Thoughts

Link based "commentary" style posts, commenting on a web link

844 posts latest post 2026-04-16
Publishing rhythm
Apr 2026 | 18 posts

interesting pricing model from popsql (pronounces Popsicle). At a glance you pay for data retention, want the abiltiy to recall all the queries you ran within the last year, run at a higher frequency, you jump a pricing tier.

such a sick episode with dax.

SST’s free tier will be free as long as aws allows a free tier, their free tier literally costs them nothing.

They talked about keeping SST small, the limitations that brings, but also the number of problems that just go away when you only have 3 people building. Lots of process disappears, everyone can trust everyone, no one needs to wait for approval, everyone is their own PM and just builds cool shit. They don’t have to worry about big costs and making payroll because they are profitable so much higher than their costs.

If they can get through phase one of just being the go platform for a very specific audience of users, and gain marketshare, the ideas of offerings on top of this are endless.

Hatch be flyin.

This new release of hatch includes support for the new package installer uv which is just mind blowing fast compared to anything else we have in python right now.

[tool.hatch.envs.default] installer = "uv"

The other features are cool too, check them out. I’ll probably be using the test runner, but I’ve been waiting for the uv support since uv launched.

Imagine waking up to a $1,300 for running an example project! That sounds like peanuts for a cloud bill but for an individual trying to learn that hits my monthly budget real hard.

That’s what happened to Marciej, make sure you check out the full article and give them a 👏 on Medium if you have an account.

The more I see things come out about aws, the more it makes me sick, and confirm my feelings that I cannot possibly use them for a side project without some real $$ planning to come out of it.

Yes, S3 charges for unauthorized requests (4xx) as well[1]. That’s expected behavior.

...

This page shows how to customize your fastapi errors. I found this very useful to setup common templates so that I can return the same 404’s both programatically and by default, so it all looks the same to the end user.

from fastapi import FastAPI, Request from fastapi.responses import JSONResponse class UnicornException(Exception): def __init__(self, name: str): self.name = name app = FastAPI() @app.exception_handler(UnicornException) async def unicorn_exception_handler(request: Request, exc: UnicornException): return JSONResponse( status_code=418, content={"message": f"Oops! {exc.name} did something. There goes a rainbow..."}, ) @app.get("/unicorns/{name}") async def read_unicorn(name: str): if name == "yolo": raise UnicornException(name=name) return {"unicorn_name": name}


This post sat in draft for months. I stumbled upon it again and found great success returning good error messages based on user...

Jerod (It’s ya boi) and Adam are my favorite tech news nerds, and have the sickest podcasts in tech. Yes plural podcasts they run seven podcasts maybe more. If you want it short and sweet they got the best 15 minutes of tech news each week this is it. My favorite is Ship it, sad to see Gerhard go, but Justin and Autumn are crushing it. Every episode is highly polished and surrounded by the sickest beats in podcasting.

Subscribe to one pod if you want, but I recommend collecting them all with the master feed.

⭐⭐⭐⭐⭐

Small web app to convert html into markdown. Pretty cool idea. I actually want to look into this for reader and see how well it would work. Right now I am just pulling descriptions, but maybe I can pull full web pages, and keep the full intent of the first 200 words or so in the cards.

Damn 2024 is such a shit show, now Devin seems to be out as a complete scam. It’s really teaching us to have skepticism for what you find on the internet. Turns out that when broken down frame by frame much of the description in the video was a straight up lie. Personally it seemed quite plausible that it was percentage points better than the competition, but I was not holding my breath for it to be a hands off engineer.

I learned about the sqlite_master table from this stack overflow answer. This helps make a lot of sense to how sqlite works. The master table contains all the sqlite objects and the sql to create them.

The .tables, and .schema “helper” functions don’t look into ATTACHed databases: they just query the SQLITE_MASTER table for the “main” database. Consequently, if you used

This is a cool snapshot testing tool that automatically creates, and updates test values for you.

Starting with some test code.

from inline_snapshot import snapshot def something(): return 1548 * 18489 def test_something(): assert something() == snapshot()

now if I run pytest my tests will fail because my assert will fail, but if I run pytest --inline-snapshot=create it will fill out my snapshot values and the file will then look like this.

Check your system to see if you are vulnerable to the xz backdoor.

I found this line most pertanent to me.

The xz packages prior to version 5.6.1-2 (specifically 5.6.0-1 and 5.6.1-1) contain this backdoor.

Also it appears that arch is not vulnerable as it does not directly link openssh to liblzma, so the known attack vecotor is not possible. read to the end of the linked article for more.

![[None]]

Install it

{ "ThePrimeagen/harpoon", branch = "harpoon2", dependencies = { "nvim-lua/plenary.nvim" }, config = function() require("waylonwalker.plugins.harpoon").setup() end, },

harpoon config

I found this statement quite intriguing.

multi-cursors are just macros.

This is quite a philisophical video and mostly prime talking about the things that make vim vim, and what prime needs in and editor vs what he can live without.

This is pretty sick, I wanted this early on when I was making lockhart. I wanted to do the git hook thing but could not figure it out and did not know that prepare-commit-msg was a hook that I could use.

Git Hooked Then I remembered! Git hooks! Lol. Why would I have that in my brain - who knows!

I asked claude again, and they whipped up a simple script that would act as a hook that triggers with the prepare-commit-msg event.

This is awesome, cuz if you want to add a git message, you can skip the hook. But if you are lazy, you exclude the message and it will call the LLM.

...

Very interesting article by Sylvain, suggested by Simon Willison.

Definitely some things that I want to come back and try later on.

Here is the TLDR of the whole post

PRAGMA journal_mode = WAL; PRAGMA busy_timeout = 5000; PRAGMA synchronous = NORMAL; PRAGMA cache_size = 1000000000; PRAGMA foreign_keys = true; PRAGMA temp_store = memory;

This is interesting, and something I need to consider. I definitely have an application with slow count queries. I am not sure how to make it better as its not a full count(*) so a count table doesn’t work, nor does counting by index.

...

Inspiring story transitioning into tech from nursing. I also came to tech through a set of circumstances that made it difficult for me to excel at my current job. Looking back it is something that I was always interested in and I was just unsure how to get in, I am so glad that I figured it out, it has been such a great benefit to my family.

I really enjoyed listening to trshpuppy’s journey in through building projects, and choosing tech not based on what she wanted to learn, but what fit the project the best.

great poll of git questions

poll: did you know that in a git merge conflict, the order of the code is different when you do a merge/rebase?

merge:

OTHER BRANCH’S CODE

...

kubeseal is a pretty simple to get started with way to manage secrets such that they can be stored in a git repo and be picked up by your continuous delivery service.

Sealed Secrets provides declarative Kubernetes Secret Management in a secure way. Since the Sealed Secrets are encrypted, they can be safely stored in a code repository. This enables an easy to implement GitOps flow that is very popular among the OSS community.

Great episode covering a seemingly simple topic. What I really benefitted from was hearing all the different use cases, from logging, debugging, to a/b testing, caching, and auth. I hadn’t even thought of it being applied to a router. I thought of it being applied for an entire application. This seems very useful for things like an admin router, all routes would need to have the admin role to get in.

![[None]]

I’ve been using these decorators to modify the behavior of specific routes. It will do things like 404 admin only routes in a way that looks just like fastapi’s default, or only allow certain roles into the route, or redirect unauthenticated users to login.

After listening to yesterday’s syntaxfm I’m now really thinking about middleware and the benefits it might have. middleware would make it easy to apply things like admin to an entire admin router, so you wont forget it on any one admin route. It will look cleaner as the admin checker is only applied once per router, not once per route.

I switched from a 60% vortex pok3r to a 40% corne June, 2021. I can relate to a lot of what Carlos talks about here. I think going from 60%-40% made my journey harder than it needed to be. There’s no going back now, but it took me a really long time to be able to hit all of the numbers and symbols, just figuring out how to do the layout was hard there’s not much space.

I didn’t touch type. I never really used my pinkies, except maybe for ESC, Shift, CTRL, Backspace et al.

I can relate to this, my typing habits were terrible. Shortly before going split ortho I worked on my speed with lots, and lots of practice on keybr and monkeytype. I took my speed from 35wpm to 80wpm with a few months of steady practice. This is one of the best things I did for myself.

Once I got split it dropped down to single digits and slowly rose back up to 80, just barely breaking my PB on monkeytype.

...