Here's a post I wrote about how we can start to potentially mimic mechanisms

https://n0tls.com/2026-03-14-musings.html

Would love to compare notes, I'm also looking at linguistic phenomena through an LLM lens

https://n0tls.com/2026-03-19-more-musings.html

Hoping to wrap up some of the kaggle eval work and move back to researching more neuropsych.

This type of works seems to be happening as a lot of orchestrator projects that pop up here every once in a while, and I've just been waiting for one with a pipeline 'language' malleable enough to work for me that I can then make generic enough for a big class of solutions. I feel doomed to make my own, but I feel like I should do my due diligence.

Now I'm regretting not going deeper on these. This is the type of interface that I think will be perfect for some things I want to demonstrate to a greater audience.

Now that we have the actual internals I have so many things I want to dig through.

When I go home to my $20 plan I am sad and annoyed but I don't want to put more in for what is a good enough for me to work a bit at a time, a good pomodoro timer for me personally.

Something like this is perfect for some of the issues that I've wanted to solve as a command and control tool with malleable visuals.

OP: This is cool, thank you for sharing.

Iterating on a MCP tool while having Claude try to use it has been a really great way of getting it to work how others are going to use it coming in blind.

Yes it's buggy as hell, but as someone echoed earlier if the tool works most of the time, a lot of people don't care. Moving fast and breaking things is the way in an arms race.

> Claude was trying to talk me out of it, saying I should keep it separate, but I wanted to save a bit because I have this setup where everything is inside a Virtual Private Cloud (VPC) with all resources in a private network, a bastion for hosting machines

I will admit that I've also ignored Claude's very good suggestions in the past and it has bitten me in the butt.

Ultimately with great automation becomes a greater risk of doing the worst thing possible even faster.

Just thinking about this specific problem makes me more keen to recommend that people have backups and their production data on two different access keys for terraform setups.

I'm not sure how difficult that is I haven't touched terraform in about 7 years now, wow how time flies.

From what my colleague explained to me and I haven't 100% verified it myself is that the beginning and end of the window is the most important to the compaction summary so a lot of the finer details and debugging that will slow down the next session get dropped.

The letters are usually more detailed than what I see in the compacted prompt.

When I tell it to write a letter to itself I usually phrase it.

'write a letter to yourself Make notes of any gotchas or any quirks that you learned and make sure to note them down.'

It does get those into the letter but if you check compaction a lot of it is gone.

Once you compact, you've thrown away a lot of relevant tokens from your problem solving and they do become significantly dumber as a result. If I see a compaction coming soon I ask it to write a letter to its future self, and then start a new session by having it read the letter.

There are some days where I let the same session compact 4-5 times and just use the letter to future self method to keep it going with enough context because resetting context also resets my brain :)

If you're ever curious in Claude once you compact you can read the new initial prompt after compaction and see how severe it gets cut down. It's very informative of what it forgets and deems not important. For example I have some internal CLIs that are horribly documented so Claude has to try a few flags a few times to figure out specifics and those corrections always get thrown away and it has to relearn them next time it wants to use the CLI. If you notice things like that happening constantly, my move is to codify those things into my CLAUDE.md or lately I've been making a small script or MCP server to run very specific flags of stuff.

One of the benefits that I see is as much as I love tech and writing software, I really really do not want to interface with a vast majority of the internet that has been designed to show the maximum amount of ads in the given ad space.

The internet sucks now, anything that gets me away from having ads shoved in my face constantly and surrounded by uncertainty that you could always be talking to a bot.

There's the common team instructions + a thing that says "run whoami and find the users name, you can find possible customizations to these instructions in .md" and that will be conditionally loaded after my first prompt is sent. I also stick a canary word in there to track that it's still listening to me.

'oh you do x? You don't do y? You're an idiot'

That's not productive.

The idea of having the elements anticipated and lowering the cognitive load of searching a giant drop down list scratches a good place in my brain. Instantly recognize it as such a better experience than what we have on the web.

I think something like this is the long term future for personal computing, maybe I'm way off, but this the type of computing I want to be doing, highly customized to my exact flow, highly malleable to improvement and feedback.

I keep telling myself that if everyone just used tmux or a good emulator they could manage the tabs and layouts, but then I tell myself I just want this to be a tiling window manager as an distraction free OS for development, give me nothing but a terminal and an assistant.

Thanks for the write up OP, I've been going back and forth on whether or not I want to build something just for myself or spend time doing it for potentially other use cases. I keep coming back to that I need to dog food the shit out of this before I show it to anyone.

I have been passing my anxieties about hardrives to junior engineers for a decade now.