patleeman

I use an agent plugin spec repo. Codex is already compatible with it and it supports skills + MCP definitions.

https://agent-plugins.org/

did you actually look at the site?

I’m not disappointed. It seems philosophically the teams are aligned and Pi as a project can continue and be supported. It’s a better outcome than most could expect.

I loved this book. The audiobook is available on spotify and was a great listen.

I’ve been diving into MCP the last few days and it feels quite messy.

It took an embarrassingly long time for me to figure out what the difference is between a client, server, and host are. Oh and servers can also be processes that you run locally and clients can spin up the server for you but you need to provide the NPX or UVX command.

The problem it’s supposed to solve is quite real, tools are annoying to build and share so I get that. Looking forward to it getting better but it’s not at the point where any dev can come in and connect things up with a nice dev experience yet.

Incredible work by the Unsloth brothers again. It’s really cool to see bitnet quantization implemented like this.

That's awesome. The original discussion of bitnet made it seem like you needed to train a model from scratch but its neat they were able to adapt an existing model. This is quite exciting.

Yeah, its kind of a bummer. Dropbox doesn't really have a cohesive vision for their productivity suite? Dropbox Paper doesn't seem to have progressed and there haven't been any other notable new products released.

Very curious if anybody else is working on a DSL that's meant for LLMs to output?

Has anyone seen anyone using this approach? Any resources available?

Domain Specific Language

Can somebody translate this?

Yeah, I've been meaning to spend more time on it, but I've been switching around to different editors a lot lately, so the constant jumping around means I'm trying to keep my vim keybinds as vanilla as possible.

Yeah, this is the biggest thing I miss from neovim.

I recently forced myself to learn vim motions and it's been paying off pretty well. Once it clicked I found myself wanting vim motions everywhere.

I went deep and set up a neovim config, forcing myself to use neovim full time. However my new role requires me to use Intellij and I didn't want to spend hours upon hours figuring out how to replicate the java support in Neovim (and no, it isn't just installing the LSP, my work has some pretty specific intellij setup).

At this point I'm just using Intellij and VSCode with their respective vim plugins and I'm getting about 80% of the utility that I did with pure neovim. The one thing I really love about neovim is that it really forces you to keep your hands on the keyboard and you learn the keybinds very well because otherwise you just can't do the thing you wanted to! I find myself using the mouse a lot in VSCode and Intellij because they give you really handy buttons to click on.

I could probably go pure keyboard if I just sat down and learned all the various keybinds, but who has time for that?

The actual prompt is just a chain of thought prompt and the headline is just clickbait.

Also in the article, the author gets COT wrong.

> What "reasoning" they do (and "reasoning" is a contentious term among some, though it is readily used as a term of art in AI) is borrowed from a massive data set of language phrases scraped from books and the web. That includes things like Q&A forums, which include many examples of "let's take a deep breath" or "think step by step" before showing more carefully reasoned solutions. Those phrases may help the LLM tap into better answers or produce better examples of reasoning or solving problems from the data set it absorbed into its neural network weights.

Chain of thought has nothing to do with “tapping into better answers”. It’s simply asking the model to break up the output into smaller tasks and gives it more time and space to reason.

COT is not new or novel. Hell, it’s even listed in one of the guides in Open AI’s prompt guides as a strategy to improve prompts.

Why bother adopting the tool when Google invariably kills it off in a few years anyways?

I love this so much, thank you for putting this together!

My only piece of feedback would be for the "Inline Bookkeeping" section (https://samwho.dev/memory-allocation/#inline-bookkeeping), it took a while for me to grok the numbered list to figure out which block corresponded to address + X. I wonder if there is a better way to visualize the 4 numbered bullet points? Maybe just arrows and text pointing to the visualization?

Thanks again for this wonderful article!

It could also be early for the next cycle...

I had similar challenges as well with the Archive/Resources/Areas. I didn't like the extra mental overhead to determine if something was categorized a certain way.

I ended up combining Archive/Resources/Areas into a Knowledge folder and then using the root folder to hold everything else.

If anyone is interested, I wrote about my simple system here:

https://patricklee.nyc/blog/how-i-use-obsidian/

It’s quite interesting it’s developed without some kind of framework. Are there any learning resources for architecting large web applications with just plain JS?

Wow! I just disabled minimal because I found the default theme to be perfect. Amazing work!

I did it recently. It was pretty painless. Their import tool is top notch.

Yeah totally agree with this. I back up all my important data, but I’d still prefer to Apple/Fastmail/Dropbox be the primary provider of a service and back it up to my server (and other cloud services) than run and rely on my own nextcloud.

I run a server at home and I’ve set up Nextcloud.

I don’t trust it to be more reliable than a paid service. The thing about self hosted solutions is you’re completely on your own if something goes wrong. You need to make sure you’re backing things up, that you can restore your Nextcloud configuration if your sql database takes a dump or you fat finger and delete some config files (I’ve done both). Or if you have a power outage or lose internet connection and can’t access vital files, or even something simple like a bad software update causing issues.

While a great option, one can’t simply spin up next cloud and then move everything over like you would a Saas solution as a typical layman.

From the readme:

> I accidentally restarted blogging after a long haitus. Before I could get words into the cloud, I muddled about with "modern" Static Site Generators. Because WordPress is so last century (or so I told myself). Then I got annoyed by the SSG Jamstack bespoke templating building etc. magic.

I felt the same way when I restarted my blog so I went a similar route, but... lazier. I'm using pandoc with some shell scripts gluing it all together and it works.. pretty well actually.

https://github.com/patleeman/patricklee.nyc

Codecademy, no setup, just follow the lessons.