, the movement in the stock price. And after a while of working and doing a bunch of research on all kinds of things, including some news research, it is done, and it gives me the last seven days for Apple and NVIDIA, showing me how their stocks developed over

days. Gives me some performance summaries here, it gives me some, uh, verbal analysis here, where it, uh, kind of summarizes its results for me to read through them, like an executive report.And all that was done by this PyCoding agent with the extensions and skills I gave it, but nothing else. No code from my side, no specific instructions which sites to visit. It did that all on its own. And you could absolutely

do that with Claude Code or Codex too, because in the end, these are all AI agents that have a- a bunch of tools to get stuff done, and whilst they're primarily built for can, of course, kind of abuse them to do all kinds of because in the end, of course, this task also involved fetching some websites and writing some scripts, which is very similar to coding projects where you also might want it to fetch some docs and write some code, right? So, it can do a bunch of other things as well.

The reason why Py, in my opinion, is a bit better for tasks like this than maybe Codex or Claude, is that it has this minimal core and can be extended to be exactly the tool you want it to be, even on a per project basis if you want to. So, you could have one project on your system another project, uh, that has, I don't know, a- a stocks research expert, and a third project that has a totally different, uh, expert, an expert for analyzing your system

utilization, where... I don't know. Anything like that. And that's also the reason, probably, why OpenClaw is using Py internally. So, here, I'm on Deep Wiki for the OpenClaw repository. In case you don't know it, learning more about GitHub repositories. It analyzes them and then gives you, like, a on-the-fly created documentation based on the code it sees there, and you can also chat, uh, with the repository, so to say. And here, I could ask, "Is OpenClaw using the PyCoding agent internally? If yes, how?" And it will analyze

that code, which it has already loaded, and it tells me, "Yes, OpenClaw is using the PyCoding agent," and it then tells me how exactly that is implemented. So yeah, that's the PyCoding agent, and I just felt like sharing it. Again, not because, uh, I'm earning anything off that. Don't have a course, have nothing. But it is a great tool, especially if you're planning or if you wanna play around with agents for noncoding tasks, but of course, just to be very clear about this too,

also for coding tasks. It can do both. It's a very versatile AI agent. You can even build your own tools on top of them, as you'll see with OpenClaw, all these things with Codex too, but again, the minimal core really is amazing here. So, in addition to Codex or Claude Code, or instead of them, this might be worth a look and, uh, yeah. It's something I have, uh, had a lot of fun with over the last weeks, and I'm excited to see where all this

and what we can do with them in a year from now. It's also a bit scary, I will say, but also very, very interesting. A- a weird mixture

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. We will have to accept AI-generated content

    Oct 15, 202612 minAudio

    I think we'll have to get used to (and accept) AI generated content. Not to slop though.

    • Apple Podcasts

    Credit the people on this episode

  2. Is software dead?

    Oct 13, 202612 minAudio

    There's a lot of talk about software being dead. Unsurprisingly, it's complicated.

    • Apple Podcasts

    Credit the people on this episode

  3. GPT 6 Astra - Good Beyond 3D?

    Sep 9, 20268 minAudio

    GPT 6 Astra was released. It's a good model, I have to say. Especially for 3D work it seems to be a real game changer. But due to all the hype, you may be disappointed...

    • Apple Podcasts

    Credit the people on this episode

  4. You may ignore Omarchy but not its impact

    Aug 26, 20268 minAudio

    Omarchy got a lot of hype and visibility recently. But even if you don't care or don't use it, you shouldn't ignore it.

    • Apple Podcasts

    Credit the people on this episode

  5. SvelteKit 3, Solid 2 - Do they matter anymore?

    Aug 18, 202614 minAudio

    We're not going back to the JS framework wars. BUT new frameworks and framework versions are important. For a different reason now!

    • Apple Podcasts

    Credit the people on this episode