ok i’m not saying do this

i recently setup an API proxy, C&C server, Grafana and Prometheus, and Discord bot. now i can send pings via Grafana or with a simple request (provided it’s authed via VPN or proxy) and have my Discord bot use a local LLM on my network to deliver the alert to a Discord channel in the voice of Ultron.

but LLMs do represent a significant technological leap forward. i also share the skepticism that we haven’t “cracked AGI” and that a lot of these products are dumb. i think another comment made a better analogy to the dotcom bubble.

ETA: i’ve been working in ML engineering since 2019, so i can sometimes forget most people didn’t even hear about this hype train until ChatGPT, but i assure you inference hardware and dumb products were picking up steam even then (Tesla FSD being a classic example).

i know it’s popular to be very dismissive, but a lot of “AI” has already been integrated into normal workflows. AI autocomplete in development text editors, software keyboards, and question asking bots isn’t going away. speech-to-text, “smart eraser”, subject classification, signal processing kernels like DLSS and frame generation, and so many more will be with us and improving for a long time. Transformers, machine learning optimized chips, and other ML fields are going to be with us for a long time. the comparison to NFTs is either angst or misunderstanding.

the thing that finally got me using a terminal multiplexer was zellij. you can think of it as tmux with training wheels, but i don’t see a reason to go back.

i don’t know the full nature of the exploit, but zlib has an exploitable integer overflow via the MiniZip project. even though our images don’t use that project.

https://github.com/madler/zlib/issues/868

i know people usually are like, “oh cool new features”

but this has a security patch that will literally unblock my pipelines at work lol 🎉

yeah overall bandwidth was probably a consideration

sure!

1. it wasn’t tough to get started. it generally reads like a normal Unix shell with some exceptions. i don’t think many Linux power users would have a hard time doing basic file system tasks or launching programs, etc. there are going to be some issues, like you can’t just paste bash commands in because && isn’t supported, multiline strings don’t require the \ character, and string escaping is totally different. those are intentional deviations that i personally agree with, but they take some getting used to. and then obviously stuff that is specific to nushell like working with tables.
2. definitely the killer feature out of the box is manipulating, parsing, and reading structured data. the “aha” moment for me was when i needed to change a value over a thousand or so JSON objects and did it with a one liner. then i use it with some extra overlays to do stuff like connect to a k8s cluster like k8s connect (helm stage dev.0) which reads my YAML config and connects to the cluster specified in that file. or making a call to our internal package store to get the latest version by parsing the returned JSON.
3. it works out of the box with your existing PATH (or Path if you’re nasty). you can just drop into it and it will have all the path stuff inherited just like if you launched zsh or bash. you’ll have to set that up if you want to use it as a system shell—like i do—, but otherwise it’s pretty seemless.

you can check out my collection of scripts here: https://github.com/covercash2/dotfiles/tree/main/nuenv

ETA: if you do have compatibility problems or need your old muscle memory to do something quick, it’s easy enough to use bash -c old_script.sh or just drop into a different shell

i’m a big nushell fan.

i was once sitting where you are. when PowerShell was released on Linux i thought about switching and read the manual. i really liked some of the philosophy:

• descriptive names for commands. cat and ls have canonical short names to save disk space on the systems they were created for. this is no longer a constraint and aliasing a longer command name is better than “git gud n00b” when it comes to discoverability.
• structured data. “everything is a string” is great when programs play nice. it breaks apart when programs prefer human readable output or worse don’t provide structured output, like —format=json or whatever.
• modern control flow semantics. yes, pipes are great, let’s keep those, but why do i have to rtfm every time i want to bang out a simple script with an if-else control flow?

i looked around at a few solutions. xonsh uses Python. eshell is integrated into emacs and uses Elisp. i briefly tried to hack something together using Kotlin Script. and yeah, i tried PowerShell.

i settled on nushell not just because it fulfilled the above requirements, but also:

• simple data types. string, number, list, record, and table are about the only types you deal with.
• wide support for structured data. JSON, YAML, TOML, CSV, etc have parsers built in. jq and other such tools are made irrelevant because you just load it into nushell query with a unified DSL using common syntax like select and where.

honestly, these are the killer features. there are so many more. context aware autocomplete, modules and overlays, super easy custom completions, extension functions (one of my favorites is git remote open), cross platform (if you’re forced to use Windows), plugins, and i can contribute since i do Rust development for work.

give PowerShell a shot, but i think nushell is the happy medium

i mean, JSON is ugly and has its own problems, but it doesn’t have 6 legal values for boolean types. my opinion is that if you can’t do it in TOML you shouldn’t be doing it in config

ETA: if you like this and it’s useful to you, i don’t want to discourage you. i’ve just been fighting with yaml config all week

loving yaml-based CI/CD

you must be new at this or have some sort of Stockholm syndrome

it’s fine as far as laptop keyboards go. i’ve pretty much given up on laptop keyboards being really satisfying. i use a mechanical when possible

it’s not worth it to me. the battery life is a huge feature, and it does feel like Asahi development has slowed. i have enough computers to tinker with. i bought my Macbook specifically to be an entry point into my other machines, i.e. from the airport or brewery or coffee shop.

maybe when it makes sense to buy a new laptop i’ll find some time and motivation to contribute, but just using Asahi doesn’t really appeal to me.

no it totally does. i use Ghostty even

i completely understand. as a Rust developer that uses Neovim, i have some hills like that too. and if i was more of an OS dev and/or had the time i might be interested to help improve the platform. my last attempt was a Thinkpad, but i had to have an external mouse for that thing, the fans were causing me to fail stealth checks, and the battery was basically a UPS.

i know a laptop that’s amazing in almost every aspect except that it doesn’t run Linux. the Macbook Pro. to me there’s barely any real comparison to be made unless Linux or Windows or the keyboard layout is a hill worth dying on to you.

i have servers and my gaming PC on Linux, but i wouldn’t trade my Macbook with its unified memory, incredible battery life, best in class touchpad, and top notch screen for anything else. Windows is dying, and chip designers (outside of Apple) seem more interested in cashing in on AI than providing a user experience. i was excited to see what Qualcomm would do, but it doesn’t seem like OEMs or Windows are particularly interested in supporting that platform as a next leap forward, while Intel is bleeding on the side of the road and AMD is constantly side-eyeing Nvidia. i think it would be peak irony for Nvidia to come out of left field with a desktop class ARM processor that’s Linux native, but that’s a pipe dream. what the ecosystem needs is a real competitor to Apple that is more focused on desktop machines than enterprise contracts. maybe RISC-V Frameworks will break out in a meaningful way. but it just seems like anything else these days in a compromise based on some biased preference or moral judgement.

anyway all that said i’m glad there’s an ecosystem of people who are stubborn enough to work on this platform. i have my own stubbornness, but i just don’t have the motivation to apply it here

academic fraud has always existed

honestly 8 space indents always felt a bit ridiculous to me. i usually use 4 since it’s more conventional in most languages but could also be happy with 2.

weird hill to die on. use default setting unless you have a good reason not to. the argument itself is a waste of time on projects that want to get things done.